GMI Computing US Inc. announced that it has raised $223 million in a round of funding led by new investor ARCHIV on September 30, 2026. The transaction also included participation from new investors NVIDIA Corporation, DSC Investment, Trend Micro Incorporated, KB Investment Co., Ltd, Kyobo Life Insurance Co., Ltd., KT Corporation and others. The company issued convertible preferred shares in this transaction.

Delta Electronics announced a collaboration with NVIDIA to accelerate the development of smarter, safer, and more efficient next-generation autonomous driving systems. This collaboration combines Delta's superior expertise in automotive energy management, xEV powertrain systems, in-vehicle high-performance computing (HPC) capabilities, and system integration with NVIDIA Hyperion and the NVIDIA Halos safety system. This collaboration is expected to strengthen the synergies between the two companies in autonomous driving across technology, ecosystem development, and market opportunities. Delta will combine its automotive prowess and system integration capabilities with NVIDIA Hyperion supported by NVIDIA's dedicated engineering team, to elevate the development of autonomous driving solutions. Together, the two companies aim to develop AI-driven autonomous driving solutions and accelerate growth in the smart mobility market.

NVIDIA Corporation Presents at SCTE Tech Expo26, Sep-29-2026 . Venue: Atlanta, Georgia, United States.

NVIDIA Corporation Presents at American Growth Summit, Sep-29-2026 . Venue: The Willard InterContinental Grand Ballroom, 1401 Pennsylvania Ave NW, Washington Dc, District Of Columbia, United States. Speakers: Bruce Andrews, Chief External Affairs Officer.

NVIDIA announced NVIDIA Open Agent Safety Platform, an open software platform and reference system design to strengthen AI security from agent testing to deployment, with full-stack governance and control across software and the hardware, compute and robotics systems that run agents. NVIDIA Open Agent Safety Platform enables full-stack governance and control across the software that runs agents, the hardware and compute layers that power their work, and the robotics systems that execute tasks in the physical world. Organizations can deploy elements of NVIDIA Open Agent Safety Platform according to their unique requirements. It includes NVIDIA OpenShell secure runtime software that sets boundaries for agents running on CPUs. As agents take on more work across more systems, enterprises need an enforceable boundary outside of the model and agent harness. Now broadly available, OpenShell provides a secure runtime boundary for controlling how autonomous AI agents execute tasks across open and closed models. OpenShell delivers this protection with minimal overhead on NVIDIA Vera, the first purpose-built CPU for agentic AI. Together, OpenShell and Vera enable agents to operate securely while completing their work as quickly as possible. As open source software, OpenShell can also be extended to work with third-party compute platforms, including those from Arm and Intel. The NVIDIA Open Agent Safety Platform reference system design features NVIDIA Sentry, an out-of-band watchdog that runs on NVIDIA BlueField®-4 DPUs to continuously monitor agent behavior. Sentry provides in-silicon security enforcement, meaning that if an AI agent attempts to move outside its software boundary, Sentry quarantines and stops it in milliseconds. Running on BlueField-4 DPUs, Sentry continuously monitors agent activity and enforces security policies independently in silicon. It combines threat detection, hardware-based agent governance and enforcement and data access protection from an isolated, out-of-band trust domain that is responsive in real time and invisible to agents and attackers. Sentry is built on NVIDIA DOCA software, which provides the programmable capabilities Sentry uses to inspect agent requests and responses, provide attested telemetry, verify agent identity and enforce granular, zero-trust access policies for data, tools, application programming interfaces and services.

TrendAI announced its support for the NVIDIA Agent Safety Platform, NVIDIA's full-stack infrastructure for running autonomous AI agents safely at enterprise scale. TrendAI complements that enforcement layer with the intelligence that informs what the policy should be, the visibility to see what agents are doing, and the threat detection and response to act on it. TrendAI Vision One™ extends NVIDIA's agent safeguards across the whole AI factory, covering agents, models, containers, storage and networks on one platform: Model protection. Inline inspection of prompts and responses blocks prompt injection and jailbreaks and redacts sensitive data before it leaks. Harness and tool governance. Security teams control which tools and MCP servers agents can reach, inspect every tool call, and analyze how skills and AI artifacts behave at runtime. Data and identity controls. TrendAI Vision One™ finds and classifies the sensitive data that feeds AI, stops agents from reaching data they shouldn't, extends file security and data loss prevention to the files agents read and write, and strips excess privilege from agents and other non-human identities. Hardware-rooted detection and response. Using NVIDIA BlueField DPU and DOCA telemetry, TrendAI Vision One™ detection stays trustworthy even if the host OS is compromised, and correlates AI signals with endpoint, network, and identity telemetry so a hijacked agent cannot move laterally unseen. Together, TrendAI™ and NVIDIA are leading the future of AI with proactive security designed to inspire innovation and eliminate risk.

Bedrock Data announced it has extended Agent DLP™, its runtime data loss prevention for AI agents, to NVIDIA OpenShell, the open source agent runtime that is part of the NVIDIA Open Safety Platform. NVIDIA OpenShell gives autonomous agents a secure environment to run in with strong isolation and a zero-trust architecture where nothing is permitted by default. This integration offers an additional policy evaluation layer where Bedrock Data evaluates a request payload once the request is authorized by OpenShell. Bedrock Data’s authorization layer evaluates the payload in the request against an enterprise’s data policies. Bedrock Data evaluates an agent's action against data policy and tells OpenShell when the data involved is or isn't cleared for the destination; OpenShell can block the action accordingly. NVIDIA OpenShell gives each agent a sandbox to do their work, with kernel-level enforcement over their processes, filesystem and outgoing connections, and checks outbound actions against policy before they take effect. OpenShell’s Supervisor Middleware is the extension point Bedrock Data leverages for the additional data-based policy enforcement. Bedrock Data evaluates each request against the enterprise's data policies, which are defined on Bedrock’s Metadata Lake: an enterprise knowledge graph that connects what data is, where it lives, where it came from, who can access it and how it is used. The decision is based on what the graph knows about the data in the request and where it is going, not on patterns in the payload, so a restriction travels with the data as it passes from one agent to another. Bedrock Data performs three functions for OpenShell: At runtime, it evaluates each action and returns the decision OpenShell enforces. A write of sensitive data to a shared scratch table is stopped, while the same write to a table restricted to that data's approved audience proceeds. At design time, it shows security teams an agent's blast radius, such as what data its permissions can access, what that data combines into and where policy would be violated. It makes OpenShell's enforcement data-aware. As data and access change, the decisions OpenShell enforces change with them, without anyone rewriting a rule. Policy applies at the OpenShell sandbox boundary however an agent acts, through an MCP tool, a command-line invocation or code it writes itself, and the same data policies govern every agent that runs there. Teams can run it in observe-only mode first, recording the decisions Bedrock Data returns without enforcing them, and turn on enforcement when they are ready.

On September 28, 2026, the company announced an increase in its equity buyback. The company increased its authorization by $150,000 million, thereby increasing the total authorization to $391,278.77million.

IonQ, Inc. announced that it will be the first on-premise quantum computing deployment at the NVIDIA Accelerated Quantum Research Center (NVAQC). The NVAQC is designed to address the fundamental challenges of scaling quantum processors into a new class of accelerated quantum supercomputers. Connecting IonQ’s full-stack quantum systems directly with NVIDIA AI infrastructure will help drive the new breakthroughs needed to turn qubits into useful, large-scale quantum supercomputers capable of solving the world's hardest problems. As part of the partnership, an IonQ Superion 256 system will be installed at the NVAQC and directly linked with an NVIDIA GB200 NVL72 system via NVIDIA NVQLink, with workloads orchestrated by the open NVIDIA CUDA-Q platform. The joint research program will focus on hybrid software development and large-scale system prototyping, generating open results and a guide for future AI use cases and quantum-GPU co-design. Alongside quantum-GPU integration, joint research is expected to target quantum algorithms portfolio optimization and risk modeling in financial services; materials science; and computational chemistry for drug discovery. These types of complex use cases will benefit from research conducted using the Superion 256 system. The Superion 256 system is available to order now, with first customer deliveries expected in 2027. Installation at NVAQC is scheduled for next year.

NVIDIA Corporation Presents at Advanced Vision & AI Conference 2026, Sep-23-2026 . Venue: Santa Clara, California, United States. Speakers: Alvin Clark, Product Marketing Manager.

Basecamp Research Ltd announced that it has received $140,000,000 in a round of funding led by returning investor, Section 32, LLC on September 23, 2026. The transaction included participation from new investors, Catalio Capital Management, LP, European Tech Collective, Firebrand River Capital, LLC, Inception Fund, King Philanthropies, Inc., NATO Innovation Fund a fund managed by NATO Investment Fund, NVIDIA Corporation, PostScriptum Ventures Limited, Redalpine Venture Partners AG, The Rockefeller Foundation, Singular Capital Partners, Sovereign AI, Menlo Anthology Fund a fund co-managed by Anthropic, PBC, Investment Arm, and Menlo Ventures Management, L.P., returning investor, True Venture Management, LLC, senior leaders from across pharma, biotech, and global industry, including André Hoffmann. The company has issued convertible preferred shares in the transaction. As a part of the transaction Andy Conrad from Section 32, LLC will join the board of directors of the company. The round is raised at a post money valuation of $800 million.

QAD | Redzone unveiled that it plans to integrate technology from NVIDIA, combining NVIDIA’s accelerated computing and AI with QAD | Redzone’s manufacturing expertise and operational context spanning ERP, connected workforce, supply chain, production, and quality. This combination of technology and manufacturing expertise will help shape a federated intelligence layer that can understand manufacturing context across systems without requiring manufacturers to replace the technology they already run, connecting intelligence wherever it lives. Manufacturing Intelligence is expected to be in market in the Spring of 2027.

NVIDIA Corporation Presents at Humanoids Summit, Sep-22-2026 . Venue: COEX Convention Center, Seoul, South Korea.

Einride AB announced a collaboration with NVIDIA Corporation to further accelerate the transition of the Company’s extensive freight network to autonomous operations. Einride AB will adapt the NVIDIA DRIVE Hyperion platform for heavy-duty trucking, working directly with NVIDIA Corporation to extend the platform's compute, sensor, software and safety architecture to the demands of heavy-duty freight. Building the next generation of the Einride Driver on NVIDIA Hyperion platform will enable Einride AB to scale autonomous deployment across a freight network that's already serving customers. Einride AB designs, builds, and operates the autonomous-driving system end-to-end, including safety validation, regulatory approval, and contracted customer deployments, while NVIDIA Corporation provides the underlying compute platform and AI development tools that support the system. Built on the NVIDIA Hyperion platform, a production-ready compute and sensor reference architecture for Level 4 autonomy, and integrated with the NVIDIA Halos safety system, the collaboration creates a scalable foundation that accelerates safe, production-ready autonomous freight operations for Einride AB. Einride AB intends to deploy NVIDIA Blackwell architecture at scale with an NVIDIA Exemplar Cloud partner, a cloud provider whose infrastructure NVIDIA Corporation has validated for large-scale AI training performance, to train, test and refine its autonomous driving models. Einride AB will adapt the NVIDIA Hyperion platform for heavy-duty trucking, working directly with NVIDIA Corporation to extend the platform's compute, sensor, software and safety architecture to the demands of heavy-duty freight. As part of its AI development workflow, Einride AB is using NVIDIA Cosmos to help improve the safety and reliability of its autonomous vehicles. NVIDIA Cosmos enables Einride AB to search and curate camera data to identify complex edge cases, as well as augment real-world data with photorealistic synthetic scenarios that increase the diversity of data used for AI training and validation.

ASUS and Poesis announced the results of a week-long experiment that successfully deployed agentic AI to trade autonomously in financial markets. The system ran entirely on the ASUS ExpertCenter Pro ET900N G3, a deskside supercomputer built on the NVIDIA DGX Station platform, powered by NVIDIA GB300 Grace Blackwell Ultra Desktop Superchip. ASUS and Poesis completed a week-long experiment in which agentic AI agents independently conducted investment research, managed risk, and executed trades using live capital and predefined mandates. Throughout the test, all models, agents, and workflows ran locally on the ASUS ExpertCenter, showing that agents can function in live markets without relying on cloud-based AI infrastructure. The collaboration offers early insights into how local, high-performance computing can support autonomous investment workflows and the ongoing development of agentic AI systems for financial markets. The week-long experiment ran on the ASUS ExpertCenter Pro ET900N G3, powered by NVIDIA GB300 Grace Blackwell Ultra Desktop Superchip. Featuring up to 20 petaFlops of AI performance and 748GB of coherent memory, the system enabled Poesis to deploy a multi-agent AI workflow locally on a single machine. Using live capital, the agents autonomously handled investment research, risk management, and trade execution while operating within predefined mandates.

Saturn Cloud announced an integration with NVIDIA Run:ai, the AI workload and GPU orchestration platform. The integration helps GPU cloud operators earn more from their installed capacity by turning fleets that would otherwise be rented by the GPU-hour into multi-tenant inference products sold by the token. NVIDIA Run:ai is one of several NVIDIA technologies Saturn Cloud builds on, following its integration of NVIDIA DSX AI Factory Platform NVIDIA Run:ai maximizes GPU utilization across the fleet through NVIDIA KAI Scheduler, providing gang scheduling for NVIDIA Grove-managed distributed workloads and policy-driven governance. Saturn Cloud runs on top, giving operators the commercial layer between their fleet and customers so they can sell inference without building it themselves. Beyond per-token serving, operators get multiple products to sell from the same infrastructure — GPU hours, per-token inference endpoints, and managed fine-tuning jobs — without building any of it in-house. They can offer dedicated GPU capacity to customers who bring their own stack, per-token model-as-a-service to customers who just want an API endpoint, and self-service development environments for teams that build and fine-tune their own models. Saturn Cloud also handles model onboarding, shared and dedicated isolation tiers for regulated customers, and the identity, access, and governance controls enterprise buyers require. Underneath, the serving layer runs on NVIDIA Dynamo, handling distributed inference with disaggregated prefill and decode, and drives vLLM, SGLang, or NVIDIA TensorRT-LLM. NVIDIA Grove orchestrates the multi-node workloads, and NVIDIA KAI scheduler places them with GPU awareness, allocates fractional GPUs, and enforces quotas across tenants. NVSentinel and NVIDIA Fleet Intelligence handle fleet-wide health monitoring and fault remediation, so degraded hardware is automatically pulled from placement. The NVIDIA Run:ai-integrated Saturn Cloud platform is available now.

Element451's collaboration with NVIDIA gives it a direct hand in shaping the AI models built for higher education, beginning with a smarter approach to student data management. The first initiative to come out of that collaboration is centered on agentic deduplication capabilities, built to help resource-constrained institutions manage student data at scale. Element451 is working with NVIDIA to bring open-source AI models into its agentic deduplication engine. Duplicate records aren't just a data hygiene problem. When a student appears twice in a system, they can get conflicting messages, miss a financial aid deadline, or be routed to the wrong advisor. Resolving those records faster means each student sees one accurate view of their status and one consistent set of next steps. Since the introduction of deduplication, Element451 has seen a 53% increase in user-confirmed merges as staff move through duplicate records with more speed and confidence. By integrating with NVIDIA's Nemotron 3 Super model, Element451 can build on that success with an open-source approach that reduces the cost of running the underlying models by up to 7x, making it easier for resource-strapped institutions to put AI to work. For a small admissions or student success team this can make a meaningful difference, as every hour staff spends chasing down a mismatched record is an hour they aren't spending with a student. Cleaning up that administrative overhead frees advisors and enrollment teams to focus on the work only they can do, helping students choose a major, map out a degree plan, and stay on track to graduate. Deduplication no longer only lives in its own silo as a separate cleanup task. It's now part of everyone's everyday workflow. The new match signals, confidence score, Bolt Analysis, and AI-recommended master record give much more context to make decisions faster and with more confidence. Element451 and NVIDIA are exploring further collaboration to source and tune other models in the Nemotron family for additional use cases, including Element451's fraud detection capabilities, which help prevent malicious application activity, an increasingly pervasive and challenging problem for colleges and universities nationwide. Element451 delivers AI products purpose-built for higher education. Bolt is the AI Agent Engine for higher education, deploying specialized agents across every stage of the student lifecycle to drive outcomes for institutions and students alike. Element is the AI-native CRM for higher education, connecting data, workflows, and AI agents across every team and every stage of the student journey. Element451 supports more than 350 institutions in powering over 60 million student journeys, including Columbia University, Louisiana State University, Tulane University, the North Carolina Community College System, and the University System of Georgia.

On September 17, 2026, Crusoe Technologies LLC closed the transaction. The company raised $3,900,000,000 in a round of funding co-led by new investors Atreides Management, LP, Valor Management LLC and Mubadala Investment Company PJSC. The transaction included participation from new investors The Founders Fund, LLC, TPG Capital, L.P. Baillie Gifford & Co., GIC Private Limited, BDT & MSD Partners, LLC, Investment Arm, Clal Insurance Enterprises Holdings Ltd., Darsana Capital Partners LP, Era Funds, Fundrise, LLC, Robinhood Ventures Fund I, fund managed by Robinhood Ventures DE, LLC, Semianalysis, XN LP, Qatar Investment Authority, individual investor, Van Eck and returning investor, NVIDIA Corporation, Radical Ventures Investments Inc., Accounts advised by T. Rowe Price Group, Inc., 1789 Capital Management, LLC, Activate Capital Partners, Altimeter Capital Management, LP, ARK Investment Management LLC, Avenir Management Company, LLC, BAM Elevate Manager, LLC, DPR Construction, Inc., Fidelity Management & Research Company LLC, Galvanize Climate Solutions LLC, M37 Management LP, Oman Investment Authority, Polychain Capital LP, Ribbit Management Company, LLC, Salesforce Ventures, LLC, Squarepoint Capital LLP, StepStone Group Inc., Tiger Global Management, LLC, Upper90 Capital Management, LP and Zigg Capital. The company issued convertible preferred shares in the transaction. The company received $3,108,543,889, which include the amount of certain previously issued simple agreements for future equity, pursuant to exemptions under Regulation D from 108 investors. The transaction was raised at a post money valuation of $30,900,000,000.

Snap Inc. is working with Salesforce, Amazon Web Services (AWS), NVIDIA, Trifork, Hololight, and more to bring SPECS into workplace environments. Together with these partners, Snap Inc. is enabling hands-free access to information, guidance and collaboration across retail floors, factories and field operations, while also building the management and security capabilities enterprises need to deploy SPECS AR glasses at scale.

Matic Inc. has become among the first to adopt, explore, and scale on NVIDIA Corporation Jetson Orin Nano 2, a new entry-level robotics computer for entry-level edge AI coming in early 2027. Matic Inc. will use it to enable more intuitive interactions with its home cleaning robots, including conversational AI, gesture detection, deeper context of the home through precision mapping and semantic understanding, and autonomous cleaning. NVIDIA Corporation Jetson Orin Nano 2 is a robotics computer that enables Matic Inc. to be the first completely unsupervised, fully autonomous, vision-first home robot that processes all camera data on the edge device, not cloud. Jetson Orin Nano 2 delivers 78 trillion operations per second of AI compute, twice the inference performance of its predecessor in the same form factor, and consumes 40 percent less power at the same performance in 15-watt mode. The NVIDIA Corporation Jetson Orin Nano 2 module and developer kit are expected to be available in the first half of 2027. Matic Inc. will begin working on that once they are available and incorporate them to be included 6-9 months after they are available. Matic Inc.'s current products ship on NVIDIA Corporation Jetson today.

IonQ, Inc. announced joint research with Oak Ridge National Laboratory, NVIDIA Corporation, and the University of Tennessee, Knoxville showing that a trained generative model can write quantum optimization circuits directly, eliminating the trial-and-error parameter-tuning loop that has made the most accurate approach too costly to run. The paper is being presented at IEEE Quantum Week in Toronto. Hybrid quantum optimization breaks a large problem into smaller pieces, solves each one, and recombines the results. Each piece needs a tailored quantum circuit, which traditionally required trial-and-error parameter tuning: run, measure, adjust, and repeat, often hundreds of times. Larger pieces can improve answers, but they also raise tuning costs, limiting the size of problems researchers could solve. In this benchmark, generative AI replaced the iterative tuning loop, and as the quantum subproblems grew the solution quality improved. The result provides a potential path toward scaling hybrid quantum optimization, unlocking completely new capabilities and scales that align with IonQ’s existing and future quantum computing hardware generations. To teach the generative model how to tune circuit instructions, the team first showed it what strong results looked like. They ran the conventional trial-and-error method across many sampled problems, kept only the near-optimal circuits, and used those examples to train a transformer, the same class of model behind large language models but trained on circuits instead of text. The trained model generates candidate quantum circuits directly, without the repetitive parameter-tuning loop used by conventional methods. In the reported experiments, the model sampled ten candidate circuits for each subproblem; all ten were simulated and scored, and the best-scoring candidate was used to update the global solution. On a dense, higher-order benchmark problem with 100 decision variables, model-generated answer quality roughly doubled as subproblems grew. Under the prior state-of-the-art method, circuit-finding time rose sharply from about 34 seconds on 4 qubits to more than 11 minutes on 12 qubits. The generative approach held nearly 28 seconds for every size previously tested. Both approaches are quantum methods; the study compares circuit-generation approaches, not quantum against classical solvers. Oak Ridge National Laboratory led the study. Co-authors span Oak Ridge National Laboratory's National Center for Computational Sciences and its Materials Science and Technology Division, IonQ, Inc, NVIDIA Corporation, and the University of Tennessee, Knoxville. Abhinav Rijal, a graduate researcher in the Department of Physics and Astronomy at the University of Tennessee, Knoxville is a co-author of the study. The paper describes its results as benchmark-scale validation. Every circuit in the study was simulated rather than executed on quantum hardware, using the NVIDIA cuQuantum SDK through the NVIDIA CUDA-Q open platform for quantum-GPU supercomputing on a single NVIDIA H200 GPU in the Oak Ridge Leadership Computing Facility’s Defiant2 system. That environment is what makes the comparison meaningful: both the trial-and-error method and the generative one ran on identical GPU-accelerated infrastructure, providing a controlled comparison of the two end-to-end workflows. The measured difference principally reflects the distributed quantum approximate optimization algorithm (DQAOA)-GPT’s replacement of iterative variational parameter optimization with generative circuit synthesis and a fixed number of candidate evaluations. The paper is available at arXiv:2607.20225. It is one of nine IonQ, Inc. papers accepted at IEEE Quantum Week 2026, held September 13–18 at the Metro Toronto Convention Centre, and the winner of a best paper award.

Emerald AI, Google LLC, and NVIDIA Corporation announced the launch of the AI Energy Management Alliance (AEMA). Led by founding members representing the full AI stack, AEMA advances flexible AI data centers that are good citizens on the electricity grid, dynamically adjusting their power draw to provide relief during periods of peak grid stress. The founding members are joined by a group of 18 launch partners, spanning AI and semiconductor leaders, such as Anthropic and Analog Devices, and utility and power leaders, such as National Grid, AES, RWE, Constellation, and NRG. Together, the membership will develop diverse approaches to power-flexible AI to unlock power capacity on today’s underutilized power grid, improve its reliability, and promote affordability by better using the grid we already have. AEMA will promote accelerated interconnections across the United States for flexible, grid-enhancing data centers using a diverse range of demand management technologies including colocated generation, storage, computational flexibility, and other approaches to dynamically respond to the grid’s needs and better utilize the existing power system. By championing policies that recognize and value flexible AI demand, AEMA supports the nation’s competitiveness in AI innovation and deployment while strengthening grid reliability and protecting energy affordability for ratepayers. Joining founding members Google LLC, NVIDIA Corporation, and Emerald AI as inaugural members and launch partners are AES, Analog Devices, Anthropic, Calibrant Energy, Camus, ClearPath, Constellation, Encoord, Fluence, Generate Capital, GridUnity, National Grid, NRG, PassKey, RWE, Splight, Verrus, and Voltus. Frank Lacey, an experienced energy industry executive, will serve as AEMA’s Executive Director. AEMA is the next evolution of the Advanced Energy Management Alliance, founded in 2014 to advance energy solutions while improving the efficiency of the electric grid.

Salesforce, Inc. announced that it has developed Koa, Salesforce’s first CRM reasoning model for Agentforce, built on NVIDIA Nemotron. Developed through deep technical collaboration with NVIDIA Corporation, Koa is purpose-built to help agents reason through complex, multi-step workflows and use the right tools to get work done. Koa was built by post-training NVIDIA Nemotron 3 Super with a proprietary synthetic dataset modeled on enterprise knowledge from nearly three decades of CRM deployments. Salesforce, Inc. and NVIDIA Corporation are also bringing Nemotron-based models and accelerated computing into Missionforce, extending that same control over model, data, and deployment environment to government and regulated organizations. Together, the companies are bringing mission-specific AI to private clouds, air-gapped networks, and other secure environments. Salesforce, Inc. applied Supervised Fine-Tuning (SFT) and reinforcement learning with Group Relative Policy Optimization (GRPO) with NVIDIA NeMo RL, NeMo Gym, and NeMo AutoModel. NVIDIA Nemotron open models gave Salesforce, Inc. the control to build CRM intelligence into the model itself. Salesforce, Inc. controls the weights and runs Koa within its own infrastructure, ensuring that no customer data crosses the trust boundary during inference. Salesforce, Inc. is bringing NVIDIA Corporation models and accelerated computing into Missionforce to enable customers in highly sensitive industries to train, tune, and deploy mission-specific models on their own critical data. Post-trained NVIDIA Corporation models will power agents for Missionforce Operations, a product that digitizes and automates complex government workflows, including procurement, supplier management, and logistics. The companies are also working to bring post-trained NVIDIA Corporation models into additional Missionforce capabilities to enable customers to build and run models on their own infrastructure for specialized workloads.

Synaptics Incorporated announced its first CTS module powered by the SN6012T touch controller is now supported by the open robotics simulation framework, NVIDIA Isaac Sim. The Synaptics CTS module is designed for dexterous robotic hands, grippers, and other Physical AI systems to detect contact and pressure distribution, supporting more precise grip control, object manipulation, and responsive interaction. The CTS module works effortlessly with Synaptics Astra® Edge AI processors. These processors serve as a multimodal sensing, inferencing, and connectivity hub unifying touch, vision, motor control, and other sensory signals into synchronized streams to facilitate hierarchical AI systems with support of NVIDIA Holoscan. As Physical AI platforms increasingly operate in highly dynamic environments and interact directly with humans and objects, developers need simulation environments that capture how contact and force behave at the point of touch. This simulation environment delivers a way to prototype tactile interactions and generate synthetic sensor data. It helps developers expedite validation cycles by enabling perception algorithms before deploying on physical hardware. The CTS platform is expected to provide: An optimized standard tactile module: High 2.5mm pitch taxel density with a 5x12 taxel array, supporting force ranges up to 100N with 0.1N resolution. Faster response, low-hysteresis proprietary sensor design: 10x faster rebound response than traditional foam sensors. Native Newton engine support: Built from the ground up for NVIDIA’s Newton physics simulation engine. The Astra AI-native processor platform is designed to: Aggregate touch, vision, and motion data into unified streams, delivered to NVIDIA Holoscan over standard Ethernet. Offload sensor collection and synchronization from central compute, so robotic platforms can interpret contact, grip, and movement locally. Bring localized Edge AI intelligence to humanoids, grippers, and industrial systems, cutting integration complexity.

Nscale Limited entered into subscription agreement and will receive in an aggregate principal amount of a minimum of $3.1 billion, comprising of $2.1 billion of unsecured convertible loan notes and a further $1.0 billion of unsecured convertible loan notes or non-voting shares on September 15, 2026. The transaction will include participation from new investor NVIDIA Corporation for $1.0 billion. The transaction is expected to close on November 16, 2026. The company will issue unsecured convertible notes in this transaction. The convertible loan notes are convertible into ordinary shares automatically upon completion of the initial public offering. Unless earlier converted or repurchased, the convertible loan notes mature on June 15, 2028.

Qedma Quantum Computing announced an integration of QESEM software with the NVIDIA CUDA-Q platform. Qedma's integration with CUDA-Q is currently available for use with Quantinuum hardware and will be available for other quantum platforms in the future. Qedma's advanced error mitigation software, QESEM, mitigates the impact of noise present in today's quantum computers. CUDA-Q users can now incorporate Qedma's advanced error mitigation capabilities to their quantum workloads directly within the CUDA-Q environment, without changing their existing programming interface or workflow. Qedma recently demonstrated quantum advantage together with IBM, accurately resolving the dynamics of quantum materials unseen in classical simulation. The results marked the first time quantum advantage has been achieved with commercially available hardware and software, reaching a paradigm where multiple classical approaches failed to provide consistent, reliable answers.

NVIDIA announced an expansion of the NVIDIA CUDA-Q open source platform with CUDA-Q Logical, an orchestration layer that provides a programmable, verifiable approach to developing useful applications for fault-tolerant quantum computers. The NVIDIA CUDA-Q platform now includes the CUDA-Q Logical orchestration layer, giving the quantum industry an open, programmable way to design and test fault-tolerant quantum computing applications in areas such as drug discovery, financial modeling and materials development. Fault-tolerant quantum processors with logical qubits are essential for useful quantum computing, allowing systems to overcome the errors inherent in physical qubits so they can execute the larger computations needed for practical applications such as drug discovery, financial modeling and materials development. With CUDA-Q Logical, researchers can now design and orchestrate the many components required for fault-tolerant quantum computing applications, easily switching between options to identify optimal system configurations for performance with logical qubits. CUDA-Q Logical is already being used by QPU makers and labs including Fermi National Accelerator Laboratory, Infleqtion, IQM Quantum Computers, QCDesign Quantum Motion and Sandia National Laboratories. Using CUDA-Q Logical, Iceberg Quantum modeled its fault-tolerant architecture for Diraq’s qubits, showing how 1,000 logical qubits can be created with just 150,000 physical qubits, roughly 10x fewer than Diraq’s previous estimates. CUDA-Q Logical gave an unprecedented way to rapidly assess potential implementations of Iceberg’s architecture, helping dramatically reduce the hardware needed for useful, fault-tolerant quantum computing. Fermilab used CUDA-Q Logical to accelerate fault-tolerant architecture development from five months to three weeks — a 7x speedup. Fermilab transformed complicated fault-tolerant system designs into a repeatable and verifiable computational workflow, accelerating fault-tolerant algorithm development from five months to three weeks, a 7x speedup. QUOPS is a benchmark for the future of quantum computing, measuring the readiness of fault-tolerant quantum hardware for practical applications. Developed by Sandia National Laboratories, QUOPS is now available in NVIDIA CUDA-Q. QUOPS is a new, independent cross-platform benchmark developed by Sandia National Laboratories that measures the progress quantum computing systems are making toward utility-scale applications. Sandia shared early results in a preprint posted ahead of IEEE Quantum Week, reporting initial QUOPS benchmarks for QPUs from Google, IBM and Quantinuum. A QUOPS reference implementation is available in NVIDIA CUDA-Q. Diraq used NVIDIA Ising, the world’s first family of open models for building and deploying AI for useful quantum computing, to calibrate its silicon-based qubit processor. Companies are integrating their quantum processors with GPU supercomputing using NVIDIA NVQLink, the open system architecture for tightly coupling quantum processors with GPU supercomputers. Anyon Computing developed a new quantum control system. Quandela developed a QPU-GPU architecture. Quantum Machines ran a demo of integrating supercomputing resources with qubits at the Israeli Quantum Computing Center. The NVIDIA CUDA-Q open development platform continues to be adopted across the quantum computing ecosystem. BlueQubit launched the Quantum Flywheel grant program for researchers, to give them access to NVIDIA accelerated computing through CUDA-Q. Qedma Quantum Computing and QCentroid each integrated their technology with CUDA-Q to support quantum error correction and mitigation, as well as the deployment of quantum applications. IonQ announced progress on DQAOA-GPT, a quantum generative AI framework, using NVIDIA accelerated computing. MITRE published work on building GPU-accelerated digital twins of quantum sensors. Phasecraft is using NVIDIA cuQuantum to generate the largest-known variational quantum eigensolver emulated molecular database. UCLA and Caltech are advancing the design of control sequences needed to run quantum applications. CUDA-Q Logical is now available through GitHub. QUOPS repository is available. Papers on CUDA-Q Logical and QUOPS are available.

On September 14, 2026, Nuance Labs, Inc. closed the transaction. The company amended the terms of the transaction. The company has received $3,300,015 in it's second and final tranche led by returning investor, Lightspeed Management Company, LLC. The company has raised $50,000,000 in total funding in the transaction. The transaction included particpation from returning investors, Accel Management Co, Inc., Southpark Commons LLC and new investors, Define Ventures, NVIDIA Corporation.

NVIDIA Corporation Presents at Goldman Sachs Communacopia + Technology Conference 2026, Sep-10-2026 08:50 AM. Venue: The Palace Hotel, California, United States.

d-Matrix Inc. announced a new collaboration with NVIDIA Corporation that includes a multi-year product roadmap, providing d-Matrix Inc. XPUs entry into the widely deployed NVIDIA Corporation AI factory ecosystem. The centerpiece of the collaboration is an NVLink Fusion enabled rack-level system that AI labs, hyperscalers, and neoclouds can seamlessly deploy for ultra-low latency premium-level token services. As an NVIDIA Corporation NVLink Fusion partner, d-Matrix Inc. will work closely with NVIDIA Corporation to incorporate d-Matrix Inc.'s next-gen inference XPUs, starting with d-Matrix Inc. Raptor™, directly into the latest NVIDIA Corporation rack reference architecture design featuring NVIDIA Corporation Vera CPUs, NVIDIA Corporation NVLink switches, NVIDIA Corporation BlueField-4 DPUs, NVIDIA Corporation ConnectX-9 SuperNICs, and NVIDIA Corporation Spectrum-X Ethernet networking. d-Matrix Inc. is also partnering with Astera Labs, a connectivity solution leader within the NVIDIA Corporation NVLink Fusion ecosystem, to deliver custom solutions to ensure high-throughput, seamless data flow throughout the system. The d-Matrix Inc. rack, enabled by the NVIDIA Corporation MGX platform, will feature modular cable-free trays built with NVIDIA Corporation's mature, proven MGX ecosystem and supply chain for fast and seamless deployments. NVLink Fusion gives d-Matrix Inc. a mature, high-bandwidth, low-latency scale-up foundation for connecting d-Matrix Inc. XPUs to NVIDIA Corporation rack-scale infrastructure. With NVLink Fusion and the MGX ecosystem, d-Matrix Inc. can build around the same rack architecture, networking and supply-chain used across the NVIDIA Corporation platform. The first engagement point in the collaboration will be d-Matrix Inc. Raptor XPUs plugging into the NVIDIA Corporation MGX rack resulting in higher performance, greater deployment flexibility for customers, and a scalable architecture for expanding Raptor-based inference clusters as demand grows. d-Matrix Inc. has designed Raptor, which is expected to tape-out before the end of the year, from the ground up for integration with NVIDIA Corporation NVLink Fusion and the NVIDIA Corporation MGX rack-scale ecosystem, reflecting d-Matrix Inc.'s commitment to building purpose-built inference silicon that works seamlessly alongside the NVIDIA Corporation stack. Raptor is being actively evaluated at AI hyperscalers and frontier labs for its unique memory stacking solution and is backed by more than 100 patents. Initial availability of d-Matrix Inc. Raptor XPUs integrated into the NVIDIA Corporation MGX rack is expected Fourth Quarter 2027.

Rackspace Technology, Inc. announced that it has joined the NVIDIA Cloud Partner Program. The program deepens Rackspace’s ability to deliver and operate NVIDIA accelerated computing for regulated and sovereign enterprises. The relationship advances Rackspace’s AI Cloud strategy, giving organizations the infrastructure and operational model to maintain control of their data, models and intelligence while putting AI into production. Rackspace fills that role as the governed operator, bringing together NVIDIA accelerated computing with governed private and hybrid cloud, data, AI platforms and managed operations. Rackspace is putting this model into practice through its Institutional Sovereign Pod, a governed AI infrastructure architecture designed to give enterprises control over their data, models and AI environments while supporting the performance required for production workloads. The Pod will bring together the NVIDIA Blackwell architecture with Palantir Foundry and AIP in an environment Rackspace operates. Enterprises can use the pod to deploy AI at scale while maintaining sovereignty, security and operational control, including the Sovereign AI supply chain solution announced by NVIDIA Corporation and Palantir Technologies Inc. Rackspace Technology, Inc. provides forward deployed engineering and managed operations for the environment, helping customers move from deployment to production faster and reach value sooner. As an NVIDIA Cloud Partner, Rackspace Technology, Inc. will support NVIDIA technologies across all layers of the enterprise AI stack, with NCP competencies in Compute, Networking, Visualization and NVIDIA Enterprise Software. These capabilities strengthen Rackspace Technology, Inc.’s ability to integrate NVIDIA Corporation infrastructure and software with governed private and hybrid cloud and managed operations, providing customers with one accountable operator for the production environment.

Palantir Technologies Inc. and NVIDIA Corporation announced a collaboration to bring sovereign AI to critical supply chains, starting with NVIDIA Corporation’s own operations. The deployment creates an AI stack that brings NVIDIA Nemotron open models into Palantir Foundry and Artificial Intelligence Platform (AIP), grounded in the Palantir Ontology, which aim to create unprecedented supply-chain visibility, identify constraints, continuously codify operational expertise and guide decisions at machine speed — maintaining the reliability and efficiency of NVIDIA Corporation’s AI infrastructure supply chain while retaining control and ownership of proprietary data. The AI stack powering this work and its applications across industries will be showcased in depth at Palantir Technologies Inc.’s AIPCon 11. Organizations across complex industries — including agriculture, manufacturing, pharmaceutical, retail, technology and governments — can use the AI stack to optimize their own unique supply chain operations. The new AI stack gives NVIDIA Corporation supply chain teams a shared command center to reduce time of ownership, beginning with materials allocation decisions that affect how quickly parts move through the supply chain. The system helps teams identify potential constraints earlier, evaluate alternatives more quickly and allocate materials based on end-to-end production impact. Open models are central to this approach. By customizing Nemotron open models with proprietary operational data, prepared and augmented using NVIDIA NeMo Data Libraries, enterprises can build AI that reflects how their business operates while retaining control over their models, data and deployment environment. Each recommendation, planner action and production outcome strengthens the system. Integrated with NVIDIA NeMo AutoModel and NeMo RL, Palantir Autopilot operationalizes that feedback to support continued model improvement. This creates a governed learning loop that preserves operational knowledge, measures decisions against real-world results and improves the specialized AI models supporting the workflow. The companies plan to extend the learnings from NVIDIA Corporation’s deployment to help enterprises build more resilient, responsive and intelligent supply chains. Given the sensitivity and criticality of NVIDIA Corporation’s supply chain, the deployment runs on NVIDIA Corporation reference architectures and the jointly developed Palantir Technologies Inc. Sovereign AI Operating System Reference Architecture, which is supported by Dell Technologies and Cisco. Every supply chain is different. This AI stack can be deployed on premises with leading system manufacturers including Cisco and Dell Technologies, or in the co-lo and cloud with Rackspace and Nebius — enabling organizations to run AI where their data, systems and operational requirements demand it.

NVIDIA Corporation, $ 0.25, Cash Dividend, Sep-10-2026

NVIDIA Corporation announced that it is collaborating with a growing ecosystem of Australian NVIDIA Cloud Partners and AI infrastructure partners to expand land, power and shell capacity designed to host multiple generations of NVIDIA DSX AI factories, supporting the nation’s growing demand for AI compute — with up to a 2-gigawatt buildout by 2027. The participating providers will operate the AI factories, with NVIDIA Corporation delivering the DSX platform, accelerated computing, networking, software and ecosystem support. NVIDIA DSX is a full-stack AI factory platform spanning facilities infrastructure, computing, networking, software and reference designs. Compatible with the NVIDIA CUDA ecosystem and enhanced through software over the life of infrastructure, DSX helps operators serve a broad range of workloads and customers, improve performance and utilization, and extend the useful life of their assets — strengthening productivity, fungibility, durability, rentability and long-term economics. This expanded capacity will give Australian innovators greater access to accelerated computing and NVIDIA Nemotron open models to develop locally relevant models, applications and agents, while supporting regional and global demand for AI compute. Across Australia, NVIDIA Cloud Partners and AI infrastructure partners Firmus, Sharon AI, IREN, Megaport, ResetData, CDC, NEXTDC and AirTrunk are already investing in NVIDIA DSX AI factories to deliver high-performance AI compute to organizations nationwide. Firmus is expanding its Project Southgate initiative, developing NVIDIA-powered AI factories designed to support hyperscale and AI-native demand across Australia and Asia-Pacific. Sharon AI is deploying up to 68,000 NVIDIA GPUs connected with NVIDIA Quantum InfiniBand and NVIDIA Spectrum-X Ethernet networking, using the NVIDIA DSX platform to support startups, enterprises and research organizations. IREN is partnering with NVIDIA Corporation to accelerate deployment of large-scale AI factories by combining NVIDIA Corporation’s DSX AI factory architecture with IREN’s expertise across power, land, data centers, GPU deployment and infrastructure operations. Megaport is expanding access to NVIDIA Corporation accelerated computing through its wholly owned subsidiary, Latitude.sh, to support the inference needs of leading AI enterprises. Combined with Megaport’s global software-defined network, Latitude.sh compute platform brings AI processing closer to users, reducing latency and improving responsiveness. ResetData is building AI infrastructure in Australia, providing NVIDIA Corporation accelerated computing through its AI cloud platform for enterprises, government and research customers that require high-performance compute and Australian data residency. CDC has developed AI-ready data center infrastructure across Australia and New Zealand to support high-density AI and high-performance computing workloads. Its facilities feature advanced cooling technologies, including direct liquid-to-chip cooling, designed to support the deployment of NVIDIA Corporation accelerated computing infrastructure at scale. NEXTDC has developed AI-ready data center infrastructure designed to support high-density AI factories, providing the power, cooling, connectivity and physical infrastructure required to deploy increasingly compute-intensive AI systems. AirTrunk is expanding AI-ready powered shells across the region, including direct-to-chip liquid-cooling infrastructure designed to support high-density AI workloads.

On September 10, 2026, Ayar Labs, Inc. has amended the terms of the transaction and it has received $649,999,973 in the transaction. The company has now received $150,000,000 in the second and final tranche from new investor, Wiwynn Corporation. The round is raised at a post money valuation of $5,000,000,000.

Ultima Genomics announced a collaboration with NVIDIA to enable pangenome-aware whole genome sequencing analysis at unprecedented scale. By combining Ultima's high-output, cost-efficient sequencing platform with accelerated GPU-enhanced computational approaches from NVIDIA and Google's open source DeepVariant for pangenome graph alignment and variant calling, Ultima aims to make pangenome-aware whole genome sequencing variant calling practical for large-scale whole genome sequencing efforts, including population studies, transitioning from panel-based approaches to whole genome sequencing, and frontier biological research. The work includes joint benchmarking and validation of pangenome-aware approaches using genetically diverse human genomes, comparing workflows based on pangenome graph alignment with traditional linear-reference pipelines against high-confidence ground truth data. The teams are evaluating workflows based on community pangenome resources such as the Human Pangenome Reference Consortium and Google DeepVariant best practices for pangenome-aware variant calling, with the goal of identifying robust, scalable approaches suitable for real-world deployment. In parallel, Ultima is working with partners on infrastructure and system design optimized for cost-effective pangenome and whole genome sequencing analysis, enabling researchers to efficiently process massive sequencing datasets while retaining flexibility to adopt future methods and data models. Ultima's internal implementations achieve significant computational efficiency gains and acceleration by combining on-board CPU approaches with NVIDIA accelerated computing. Additionally, NVIDIA Parabricks provides GPU-accelerated versions of Giraffe and DeepVariant to meet the computational demands of population-scale sequencing projects. The joint work reflects a shared commitment to open, future-proof genomic ecosystems and lays the foundation for expansion beyond DNA sequencing into additional modalities over time.

HP Inc. announced a collaboration with Red Hat to give organizations more choice in where AI workloads run, whether locally, in the cloud or across both environments. In collaboration with Red Hat, HP is developing an open, enterprise-grade AI platform to deliver purpose-built AI infrastructure powered by Red Hat AI Factory with NVIDIA. HP’s open enterprise-grade AI platform aims to help companies maximize local AI inference throughput with up to 20 PFLOPS FP4 AI performance, reduce environment setup time and deployment risk, and improve GPU utilization through optimized NVIDIA CUDA libraries, scheduling, and multi-GPU workload orchestration. Red Hat AI Factory with NVIDIA is an integrated AI platform, built on the infrastructure of Red Hat Enterprise Linux and Red Hat OpenShift, for deploying and managing AI models, agents and applications across the hybrid cloud. The collaboration provides the ability to accelerate AI development by reducing setup time, enabling local agentic coding, and allowing companies to offload compute to the ZGX Fury without altering existing workflows. Running on Red Hat AI Factory, the solution is bringing together the co-engineered and jointly validated power of NVIDIA AI Enterprise and the scalability of Red Hat AI Enterprise. The initial solution will bring together HP ZGX Fury, based on the NVIDIA DGX Station platform, powered by NVIDIA GB300 Grace Blackwell Ultra Desktop Superchip, as the physical infrastructure for demanding local AI development and inference. Enterprise lifecycle management and support pathways are designed to improve consistency from developer environments to production deployment. HP ZGX Fury is now available to order. HP ZGX Fury is certified to run on Red Hat Enterprise Linux and available via the Red Hat Ecosystem Catalog. Customers will also be able to evaluate the planned solution in a sandboxed environment delivered on HP devices with Red Hat AI Factory with NVIDIA. Details on timing, locations, eligibility, supported configurations and access will be shared when available.

On September 9, 2026, Verily Life Sciences, LLC closed the transaction. The company has received its second and final tranche. The transaction also included participation from new investor, NVIDIA Corporation, and returning investor, CU Healthcare Innovation Fund II, L.P., a fund managed by CU Healthcare Innovation Fund, and University of Colorado School of Medicine, Endowment Arm.

NVIDIA Corporation Presents at 2026 Q2B Copenhagen Conference, Sep-09-2026 . Venue: Cophenhagen, Denmark.

Orange Logic Inc. announced it is embedding NVIDIA Nemotron open models within Orange Logic Agent Studio, expanding the AI capabilities customers can use to build and deploy agents across enterprise content operations. This includes powering the complex media workflows used by broadcasters, studios, streaming services, and other content intensive enterprises. Agent Studio gives AI agents access to the context surrounding enterprise content, including metadata, rights, permissions, business rules, approvals, and workflow state. With NVIDIA Nemotron open models, agents can apply advanced reasoning and multimodal AI to that context to understand content, make decisions, and act within established enterprise controls. Agents powered by Nemotron can work across images, video, audio, documents, and metadata. For media companies, this opens new opportunities across content discovery, media understanding, metadata enrichment, rights and compliance, content readiness, archive activation, and workflow automation. Unlike AI tools that operate outside the systems managing enterprise content, Agent Studio enables agents to operate within existing permissions and workflows. Customers can build specialized agents, connect them to approved tools and services, coordinate multiple agents across a process, and choose the model best suited to each task while maintaining consistent governance and orchestration. Orange Logic Inc. will showcase NVIDIA Nemotron within Agent Studio at IBC2026.

Beamr Imaging Ltd. announced that it has joined the VAST Cosmos Community, a global ecosystem of developers, builders, and innovators developing next-generation AI solutions, as a Technology Partner. The strategic collaboration with VAST Data, the AI Operating System company, will help deliver new AI-powered video workflows for broadcasters and content owners. The partnership will debut at IBC 2026, September 11-14 in Amsterdam, with a demonstration that brings together the VAST AI Operating System, NVIDIA Video Super Resolution and Beamr's content-adaptive video technology. The workflow enables broadcasters and content owners to upscale, AI-enhance and modernize their media archives, while keeping the resulting storage and delivery economics under control. Running on the VAST AI Operating System for Broadcast, the demonstration addresses these challenges where the video archives are already stored and managed. NVIDIA Video Super Resolution upscales lower-resolution content to 1080p or 4K, or improves quality at the same resolution. Beamr's content-adaptive bitrate (CABR) technology then delivers the result up to 50% smaller than standard encoding along a cost-friendly path that fits the existing delivery chain. The same pipeline can modernize media assets to HEVC or AV1 codecs - bringing older material up to formats that carry premium delivery standards, and opening new paths to monetize the media library. The entire pipeline runs end-to-end on NVIDIA RTX PRO GPUs and can be deployed in the cloud or on-premises. As a VAST Data Technology Partner, Beamr plans to collaborate with VAST Data on joint go-to-market initiatives around complementary video and data workflows. The IBC demonstration represents the first showcase of the collaboration for the broadcast and media market.

On September 8, 2026, Cognition AI, Inc closed the transaction. The company has received over $1,999.999861 million in a round of funding co-led by new investors Andreessen Horowitz LLC and Accel Management Co, Inc. The transaction also included participation from new investors Benchmark, Deer Management Co. LLC, Kleiner Perkins Caufield & Byers LLC, Greylock Partners, Lightspeed Management Company, LLC, Altimeter Capital Management, LP, Bond Capital Management LP, Meritech Sub-Management Corporation, Valor Management LLC, T. Rowe Price Group, Inc., DST Global, HOF Capital Management, LLC, BoxGroup Services, LLC, Battery Ventures L.P., Long Journey Ventures, A-Star Partners, LLC, Alkeon Capital Management, LLC, Diffusion, NVIDIA Corporation and returning investors The Founders Fund, LLC, General Catalyst Group Management, LLC, Avenir Management Company, LLC, Atreides Management, LP, Lux Capital Management, LLC, Eight Partners VC, LLC, D1 Capital Partners L.P., 137 Ventures Management LLC, Positive Sum, Soma Capital Management, LLC, Ribbit Management Company, LLC, Swish Ventures, Stripes, LLC, Definition Capital Management LLC, Hanabi Capital Management, Bain Capital Ventures,LP, Layer Capital Management, L.P., and Alpha Wave Global, LP. The transaction is being raised at a post-money valuation of $48,000 million. The company issued convertible preferred shares in this transaction.

Mistral AI SAS announced that it has raised €3 billion in a round of funding co-led by new investors Samsung Electronics Co., Ltd., Scaleup Europe Fund, a fund managed by EQT AB (publ), and existing investor PSG Equity L.L.C. on September 8, 2026. The transaction included participation from new investors Advent International, L.P., funds and accounts managed by BlackRock, Inc., The Government Of The Grand Duchy Of Luxembourg, existing investors Andreessen Horowitz LLC, ASML Holding N.V., Belfius Bank SA/NV, BNP Paribas Corporate & Investment Banking, Bpifrance Investissement SAS, Carmignac Gestion SA, DST Global, Eurazeo SE, General Catalyst Group Management, LLC, Headline, Hillspire LLC, Index Ventures SA, Korelya Capital SAS, Lightspeed Management Company, LLC, NVIDIA Corporation, Solar, a fund managed by Localglobe LLP, and Salesforce Ventures, LLC. The company has issued Convertible Preferred Stock in the transaction. The transaction has been raised at a post money valuation of above €21 billion.

Cognition AI, Inc. announced that it will issue 9,419,752 series E preferred stock at an issue price of $212.3198 per share for gross proceeds of $1,999,999,860.6896 on September 4, 2026. The shares are issued at $0.00001 par value per share. The preferred shares will be convertible to common shares at its original issue price.

NVIDIA Corporation (NasdaqGS:NVDA) entered into a definitive agreement to acquire an unknown majority stake in Hugging Face, Inc. for $12.9 billion on September 2, 2026. The transaction includes an approximately $11.9 billion purchase price payable to Hugging Face stockholders, subject to certain adjustments, and an equity-based retention program of up to approximately $1 billion for Hugging Face employees joining NVIDIA. As part of consideration, $12.93 billion is paid towards common equity of Hugging Face, Inc. NVIDIA expects the acquisition to provide additional resources to Hugging Face and support developers building, sharing and deploying open models and applications. NVIDIA has committed to, among other things, keep Hugging Face’s platform open, consistent with Hugging Face’s existing practices. The transaction is subject to the satisfaction or waiver of customary closing conditions, including receipt of required regulatory approvals. The transaction is expected to close in the first half of 2027.

Equinix, Inc. announced a significant expansion of its longtime collaboration with NVIDIA Corporation to deliver Equinix Inference Exchange, a distributed AI inference program for global enterprises, alongside a new collaboration with Together AI. This collaboration brings together NVIDIA Corporation's validated Enterprise Reference Architectures with Together AI's inference platform, supporting more than 200 open-source models. Delivered through Equinix, Inc.'s global data centers, it will provide connectivity to clouds, networks and AI providers through Equinix Fabric. The solution will be announced at Equinix Horizon, the company's inaugural customer and partner event, alongside Equinix Fabric One, which will make it easier for enterprises to connect across globally distributed AI environments. Equinix Inference Exchange will give enterprises a faster path from AI experimentation to production, with secure, low-latency connectivity to the data, users and ecosystem they depend on. The solution combines three complementary layers designed to simplify distributed AI inference: Equinix, Inc. provides the infrastructure foundation, including power, advanced cooling and day-two operations, connected through Equinix Fabric to the clouds, networks and AI providers that inference depends on. NVIDIA Corporation anchors the build with its Enterprise Reference Architectures and AI infrastructure purpose-built to maximize AI factory throughput and minimize token cost. Together AI runs the platform on top, supporting both multitenant deployments for shared efficiency and dedicated single-tenant environments for workloads that require dedicated capacity. Built on Equinix Fabric, the solution will connect to inference providers across major metros worldwide, cutting time-to-first-token. It also will connect to an expansive ecosystem of clouds, networks and AI providers, reducing deployment complexity. The solution aims to support a broad range of enterprise inference scenarios, including metro edge inference, open model migration, and sovereign AI. Equinix Inference Exchange will be available starting in the First Quarter 2027.

Winamp Group SA provided an update on the U.S. legal proceedings initiated by its subsidiary Jamendo SA against NVIDIA Corporation and Suno, Inc. in relation to the alleged unauthorized use of musical content and associated data in the development of artificial intelligence technologies. As previously announced, Jamendo initiated federal proceedings against NVIDIA and Suno in June 2026 as part of its actions to protect its intellectual property rights and those of the artists it represents. Since these initial filings, Jamendo has completed a number of additional legal and procedural steps, allowing both cases to progress on a more comprehensive basis. NVIDIA proceedings move forward: Jamendo has amended its complaint against NVIDIA to incorporate additional elements supporting its claims. The amended complaint has now been formally served on NVIDIA, marking an important milestone and allowing the case to move into its next phase before the U.S. federal court. The first hearing is currently scheduled for January 12, 2027. Suno: updated complaint to be refiled - Jamendo has also completed additional work on its claims against Suno. For procedural reasons, Jamendo has decided to withdraw the current complaint and refile an updated and more comprehensive complaint shortly. This decision is procedural in nature and is intended to ensure that all relevant elements are incorporated into the new complaint from the outset. A broader approach to AI-related uses - Beyond the proceedings against NVIDIA and Suno, Jamendo has identified, notably based on information made publicly available by the companies concerned, six additional companies active in the technology and artificial intelligence sectors that have used its content and/or data in the development of artificial intelligence technologies. Jamendo is currently reviewing these situations and has already contacted some of the companies concerned. Jamendo’s priority is to seek a commercial resolution, including appropriate compensation for past uses where applicable, as well as licensing arrangements or partnerships for future uses. Where these discussions do not lead to a satisfactory resolution, Jamendo intends to take the necessary steps to enforce its rights, as it has already done in the NVIDIA and Suno matters. This approach is part of a broader strategy to protect and create value from Jamendo’s catalogue and data, as music content becomes an increasingly important resource for the development of artificial intelligence. Recent developments in U.S. copyright litigation involving artificial intelligence have further highlighted the growing importance of these issues. While each case depends on its own specific circumstances, Jamendo believes that its claims are supported by a solid factual and legal basis.

On September 01, 2026, NVIDIA Corporation (NasdaqGS:NVDA) is reportedly in advanced negotiations to acquire AI startup Hugging Face, Inc. for approximately $12.9 billion, with the total deal potentially reaching $14 billion. This acquisition would mark NVIDIA CEO Jensen Huang's largest initiative to date aimed at expanding the adoption of AI applications and broadening its customer base.

iPronics Programmable Photonics S.L. announced that it has received $125 million in a round of funding co-led by new investors Light Street Capital Management, LLC, and Maverick Silicon GP, LLC on September 2, 2026. The transaction includes participation from new investors NVIDIA Corporation, IOWN AI Fund, a fund managed by Catalight Capital, European Innovation Council (EIC) Fund, Tate Family Trust, BC Mgmt, Inc., and returning investors Amadeus Capital Partners Limited, Criteria Venture Capital, SICC, S.A., a fund managed by Criteria Capital Risc, Triatomic Capital Private LP, Robert Bosch Venture Capital GmbH, Fine Structure Ventures. The company issued convertible preferred shares in this transaction. The company has raised $177 million in funding till date. As part of the transaction, Manish Muthal of Maverick Silicon and Geoffrey Tate of Light Street Capital Management, LLC joins the company's board of directors.

Marvell Technology, Inc. announced a quarterly dividend of $0.06 per share of common stock, including preferred stock on an as converted to common stock basis, payable on October 29, 2026 to stockholders of record as of October 9, 2026.

Qualcomm Incorporated announced the renewal of its global patent license agreement with Apple, effective April 1, 2027.

Intel Corporation Presents at Advanced Vision & AI Conference 2026, Sep-23-2026 . Venue: Santa Clara, California, United States.

AT&S Austria Technologie & Systemtechnik AG announced an expanded collaboration with Marvell Technology to increase advanced IC substrate capacity for next-generation AI and cloud infrastructure. Long-term agreement expands advanced IC substrate capacity to support growing demand for AI infrastructure. The agreement builds on the companies’ existing relationship and supports Marvell Technology’s growing demand for advanced substrates as AI deployments continue to scale. Marvell Technology is the additional customer identified in AT&S’s previously announced expansion of its Kulim manufacturing site. The agreement represents the next phase of the strategic collaboration between AT&S and Marvell Technology and underscores the growing importance of advanced IC substrate technologies as semiconductor designs become larger and more complex. As previously announced, the expansion of the Kulim site includes the fit-out of Plant 2 and the construction of a new manufacturing facility for IC substrate cores and advanced packaging. Supported by long-term customer commitments, the investment will provide additional capacity to meet growing demand for advanced IC substrates and packaging. The expansion in Kulim is supported by long-term customer agreements that provide visibility for future growth while supporting a disciplined investment strategy.

QUALCOMM Incorporated Presents at Humanoids Summit, Sep-22-2026 . Venue: COEX Convention Center, Seoul, South Korea.

MindWalk Holdings Corp. announced results from a production deployment of the open-source OpenFold3 structure prediction model, run with engineering partners Advanced Micro Devices Inc. and Vultr Holdings LLC on Vultr Kubernetes Engine and Vultr Cloud GPU powered by AMD Instinct MI325X GPUs and delivered through AMD Inference Microservices. It also served as an independent validation of AMD Inference Microservices for OpenFold3 in a production life-sciences setting. MindWalk Holdings Corp. designed the evaluation around its own production workloads, spanning protein, RNA, DNA, ligand, and antibody-antigen prediction, and ran the complete path from preprocessing through inference inside its discovery workflows, under its governance and human scientific review. A production-grade prediction environment deployed in minutes, and antibody-antigen inference time fell by approximately 5x against MindWalk Holdings Corp.'s prior environment. The deployment expands the compute capacity available beneath ReefIQ as MindWalk Holdings Corp. onboards enterprise partners. MindWalk Holdings Corp. ran single- and multi-chain proteins, protein-RNA and protein-DNA interactions, protein-ligand structures using CCD and SMILES inputs, and antibody-antigen complexes drawn from its own discovery programs, exercising the complete path including CPU-intensive preprocessing and multiple sequence alignment alongside GPU-accelerated inference. The evaluation also surfaced practical considerations around multi-GPU execution, inference controls, and deployment workflows. MindWalk Holdings Corp. shared those findings with Advanced Micro Devices Inc. and Vultr Holdings LLC to inform subsequent platform work. MindWalk Holdings Corp. deployed, ran, and measured a third-party open-source model on AMD Instinct GPUs in the Vultr cloud, inside its own discovery workflows and under its own governance and human scientific review. Partners that bring their own models onto ReefIQ follow the same path. Vultr Holdings LLC and Advanced Micro Devices Inc. form the compute fabric beneath MindWalk Holdings Corp.'s platform. Inference efficiency at this layer bears directly on the cost of delivering AI workflows to partners at enterprise scale. The standardized, Kubernetes-native path from open model to production workflow that Advanced Micro Devices Inc. and Vultr Holdings LLC have built shortens the distance between a research question and an answer inside a regulated environment.

Marvell Technology, Inc. announced industry-first 2nm optical interconnect demonstrations that advance the next generation of AI data center connectivity. The technologies will be showcased at the European Conference on Optical Communication (ECOC) 2026, September 20–24, 2026, in Málaga, Spain. Demonstrations include first 2nm 400G/lane optical PAM4, first 2nm 800G ZR/ZR+ pluggable with MACsec and first 2nm 1.6T ZR and O-band coherent-lite technologies. Marvell advanced optical and silicon design leverages leading-edge process nodes to deliver the improvements in performance, bandwidth density and power critical to scaling AI infrastructure efficiently. Marvell will showcase the breadth of its comprehensive AI infrastructure portfolio through 38 advanced technology demos. These include the Marvell 102.4T co-packaged optics (CPO) platform, alongside additional demos developed with its partner ecosystem. As AI data center architectures scale from 1.6T to 3.2T and beyond, operators require faster, lower-power and more secure optical interconnects to serve AI scale-up, scale-out and scale-across domains. Increasing bandwidth density while reducing power per bit has become critical to scaling AI infrastructure efficiently. Building upon a multi-generational history of industry firsts—including the first 5nm 200G/lane 1.6T Marvell Nova DSPs in 2023, the 3nm 1.6T Ara platform in 2024 and the 2nm 800G Libra DSPs announced earlier this year—Marvell is extending its portfolio of industry-leading 2nm connectivity solutions, including optical PAM4, coherent-lite PAM4/IMDD and ZR/ZR+ coherent solutions in the 2nm node. These technologies are optimized to enable ultra-high-speed interconnects that deliver faster communication between XPUs, CPUs, accelerators, and memory resources and improve energy efficiency to help data center operators manage the ever-increasing power demands of large AI deployments. At ECOC 2026 Marvell will highlight: First 2nm 400G/lane optical PAM4 demonstration: Highlighting the company’s multi-generational PAM4 leadership, the technology provides low-power, ultra-high-speed optical connections to help scale AI infrastructure, enabling the transition to 3.2T data center connectivity. First live demonstration of 2nm 800G ZR/ZR+ with MACsec: Powered by the Marvell Libra 2nm coherent DSP, the 800G ZR/ZR+ OSFP pluggable integrates media access control security (MACsec), expanding the company’s extensive coherent DSP and Marvell COLORZ pluggable portfolios, delivering efficient, high-performance and secure optical transmission for hyperscale AI and cloud data center networks worldwide. First 2nm 1.6T ZR and coherent-lite O-band demonstrations: The 1.6T ZR and 1.6T coherent-lite demonstrations are powered by next-generation Marvell DSPs announced earlier this year, optimized to deliver high-bandwidth connectivity within and between data center facilities and paving the way for 3.2T connectivity. Advanced CPO technology for next-generation AI fabrics: The Marvell CPO platform integrates optics directly with AI accelerators and switch silicon, as per-lane speeds and multi-rack AI pods push past copper’s reach limits. Demonstrated at 102.4 Tbps with 200G-per-lane silicon photonics, the platform delivers the bandwidth density and power efficiency needed to scale AI fabrics. Marvell executives will also participate in presentations and panels on optical connectivity for next-generation AI data center infrastructure. In addition to industry-first 2nm demonstrations, Marvell will also showcase solutions across its AI infrastructure portfolio at ECOC (Booth # 2002), from 224G SerDes and 1.6T optics to coherent interconnects and optical memory fabrics.

Broadcom Inc., $ 0.65, Cash Dividend, Sep-21-2026

Arista Networks, Inc.(NYSE:ANET) added to S&P 100

Marvell Technology, Inc. announced delivery of Azure Payment HSM v2, a new highly available payments security platform for financial institutions and payment service providers. The offering combines Marvell LiquidSecurity hardware security module (HSM) technology, Utimaco Atalla Payments Module and Microsoft Azure. It provides banks, payment processors and financial institutions with a hardware-backed, compliant and highly available foundation for securing critical payment workloads and transactions. Azure Payment HSM v2 delivers a cloud-native payment security platform through Microsoft Azure. The service enables financial institutions and payment providers to secure payment transactions, protect sensitive payment assets, maintain cryptographic key sovereignty and meet stringent PCI security, compliance, audit, performance and operational requirements without the cost and complexity of deploying or managing physical HSM infrastructure. Payment networks are under increasing pressure to modernize while meeting expanding regulatory and compliance obligations, including PCI PIN Security, PCI HSM and regional data sovereignty requirements. At the same time, transaction volumes and fraud threats continue to grow, exposing the limitations of traditional payment HSM deployments that rely on customer-owned data center-based infrastructure, which is costly, operationally complex and difficult to scale globally. This collaboration addresses that gap by combining: Utimaco Atalla Payments Module: Industry-proven cryptographic software supporting card issuance, PIN translation, mobile payments and key management, widely deployed across global payment networks, and the only payment software supporting multiple hardware platforms on the market. Marvell LiquidSecurity HSMs: Purpose-built, cloud-scale hardware security modules designed for dense, multi-tenant environments, offering high transaction throughput and hardware-based key protection. Microsoft Azure: A globally scaled, enterprise-grade cloud platform that delivers the integrated solution as a fully managed service, enabling customers to accelerate modernization, meet stringent compliance requirements, elastically scale with demand and reduce the operational burden of managing payment infrastructure. Azure Payment HSM v2 is available as a public preview beginning September 17, 2026, in Western U.S. and Western Europe.

GlobalFoundries Inc. and Marvell Technology, Inc. announced an expanded multi-year agreement to increase capacity for GlobalFoundries Inc.’s silicon germanium (SiGe) technology at its Burlington, Vermont, facility to support growing demand for next-generation optical connectivity. The expanded collaboration between GlobalFoundries Inc. and Marvell Technology, Inc. will increase capacity for high-performance SiGe technology used in next-generation pluggable optical transceivers, Near-Packaged Optics (NPO) and Co-packaged Optics (CPO). The expanded agreement builds on existing production and is expected to add significant capacity to support Marvell Technology, Inc.’s growing requirements. GlobalFoundries Inc.’s SiGe technology portfolio provides the high-frequency performance, signal integrity and power efficiency required for advanced optical networking applications. Its current technology supports 200G-per-lane optical connectivity, with a roadmap designed to address higher-speed generations as AI and cloud data center bandwidth requirements continue to grow. GlobalFoundries Inc. has manufactured advanced SiGe technology for over a decade and continues to co-integrate SiGe, silicon photonics and advanced packaging for new optical connectivity architectures including NPO and CPO. The agreement builds on more than a decade of collaboration between GlobalFoundries Inc. and Marvell Technology, Inc. and reflects the increasingly important role optical networking technologies play in enabling AI and cloud infrastructure at scale.

Advantech Co., Ltd. announced the expansion of its WEDA-Powered Edge AI ecosystem, centered on the theme 'Streamlining AI from Development to Deployment.' Through WEDA (WISE-Edge Developer Architecture), WEDA-Ready Edge Computing, and the Advantech Container Catalog (ACC), Advantech helps developers, system integrators, and customers reduce the complexity of Edge AI development, validation, deployment, and lifecycle management. WEDA is designed to turn Edge AI deployment from fragmented development efforts into a standardized and scalable lifecycle. By bringing together Advantech hardware, WEDA-Ready Linux, Advantech Container Catalog, WEDA Edge, and WEDA Cloud, Advantech helps developers build from Ready-to-Dev containers, scale from prototype to fleet through cloud APIs, and continuously evolve AI models through secure model management and data harvesting. Advantech is building an open Edge AI ecosystem that helps customers move faster from innovation to production. To accelerate edge AI, Intel Corporation aligns its broad product portfolio and Open Edge platform approach with Advantech Co., Ltd.'s WEDA-powered solutions, making it simpler to develop and optimize Edge AI workloads at scale across many different environments. Qualcomm Technologies, Inc. collaborates with Advantech Co., Ltd. on the WEDA ecosystem, enabling developers to accelerate the deployment of scalable Edge AI solutions. By combining Qualcomm Technologies, Inc.'s AI platform capabilities with Advantech Co., Ltd.'s WEDA-Ready infrastructure and Advantech Container Catalogue resources, AI adoption is simplified and intelligent applications are brought to market faster across industrial and enterprise environments. AMD collaborates with Advantech Co., Ltd. on WEDA-Ready Edge Computing and Advantech Container Catalog, helping developers take advantage of AMD Ryzen AI Embedded processor-based platforms with pre-validated, ready-to-develop container resources. Together, integration complexity is reduced and Edge AI workloads are accelerated from prototype to scalable deployment across industrial applications. Advantech Co., Ltd. will host a global WEDA Developer Webinar Series featuring WEDA use cases, technical resources, and deployment strategies.

Advantech Co., Ltd. announced the expansion of its WEDA-Powered Edge AI ecosystem, centered on the theme 'Streamlining AI from Development to Deployment.' Through WEDA (WISE-Edge Developer Architecture), WEDA-Ready Edge Computing, and the Advantech Container Catalog (ACC), Advantech helps developers, system integrators, and customers reduce the complexity of Edge AI development, validation, deployment, and lifecycle management. WEDA is designed to turn Edge AI deployment from fragmented development efforts into a standardized and scalable lifecycle. By bringing together Advantech hardware, WEDA-Ready Linux, Advantech Container Catalog, WEDA Edge, and WEDA Cloud, Advantech helps developers build from Ready-to-Dev containers, scale from prototype to fleet through cloud APIs, and continuously evolve AI models through secure model management and data harvesting. Advantech is building an open Edge AI ecosystem that helps customers move faster from innovation to production. To accelerate edge AI, Intel Corporation aligns its broad product portfolio and Open Edge platform approach with Advantech Co., Ltd.'s WEDA-powered solutions, making it simpler to develop and optimize Edge AI workloads at scale across many different environments. Qualcomm Technologies, Inc. collaborates with Advantech Co., Ltd. on the WEDA ecosystem, enabling developers to accelerate the deployment of scalable Edge AI solutions. By combining Qualcomm Technologies, Inc.'s AI platform capabilities with Advantech Co., Ltd.'s WEDA-Ready infrastructure and Advantech Container Catalogue resources, AI adoption is simplified and intelligent applications are brought to market faster across industrial and enterprise environments. AMD collaborates with Advantech Co., Ltd. on WEDA-Ready Edge Computing and Advantech Container Catalog, helping developers take advantage of AMD Ryzen AI Embedded processor-based platforms with pre-validated, ready-to-develop container resources. Together, integration complexity is reduced and Edge AI workloads are accelerated from prototype to scalable deployment across industrial applications. Advantech Co., Ltd. will host a global WEDA Developer Webinar Series featuring WEDA use cases, technical resources, and deployment strategies.

Marvell Technology, Inc. Presents at AI Infra Summit 2026, Sep-15-2026 through Sep-17-2026. Venue: Santa Clara Convention Centre, Santa Clara, California, United States. Presentation Date & Speakers: Sep-15-2026, Preetinder S. Virk, Senior VP & GM of the Photonic Fabric Business Unit, Rajagopal Krishnaswamy, Associate Vice President of Marketing, Cloud Data/AI Switching. Sep-16-2026, Dave Lazovsky, Executive Vice President and General Manager, Data Center Networking Business Group, Marvell, Rajagopal Krishnaswamy, Associate Vice President of Marketing, Cloud Data/AI Switching.

SiFive, Inc., and AMD showcased a demonstration of AMD ROCm running on SiFive’s recently announced BigSky Datacenter Development Platform. In a demonstration-only system, the companies showed the ability to run the modern Gemma4-E2B LLM model using ROCm 10.0 on SiFive servers powered by the SiFive Performance P870-D CPUs acting as the head node, and AMD Radeon AI PRO R9700 GPUs driving the inference offload. The SiFive BigSky SF-2U870 is designed explicitly for advanced software porting, workload tuning, and validation testing to foster ecosystem enablement. Hardware specifications include: 32 P870-D cores operating at 2.0 GHz, 256GB DDR5-5600 memory, 4x PCIe Gen5 x16 (64 lanes) and PCIe Gen3 x4, 2x 7.68TB U.2 NVMe SSD, 10/25Gb OCP 3.0 NIC. The platforms are available.

Advanced Micro Devices, Inc. Presents at AI Infra Summit 2026, Sep-15-2026 . Venue: Santa Clara Convention Centre, Santa Clara, California, United States.

Atsign announced that, with Intel Corporation, it has solved the massive encryption bottleneck that has until now held back the capabilities of autonomous Edge AI agents. In combination with Intel Corporation, Atsign has achieved a staggering performance uplift of approximately ~88x for hardware-accelerated, end-to-end encrypted Agent-to-Agent (A2A) communications. Describing the results of their collaboration, the two companies have co-published a new Smart Transportation Solution Brief, “Securing Agentic AI at the Edge for Smart Transportation”. The brief reiterates how transportation systems are evolving toward distributed, intelligent infrastructures where edge systems, operations platforms, and autonomous agents work together across intersections, traffic corridors, and mobility networks; and emphasizes how, as transportation becomes more autonomous, trust becomes essential - trust that every device, AI agent, data exchange, cryptographic key, and workload is secured to enable safe and reliable decision-making at scale. To deliver this trust, the new solution combines Atsign’s secure agent-to-agent architecture with Intel Corporation’s built-in hardware security capabilities: Atsign provides trusted identity management, namespace isolation, scoped data sharing, end-to-end encryption, and auditability. Intel Corporation provides hardware-backed key generation, accelerated encryption and hashing, protected key storage, secure boot, Trusted Compute, Total Memory Encryption, and Full Disk Encryption to help ensure platform integrity and data protection. The Smart Transportation Solution Brief details multiple new zero-trust Agentic AI performance benchmarks, showing how the performance uplift varies according to the hardware in question. For example: an ~88x performance uplift is available on Intel Xeon (up to 5 Gbps); and an ~60x performance uplift is available on Intel Core Ultra 9 (up to 6 Gbps). The new solution brief highlights how Intel Corporation’s hardware-rooted security capabilities, combined with Atsign’s secure agentic architecture, help protect data, workloads, and communications. The result for customers is the scalability and performance required for modern transportation infrastructure, alongside the security, trust, privacy required in the AI age. Intel Corporation has also added Atsign to its Industrial Builders partner program. For Intel Corporation customers, the presence of Atsign on the directory ensures de-risked deployments, through immediate access to Atsign’s pre-validated solutions. Customers know these have been engineered and tested for reliable compatibility with Intel Corporation architectures. Streamlined sourcing; the fast identification of Atsign’s workloads, presence and technical specializations eliminates the effort of navigating unverified vendor pools. Confidence that Atsign meets Intel Corporation’s compliance standards, bringing direct ecosystem backing and escalated Intel Corporation technical support to complex integrations. Direct collaboration, via a trusted gateway through which customers can connect with Atsign, whether for RFPs, reference architectures or proof-of-concept deployments.

Xanadu Quantum Technologies Limited announced the release of Backline, an open-platform solution, co-designed by Xanadu Quantum Technologies Limited and Advanced Micro Devices, Inc., for rapid communications between quantum and classical computing hardware. Backline is built using Xanadu Quantum Technologies Limited’s open-source PennyLane software platform to solve one of quantum technology’s most pressing challenges: how to establish ultra-low-latency communication between classical processors and quantum hardware. Backline reframes the quantum-classical boundary as a heterogeneous computing problem, treating a quantum processor as one more node in a high-performance distributed network. The solution seeks to provide researchers and hardware developers low-level hardware control directly through PennyLane's high-level Python interface. Backline is available as an open-source feature scane open-source platform.

From May 4, 2026 to August 2, 2026, the company has repurchased 0 shares, representing 0% for $0 million. With this, the company has completed the repurchase of 41,000,000 shares, representing 0.87% for $10,925.55 million under the buyback announced on April 7, 2025.

From May 4, 2026 to August 2, 2026, the company has repurchased 0 shares, representing 0% for $0 million. With this, the company has completed the repurchase of 0 shares, representing 0% for $0 million under the buyback announced on March 4, 2026.

Broadcom Inc. has announced a Fixed-Income Exchange Offer. Security Name: 4.926% Notes due May 15, 2037 Security Type: Corporate Bond/Note (Non Convertible) Principal Amount: $654.004 million Security Features: Callable; Senior; Subordinated; Unsecured Coupon Type: Fixed Transaction Features: Exchange Offer

Certain Performance Stock Units of Intel Corporation are subject to a Lock-Up Agreement Ending on 10-SEP-2026. These Performance Stock Units will be under lockup for 31 days starting from 10-AUG-2026 to 10-SEP-2026. Details: The directors and executive officers (such persons, the “lock-up parties”) have entered into lock-up agreements with the underwriters prior to the commencement of this offering pursuant to which each lock-up party, with limited exceptions, for a period of 30 days after the date of this prospectus supplement, will not dispose off any common stock or securities convertible into common stock.

Certain Common Stock of Intel Corporation are subject to a Lock-Up Agreement Ending on 10-SEP-2026. These Common Stock will be under lockup for 31 days starting from 10-AUG-2026 to 10-SEP-2026. Details: The directors and executive officers (such persons, the “lock-up parties”) have entered into lock-up agreements with the underwriters prior to the commencement of this offering pursuant to which each lock-up party, with limited exceptions, for a period of 30 days after the date of this prospectus supplement, will not dispose off any common stock or securities convertible into common stock.

Certain Options of Intel Corporation are subject to a Lock-Up Agreement Ending on 10-SEP-2026. These Options will be under lockup for 31 days starting from 10-AUG-2026 to 10-SEP-2026. Details: The directors and executive officers (such persons, the “lock-up parties”) have entered into lock-up agreements with the underwriters prior to the commencement of this offering pursuant to which each lock-up party, with limited exceptions, for a period of 30 days after the date of this prospectus supplement, will not dispose off any common stock or securities convertible into common stock.

Certain Restricted Stock Units of Intel Corporation are subject to a Lock-Up Agreement Ending on 10-SEP-2026. These Restricted Stock Units will be under lockup for 31 days starting from 10-AUG-2026 to 10-SEP-2026. Details: The directors and executive officers (such persons, the “lock-up parties”) have entered into lock-up agreements with the underwriters prior to the commencement of this offering pursuant to which each lock-up party, with limited exceptions, for a period of 30 days after the date of this prospectus supplement, will not dispose off any common stock or securities convertible into common stock.

Marvell Technology, Inc. Presents at Citi’s 2026 Global TMT Conference, Sep-09-2026 11:30 AM. Venue: New York Hilton Midtown, New York, New York, United States.

Broadcom Inc. Presents at Goldman Sachs Communacopia + Technology Conference 2026, Sep-08-2026 12:25 PM. Venue: The Palace Hotel, California, United States.

F-Secure Corporation and Advanced Micro Devices, Inc. Silo AI have demonstrated how F-Secure Trust technology and model routing solutions by Advanced Micro Devices, Inc. Silo AI can work together to help keep digital journeys safe, practical, private, and efficient in the agentic AI era. The uniquely developed routing approach decides, per piece of data, where it should be processed — on the device or in the cloud, and with what AI model — based on sensitivity, cost, model capability, and performance. Using this routing approach, upcoming solution F-Secure TrustPath helps keep the consumer's digital journey protected end-to-end. The two companies have developed an approach built on a single principle: trust is granted per moment, not per domain. Rather than granting blanket trust to an entire website, each step of a journey is judged on its own context. This requires multiple AI models to safeguard the agentic AI operating on the websites. F-Secure Corporation brings this trust model across the consumer journey; Advanced Micro Devices, Inc. Silo AI brings the compute that can dynamically span from on-device to cloud-deployed AI models based on the routing policy. This collaboration signals the latest in a series of AI-led announcements, showcasing F-Secure Corporation's AI-powered technology and reinforcing its position as consumer cyber security leader and pioneer of trust as a service in the agentic era. F-Secure TrustPath helps secure entire digital lifestyle journeys end-to-end, from start to completion. F-Secure TrustPath capabilities are expected to enter beta in Fourth Quarter 2026 and production in 2027.

QUALCOMM Incorporated Presents at Goldman Sachs Communacopia + Technology Conference 2026, Sep-08-2026 08:45 AM. Venue: The Palace Hotel, California, United States.

Qualcomm Technologies, Inc. announced a multi-generation collaboration with Amazon.com, Inc. to enable customized silicon at scale for large-scale AI data centers, working together on AI inference. In addition, the companies are working on optical connectivity solutions extending up to 1.6T and future-generation solutions. The collaboration brings together Amazon.com, Inc.'s comprehensive, secure, and price-performant AI infrastructure with Qualcomm Technologies, Inc.'s leadership in power-efficient processing, silicon design and system-level integration. Qualcomm Technologies, Inc. and Amazon.com, Inc. will also collaborate on high-performance optical connectivity solutions designed to support the growing scale and bandwidth demands of AI infrastructure. The collaboration will leverage Qualcomm Technologies, Inc.'s advanced SerDes and optical DSP technologies. As part of the expanded collaboration, Qualcomm Technologies, Inc. plans to deepen its use of Amazon.com, Inc.'s AI infrastructure, including Amazon Bedrock, for electronic design automation (EDA) workloads, targeting a reduction in chip design cycles.

Intel Foundry and ASML highlighted continued progress in bringing High Numerical Aperture (High NA) Extreme Ultraviolet (EUV) lithography into production and advancing the technologies that will shape its future adoption. Intel Foundry and ASML shared the following updates: High NA continues to be used in high-volume manufacturing production at Intel Foundry, with more than one million wafers processed to date across early tool certification and testing, research and development efforts, and volume production on select layers for a subset of Intel Core Ultra Series 3 processors, code-named Panther Lake. Overlay, throughput and availability are meeting Intel Foundry's expectations. Products manufactured on Intel 18A using High NA for select layers continue to deliver performance that meets or exceeds comparable layers patterned using the NXE platform (EUV with numerical aperture of 0.33). Customers can realize the benefits of High NA using the industry's current mask format, either by floor-planning within the 6-inch mask or by utilizing Intel Foundry’s stitching capabilities and process design kit (PDK) solutions. For more than three years, Intel Foundry has championed the large mask format initiative to align the mask ecosystem, working closely with ASML, mask makers, automation suppliers, EDA partners, materials providers, and semiconductor manufacturers to advance the standards, infrastructure, and technology needed to support future High NA scaling. Intel Foundry and ASML are building on years of collaboration to bring together the lithography ecosystem around larger-format masks, helping drive standards, infrastructure, and technology advances needed to support future High NA EUV scaling. Working alongside mask makers, automation suppliers, EDA partners, materials providers, and semiconductor manufacturers, the companies have helped establish the industry's path to larger-format masks while enabling customers to realize the benefits of High NA EUV.

Intel Corporation Presents at SPIE Photomask Technology + Extreme Ultraviolet Lithography Conference, Sep-08-2026 .

Arista Networks, Inc. Presents at AI Infra & Network Summit 2026, Sep-03-2026 . Venue: Ochanomizu Solacity Conference Center, Tokyo, Tokyo, Japan.

Broadcom Inc. announced that the Board of Directors of Broadcom has approved a quarterly cash dividend of $0.65 per share. The dividend is payable on September 30, 2026 to stockholders of record at the close of business (5:00 p.m. Eastern Time) on September 21, 2026.

Broadcom Inc. reported earnings results for the third quarter and nine months ended August 02, 2026. For the third quarter, the company reported sales was USD 29,591 million compared to USD 15,952 million a year ago. Net income was USD 13,088 million compared to USD 4,140 million a year ago. Basic earnings per share from continuing operations was USD 2.75 compared to USD 0.88 a year ago. Diluted earnings per share from continuing operations was USD 2.68 compared to USD 0.85 a year ago. For the nine months, sales was USD 71,089 million compared to USD 45,872 million a year ago. Net income was USD 29,747 million compared to USD 14,608 million a year ago. Basic earnings per share from continuing operations was USD 6.26 compared to USD 3.1 a year ago. Diluted earnings per share from continuing operations was USD 6.09 compared to USD 3.02 a year ago.

Broadcom Inc. provided earnings guidance for the fourth quarter ending November 1, 2026. For the fourth quarter, the company expects revenue to be approximately $34.8 billion.

Broadcom Inc. introduced VMware Private AI Cloud, a more secure, scalable, and flexible approach to AI that brings the model to the data, not the data to the model. Built on Broadcom's advanced software capabilities, VMware Private AI Cloud gives organizations a production-ready path to securely building, running, and governing inference workloads, agentic applications, and traditional enterprise workloads together on a single private cloud platform. Offering diverse hardware, model, and accelerator choices, VMware Private AI Cloud lets enterprises scale AI cost-effectively, operate more securely, and innovate rapidly. Scale AI Cost-Effectively: VMware Private AI Cloud addresses the three core AI cost drivers: hardware CapEx, operational complexity, and token economics (tokenomics). VMware Cloud Foundation (VCF) 9 lowers hardware costs via NVMe memory tiering and cluster-wide storage deduplication. VCF supports GPUs, CPUs, and accelerators from leading vendors, along with server hardware from major OEM and ODM vendors, allowing customers to run heterogeneous clusters cost-effectively. To optimize tokenomics and resource usage, it features token monitoring, multi-tenant model sharing, enhanced GPU/vGPU tracking, and an AI metrics observability dashboard. Infrastructure and operations innovations announced for Private AI Cloud include: Metal to model faster with VMware AI Factory: Broadcom announced VMware AI Factory, the software-defined foundation of VMware Private AI Cloud, providing customers a simplified path to production with new automation innovations for deploying AI-ready infrastructure and Day 2 operations. With VMware AI Factory, customers can achieve faster time to first model deployment and better manage AI tokenomics. Validated AI Models for VCF enable Model as a Service: VMware AI Factory gives enterprises a production-ready path to running leading AI models on-premises. VCF customers can run more than 150 open source and commercial models, including Nemotron 3, Gemma 4, cotomi, Qwen 3.7-Max, and GLM 5.2. Broadcom is working with the world's leading AI model providers to give organizations a clear path to data sovereignty and cost-effective AI at scale, with leading models available securely and delivered as a service to their user community through VCF's built-in services.Operate More Securely: Designed with a defense-in-depth approach aligned to NIST CSF 2.0, VCF protects against AI-accelerated threats by minimizing the attack surface and enabling continuous compliance. Automated, non-disruptive updates keep systems current, while VMware vDefend uses virtual patching and hypervisor-level lateral security with microsegmentation to enforce Zero Trust and block exploits. Furthermore, vDefend’s multi-layer threat defense and VMware Avi Load Balancer’s web application firewall and API protection prevent sophisticated attacks. Security innovations supporting Private AI Cloud include:TrueSource by Broadcom for verifiably built open source: Designed to address the acceleration in AI-enabled exploitation, Spring Enterprise delivers secure, curated Spring releases from the team that maintains it, including frontier model scanned, human-verified patches delivered simultaneously across every release line. TrueSource Trusted Artifacts extends clean-room builds across the Java ecosystem, Python, and Node.js, along with the Bitnami Secure Images catalog while TrueSource Data Services brings the same standard to the data tier: PostgreSQL, RabbitMQ, MySQL, and Valkey. Agentic Zero Trust with vDefend: New vDefend enhancements will extend Zero Trust lateral security for agentic AI workloads by identifying agentic AI components through continuous monitoring of traffic flows, detecting unauthorized usage of shadow AI, and providing distributed virtual patching through AI-generated Intrusion Detection and Prevention (IDPS) signatures. Agentic Threat Defense with Avi Load Balancer: Avi web security will broaden its protection of agentic AI workloads by restricting agents from accessing unauthorized tools and preventing misuse; flagging and isolating anomalous behavior to detect zero day attacks; and helping to prevent the unauthorized exfiltration of sensitive data with data protection guardrails. Innovate Rapidly for the Agentic AI Era: Unlike traditional apps, autonomous AI agents can act unchecked, exceed scope, or misinterpret instructions. Consequently, trust depends on robust controls and data integrity. VMware Tanzu Platform, with VMware vDefend, provides a foundation for trustworthy enterprise agents via a deny-by-default architecture, a prebuilt harness, and a curated marketplace. Agentic AI innovations announced for Private AI Cloud include: AgentMinder by Broadcom helps govern autonomous AI agents at scale: Broadcom today unveiled AgentMinder, a new solution that provides enterprises with a central control plane for autonomous AI agents. It treats agents as enterprise-grade identities, binding their authority to a specific mission, approved tools, and authorized resources. Additionally, it provides runtime policy enforcement—ensuring least-privileged access for every tool invocation—and delivers compliance-grade auditability, giving enterprises full visibility into what their agents are doing. AI-ready data foundations turn enterprise data into AI agent knowledge: New AI-ready data foundations in Tanzu Platform let data owners build and manage dynamic pipelines across structured and unstructured data, producing data products optimized for low-cost agent consumption. Those products publish to the Tanzu Platform marketplace as governed, context-rich services that both agents and developers can discover and use without data ever leaving the enterprise. Tanzu Platform agent foundations to build and run trusted AI agents: Tanzu Platform agent foundations enforce strict containment through a deny-by-default runtime; agents have zero access to APIs, networks, MCP servers, or the internet unless explicitly granted. New enhancements include an isolated credential store that shields all credentials from agents entirely. Agents can't leak or misuse what they can't see, closing off credential theft and prompt injection attacks.

Broadcom Inc. announced VMware AI Factory, the software-defined foundation of VMware Private AI Cloud. VMware AI Factory provides customers a simplified path to production AI with new automation innovations for deploying AI-ready infrastructure and supporting Day 2 operations. VMware AI Factory brings AI applications directly to enterprise private data within a secure private cloud environment. VCF’s unique infrastructure automation capabilities can reduce the time from bare metal server deployment to serving the first AI model from weeks to a matter of hours. VMware AI Factory streamlines AI infrastructure management by fully automating hardware provisioning, software stack enablement, and end-to-end lifecycle management. By integrating hardware and software operations into a unified, automated solution, organizations can rapidly scale AI workloads while minimizing operational complexity. As part of the VMware AI Factory, private AI services help make AI operational, governable, and cost-effective. VCF pools and shares GPU resources across the organization so teams can run multiple models on shared hardware instead of dedicating infrastructure to each workload. A unified model gallery gives IT and data science teams a single interface for deploying and managing model inference and RAG workflows across VMs, containers, and GPU resources, with built-in observability into token throughput, latency, and compute and memory utilization. Enterprises can pivot to new models while keeping costs low through shared infrastructure and governed models-as-a-service. New and forthcoming private AI services include multi-tenant model sharing, AI Gateway for unified model governance between on-premises and cloud environments, and secure AI sandboxes and governance. Broadcom is announcing a new partnership with MetalSoft to deliver integrated heterogeneous bare metal automation for VCF that drops bare-metal provisioning time from weeks to minutes. The integration will help IT provision or repave physical servers from multiple vendors directly through the VCF management console, unifying the software and hardware lifecycle into a single operational model and eliminating the need for vendor-specific tools for hardware and firmware management. VMware AI Factory combines VMware Cloud Foundation (VCF) with certified VCF AI ReadyNodes from Cisco, Dell Technologies, Lenovo, Supermicro and others and customers’ preferred AI software and accelerator architectures. Broadcom and AMD are collaborating to deliver a VMware AI Factory that pairs VCF with AMD Instinct GPUs and the open AMD ROCm software ecosystem. Zero-touch provisioning will orchestrate the end-to-end deployment of the entire stack, from vSphere and vSAN through Kubernetes and the AMD GPU operator, and the AMD DVX driver can attach GPUs to large VMs consumed by a VMware vSphere Kubernetes Service cluster. VMware AI Factory gives enterprises a production-ready path to running leading AI models on-premises. VCF customers can run more than 150 open source and commercial models, including Nemotron 3, Gemma 4, cotomi, Qwen 3.7-Max, and GLM 5.2. Broadcom is working with the world’s leading AI model providers to give organizations a clear path to data sovereignty and cost-effective AI at scale, with leading models available securely and delivered as a service to their user community through VCF’s built-in services.

Broadcom Inc. announced that leading AI models from providers including Google, NVIDIA, NEC, Alibaba Cloud, and Z.ai, are validated to run on VMware Cloud Foundation (VCF), enabling customers to bring these AI models on-premises and deliver “model as a service” to their users. VCF provides the AI- and Kubernetes-native platform that gives organizations across industries a production-ready path to deploying a wide variety of AI models. VCF is a unified private cloud platform capable of running inference workloads, agentic applications, containerized services, and traditional VMs together, eliminating the operational fragmentation of managing separate stacks. Independent benchmark testing under MLPerf Inference v5.1 standards confirms that VCF delivers performance on par with bare metal, making it the ideal platform of choice for enterprises deploying AI at scale on premises. VCF empowers enterprises to accelerate AI workload deployment at lower costs through an open and extensible ecosystem. Support for mixed compute across AMD, Intel, and NVIDIA frees enterprises to choose their preferred GPU and CPU hardware for their AI workloads. Leveraging vLLM as the default model runtime gives customers the ability to run more than 150 open source models on VCF. Broadcom is announcing the following models have been tested and validated to run on VCF: Nemotron 3: The NVIDIA Nemotron 3 family of open, multimodal models delivers leading accuracy and efficiency to help agents complete tasks faster. Combining hybrid Mamba-Transformer MoE architecture, 1 million context and multi-environment reinforcement learning, Nemotron 3 enables scalable, long-running agentic workflows across enterprise applications. Gemma 4: Google DeepMind's latest open source, open-weight multimodal model family, purpose-built for developers and the research community for bringing local execution, and enabling enterprises to build and deploy autonomous AI agents. cotomi: NEC's proprietary AI model optimized for Japanese language, trained on curated, highly reliable datasets. It empowers enterprises by seamlessly combining high-speed processing with a 40% improvement in token efficiency. Qwen 3.7-Max: Alibaba's Qwen 3.7-Max is a proprietary multimodal model that offers impressive one-million-token context windows, advanced multimodal reasoning, and agentic-era design, giving global enterprises sovereign, on-premises access to one of the world's most capable AI model families. GLM 5.2: Z.ai (formerly Zhipu AI)'s open source General Language Model enables enterprises to deploy coding and reasoning agents locally for multi-step autonomous workflows with data sovereignty and optimal hardware performance.

Broadcom Inc. announced TrueSource by Broadcom, a portfolio of commercially supported, verifiably built open source software for the enterprise. TrueSource brings together Spring Enterprise, the company’s flagship offering for the Spring ecosystem; new TrueSource Trusted Artifacts, which provides secure clean-room builds of the broader Java ecosystem, Python, and Node.js and incorporates a secure catalog of hardened container images; and TrueSource Data Services, a new offering that provides trusted artifacts, support, and deployment expertise for PostgreSQL, RabbitMQ, MySQL, and Valkey data engines. TrueSource Offerings are Built on Common Principles: Curated, prescriptive, enterprise-grade libraries and artifacts: Every library and artifact is selected against a reference architecture, built, and verified by Broadcom engineers, so enterprises consume open source with confidence. Remediation with maintainers, not around them: Broadcom contributes fixes upstream and backs community maintainers across the industry with engineering time and funding. Patch automation tooling and security visibility: Automation scans customer repositories, assesses the blast radius of each release before they consume it, and opens pull requests that apply the lowest-risk remediation path, with dashboards showing their security team exactly what’s fixed and what remains. Early access with collaboration: Properly licensed customers of any TrueSource offering will have the option to bring not-yet-public vulnerabilities they discover for early access to remediation. In addition, there is a special program for critical infrastructure organizations to get dedicated access to patch insights and mitigation advice. Broadcom Sets the Enterprise Standard with Spring Enterprise: Building on Broadcom’s June commitment to Spring supply chain security, this announcement arrives as AI accelerates exploitation, allowing attackers to weaponize vulnerabilities in hours. While this has fueled interest in fully automated, AI-generated patching, research indicates this approach carries significant operational and security risks. In new testing, 1Password’s Off-by-1 Labs found that only 26 percent of 6,000 AI-generated patches fixed vulnerabilities without breaking applications. They concluded that automated patches are not yet safe enough to trust without significant human oversight. Spring Enterprise provides secure, curated releases of Spring from the team that creates and maintains it. That stewardship comes with over 20 years of experience in making compatibility, performance, and security judgements that have allowed Spring to flourish. Customers receive: Proactive scanning with human-verified patches: Broadcom engineers continuously scan Spring and its dependency tree with frontier model analysis, then verify every patch by hand, finding vulnerabilities before attackers do. In the past five months, engineers have already spent more than 12 billion tokens against frontier models. Simultaneous patches across every release line: Because Broadcom maintains Spring, every supported release line is patched before a CVE is ever published. Disclosure and remediation for OSS and long term support versions arrive together, so no version is left waiting for a fix. The whole dependency tree, not just Spring: Coverage extends beyond Spring itself to its managed dependencies, including Apache Tomcat, Kotlin, and across the full dependency tree: more than 5,000 verified Java libraries, built and signed at the exact versions pinned by every supported Spring Boot release line. Security fixes without the upgrade: Full point releases bundle fixes with changes that demand testing. CVE-only patches carry the remediation alone, so security teams can push them to production in hours, not weeks. TrueSource Trusted Artifacts Extends Coverage Across Ecosystems: TrueSource Trusted Artifacts provides secure, clean room SLSA Build Level 3 builds of libraries across the Java ecosystem, Python, and Node.js. Broadcom’s curation process ensures that the libraries conform to a reference architecture and are supportable by the maintainers of record. Thousands of engineers across Broadcom’s software divisions scan, fix, contribute to, and consume them every day in the software that runs the world’s most essential businesses. The offering also includes the Bitnami Secure Images catalog, adding hardened, verifiably built container images for hundreds of commonly used open source packages to the same commercial offering.TrueSource Data Services brings it to the data tier: TrueSource Data Services extends the TrueSource promise to the data engines enterprise applications depend on: PostgreSQL, RabbitMQ, MySQL, and Valkey. A flawed patch can put the data itself at risk, so remediation takes operational judgment. Broadcom brings that judgment, from hardening and supporting these engines for the world’s most demanding enterprises, to curate a validated distribution inclusive of these data engines and the associated critical extensions, Operators and Helm Charts. The offering includes deployment automation for these engines as well as visibility into the security and operational posture. Broadcom has already invested behind this position, applying AI where it is effective. As announced in June, its Spring engineering team has scaled frontier model based scanning and validation across the dependency ecosystem, with every resulting fix authored, reviewed, and verified by engineers who know the code. That work answered the more than 1,700 percent surge in monthly security advisories reported by the Spring community and delivered the largest set of security patches in Spring’s 23-year history.Availability: Spring Enterprise, TrueSource Trusted Artifacts, and TrueSource Data Services are available with simple, tiered site licensing options.

Broadcom Inc. announced new AI-ready data foundations for the VMware Tanzu Platform, the official agent platform for VMware Private AI Cloud. The update delivers a complete, end-to-end framework enabling enterprises to transition safely from initial AI pilots to fully production-ready AI agents inside their own secure private clouds. As organizations adopt agentic AI, business leaders face critical trust and security hurdles. Unlike traditional software, AI agents act autonomously and query data independently, creating significant risks around data leakage, unexpected cloud egress fees, and inaccurate outputs caused by uncurated information. The latest VMware Tanzu Platform release solves this agent trust problem in two ways: securing the agent and securing the data. Agents run in hardened sandboxes leveraging services from a curated marketplace, so every agent operates with isolated credentials and explicit connections to the services it uses. Data is governed for access, context, and lineage. Access means agents reach the right data and only the right data. Context means data is prepared efficiently before agents consume it, improving accuracy and lowering token costs. Lineage means every agentic decision can be traced to exactly what data was used and where it came from. Key capabilities introduced in the VMware Tanzu Platform include: Hardened Agent Sandboxes: Enforces a "deny-by-default" security containment model that completely isolates credentials, helping to prevent prompt injection attacks and unauthorized network access.AI-Ready Data Foundations: Processes structured and unstructured enterprise data on-site, delivering high-precision context to AI agents to improve accuracy, reduce hallucinations, and lower token costs .Out-of-the-Box Developer Harness: Accelerates build times with pre-approved skills, step-by-step workflow buildpacks, human-in-the-loop controls, and integrated memory services. Curated Marketplace: Operates a centralized catalog where developers and agents can safely discover and connect to vetted AI models, tools, and data products. Auditable Agent Governance: Integrates an AI gateway to monitor, rate-limit, and log every action an agent takes for strict compliance and auditing. Availability: Capabilities announced here will be generally available in Tanzu Platform in Fall 2026.

Broadcom Inc. announced comprehensive security, identity, and observability capabilities for agentic AI environments running on Private AI Cloud. With AgentMinder, VMware vDefend, and VMware Avi Load Balancer, Broadcom provides advanced, multi-layer cybersecurity solutions designed to govern autonomous AI agents, defend agentic workloads, and protect enterprise Private AI Cloud workloads. Private AI Cloud delivers a more secure and cost-effective AI infrastructure to run inference workloads and agentic AI. It features the advanced security and automated compliance guardrails to help secure the workloads while helping to prevent data exfiltration and reducing cyber risk across every AI workload. Customers benefit from an agentic AI pipeline to build, run and govern trusted agents to innovate rapidly. The enterprise attack surface expands dramatically across agents, tools, datastores, and large language models (LLMs) as organizations deploy agentic AI applications using new protocols such as Model Context Protocol (MCP) and agent-to-agent (A2A) communications. Broadcom’s advanced cybersecurity solutions address these emerging blind spots by establishing trusted agent identities, enforcing Zero Trust boundaries, and providing enhanced visibility and granular control across the entire AI ecosystem. Comprehensive Defense and Governance for Agentic AI: These solutions address critical agentic AI security and operational risks across three core layers: VMware vDefend, VMware Avi Load Balancer and Broadcom AgentMinder. VMware vDefend: Agentic Zero Trust: VMware vDefend delivers Zero Trust lateral security for AI workloads running on VMware Cloud Foundation. The following new vDefend enhancements will extend its ZeroTrust lateral security to agentic AI workloads: Discovery of Agentic AI Components: Will automatically identify MCP servers, LLMs, datastores, and tools by continuously monitoring traffic flows across VMware Cloud Foundation. Shadow AI Monitoring: Will detect unauthorized AI usage and enforce strict Zero Trust policies to help maintain complete control over enterprise environments.AI-Generated IDPS Signatures: Will use an agentic AI pipeline to create Intrusion Detection and Prevention (IDPS) signatures at machine scale and speed, providing distributed virtual patching to protect workloads against the volume and velocity of AI-discovered vulnerabilities. VMware Avi Load Balancer: Agentic Threat Defense: Avi Load Balancer, with its seamless integration with Kubernetes including VMware vSphere Kubernetes Service (VKS), multi-terabit performance, elastic scale-out operation is ideally suited to deliver AI-aware load balancing, web application security and API protection (WAAP) to agentic AI workloads. The following new enhancements will broaden its protection of agentic AI workloads: Tool and Agent Misuse Prevention: Will help restrict agents from accessing unauthorized MCP tools and inspect transaction content to block malicious execution, including remote code execution (RCE) and file injection. Zero Day Attack Detection: Will establish normal agentic traffic baselines across agents, LLMs, and tools to flag and isolate anomalous behavior in real time. Sensitive Data Protection: Will help prevent unauthorized exfiltration of credentials, personally identifiable information (PII), and sensitive financial data. Broadcom AgentMinder: Agent Identity, Intent & Governance: Unveiled today, AgentMinder serves as the central control plane for autonomous AI agents (read the press release). As agents transition from content generation to active business process execution, AgentMinder delivers the following core governance capabilities: Identity and Intent Controls: Treats autonomous agents as enterprise-grade identities and binds their authority to a declared mission, permitted intents, approved tools, and authorized resources. Runtime Policy Enforcement: Features a cloud-native gateway that evaluates real-time context (identity, tool, intent, resource) and enforces least-privileged policies on every tool invocation. Compliance-Grade Auditability: Built on OpenTelemetry to deliver compliance-grade visibility into every agent session and action, delivering chain of custody, anomaly detection, and operational insight at machine speed.

Broadcom Inc. introduced AgentMinder, a new solution that acts like a traffic controller for autonomous artificial intelligence (AI) agents. AgentMinder independently verifies agent identity and authorizes each action against its declared mission, intent, context, and current risk before the action reaches an enterprise resource. AgentMinder addresses these requirements by providing a consistent framework for managing AI agents across models, tools, and deployment environments. By establishing policy-based controls, continuous authorization, and full telemetry, AgentMinder helps enterprises scale agentic AI while maintaining security, compliance, and operational trust. AgentMinder enables enterprises to turn autonomous AI agents into governed digital employees via the following capabilities: Identity and intent: AgentMinder treats agents as enterprise-grade identities and binds their authority to a declared mission, permitted intents, approved tools, and authorized resources. Every agent must declare what it is trying to do, not just who it is, before it can touch enterprise systems. Runtime enforcement: AgentMinder’s cloud-native AI gateway secures every tool call at runtime, authenticating tokens and directing traffic exclusively to authorized backends. Driven by a dynamic policy engine, the solution instantly evaluates context—from user identity to intent—to help enable total policy compliance. Observability and audit: With an observability layer built on OpenTelemetry, AgentMinder provides security, risk, and platform teams full, compliance-grade visibility into every agent session and action, delivering chain of custody, anomaly detection, and operational insight at machine speed. This combination of intent governance, runtime enforcement, and deep auditability enables enterprises to safely scale agentic AI into high-risk workflows such as Finance, HR, or IT without rewriting their existing identity or authorization stack. With a cloud-native architecture, AgentMinder is deployed alongside existing large language models (LLMs)—whether on-premises, in virtual private clouds (VPCs) or across public cloud environments. By integrating with existing authorization stacks via the AuthZEN standard, organizations can reuse current policy enforcement endpoints without routing traffic through a single software as a service (SaaS) chokepoint. AgentMinder benefits enterprises by helping secure agents with real identities and auditable controls, making AI costs visible and predictable, and deploying wherever their models run–all in a single, integrated solution. AgentMinder provides a highly flexible deployment model across platforms like VMware vSphere Kubernetes Service (VKS), Google Cloud Platform, and other standards-based, cloud-native Kubernetes platforms. AgentMinder’s multi-region, active-active architecture supports continuous uptime, effortlessly supporting peak loads of nearly 36 million customer-related and seven million workforce-related API calls every single day. AgentMinder is generally available.

AMD, Cisco and HUMAIN planned to deploy up to 250 MW of AI infrastructure powered by AMD Instinct MI400 Series GPUs, AMD EPYC CPUs and AMD ROCm open software, together with Cisco networking and critical infrastructure, through the companies’ previously announced joint venture. Deployment was planned to begin in 2027, with capacity expected to start coming online in the second half of the year. This next phase of the companies’ previously announced joint venture would significantly expand Saudi Arabia’s AI capacity. Driven by strong customer demand, the joint venture remained on track to deploy up to 1 GW of AI infrastructure by 2030. The collaboration was designed to give governments, enterprises, research institutions and developers access to advanced AI capabilities through a platform that brings together open models, open software and locally operated infrastructure. The platform would give customers greater control over where their data resides, how their models are customized, and how their AI systems are deployed and governed. Together, AMD, Cisco and HUMAIN planned to deliver the performance, scale and openness required by AI labs, model providers, enterprises and developers, bringing a new generation of sovereign AI capabilities to Saudi Arabia and the broader global market.