NVIDIA (NVDA) Company Report

NVIDIA and Dassault Systèmes announced their largest collaboration in over 25 years to integrate CUDA-X acceleration, NVIDIA AI, and Omniverse into Dassault Systèmes’ 3D design and simulation tools, enabling real-time virtual twins at 100×–1,000× scale and beyond.
The partnership spans industries—from life sciences (BIOVIA-powered virtual protein design at Bel Group) to automotive engineering (real-time crash and aero simulation at Lucid) and factory automation (software-defined AI factories at Omron)—demonstrating cross-sector virtual twin applications.
Introduction of AI “virtual companions” will let engineers seamlessly convert unstructured inputs (e.g., photos, sketches) into structured 3D models, automate compliance and manufacturability checks, and codify individual and organizational knowledge for design “shift-left” workflows.
The joint vision underpins the coming reindustrialization—an estimated $85 trillion global infrastructure buildout over the next decade—and aims to accelerate software-defined products and factories with continuous, AI-driven virtual twin operations.

1 day ago

Transcript

NVIDIA and Dassault Systèmes expand virtual twin and AI partnership

NVDA

NVIDIA and Dassault Systèmes announced their largest-ever collaboration, embedding CUDA-X acceleration, NVIDIA AI, and Omniverse into the 3DEXPERIENCE platform for real-time virtual twin capabilities.
The integration spans life sciences (AI-driven protein and drug/material discovery with BIOVIA, e.g., Bel Group’s non-dairy protein R&D) , engineering design (upstream crash and aerodynamic optimization for Lucid Motors) , and fully software-defined factory automation (e.g., OMRON’s autonomous production lines).
Jensen Huang characterized AI as foundational infrastructure akin to electricity and internet, anticipating $85–100 trillion in industrial AI and digital twin investments over the next decade.
AI-powered virtual companions will convert unstructured data (e.g., sketches, images) into structured 3D models, codify user expertise, automate design workflows, and keep proprietary knowledge on local systems.

1 day ago

Transcript

NVIDIA partners with Dassault Systèmes to deliver AI-powered virtual twin platform

NVDA

New Projects/Investments

NVIDIA and Dassault Systèmes announce their largest-ever collaboration, integrating NVIDIA CUDA-X, AI, and Omniverse into CATIA, SOLIDWORKS, SIMULIA and other Dassault tools to enable real-time, generative virtual twins across industries.
The partnership embeds physics-aware AI models (e.g., PhysicsNeMo) to accelerate simulations by orders of magnitude and shift compliance, manufacturability, and lifecycle checks upstream in the design process.
Joint focus on sector-specific use cases—from BIOVIA-powered food-science virtual twins with Bel Group to automotive generative optimization for Lucid and fully software-defined smart factories with OMRON.
NVIDIA highlights the scale of the coming AI-industrial infrastructure buildout, estimating $85–100 trillion in virtual twin-driven industrialization investment over the next decade.

1 day ago

Transcript

Cisco partners with NVIDIA to accelerate enterprise AI adoption

NVDA

New Projects/Investments

Cisco and NVIDIA detailed their AI factory vision, aiming to reinvent the enterprise computing stack across compute, storage, networking (via Cisco Nexus control plane), and security to support large-scale AI deployments.
Jensen Huang urged companies to “let a thousand flowers bloom,” encouraging broad AI experimentation before curating and concentrating on core workflows such as chip design, software engineering, and systems integration.
Emphasized an AI sensibility of abundance, applying orders-of-magnitude speed (real-time over annual cycles) and zero-mass computing to the most impactful business problems.
Outlined a five-layer AI stack—energy (chips), infrastructure (hardware & software), AI models, and applications—stressing that successful enterprise AI hinges on driving real-world applications, not just infrastructure.
Advised firms to build on-premises AI infrastructure to safeguard proprietary IP (notably their own questions) and integrate “AI in the loop” for continuous organizational learning and value capture.

3 days ago

Transcript

Cisco and NVIDIA outline AI infrastructure partnership

NVDA

New Projects/Investments

Cisco and NVIDIA announced a collaboration to develop AI factories, aiming to reinvent enterprise computing across processing, storage, networking, and security for AI workloads.
Jensen Huang urged enterprises to “let a thousand flowers bloom” by enabling broad AI experimentation in core business areas before curating and scaling the most impactful projects.
The CEOs highlighted the shift from explicit to implicit programming—treating computing as infinitely fast and leveraging AI’s “abundance” mindset to tackle large, context-dependent problems.
They detailed a five-layer AI stack—energy, chips, infrastructure, models, and applications—and stressed that success now hinges on rapid application development rather than infrastructure alone.
Recommended building on-prem AI systems to protect proprietary IP—keeping AI “in the loop” so organizational questions and insights become enduring company assets.

3 days ago

Transcript

Cisco and NVIDIA outline enterprise AI factories at AI Summit 2026

NVDA

New Projects/Investments

NVIDIA CEO Jensen Huang and Cisco’s leadership presented the AI factory partnership, aiming to reinvent the entire computing stack—processing, storage, networking and security—to accelerate enterprise AI deployment.
Huang advocated an enterprise strategy of “let a thousand flowers bloom,” encouraging broad AI experimentation before curating efforts around a company’s core value drivers like chip design and software engineering, with partners such as Synopsys, Cadence, Siemens and Dassault.
The talk underscored AI’s shift from explicit programming to implicit, self-supervised learning, scaling models from millions to trillions of parameters and enabling tasks that once took years to be done in real time.
A five-layer AI “cake” was described—energy, chips, infrastructure, AI models and applications—emphasizing that applications and embedding “AI in the loop” will define future corporate IP and competitiveness.

3 days ago

Transcript

NVIDIA and Dassault Systèmes announce industrial AI virtual-twin alliance

NVDA

New Projects/Investments

NVIDIA and Dassault Systèmes entered a long-term strategic partnership to integrate Dassault’s Virtual Twin and 3DEXPERIENCE platforms with NVIDIA’s Omniverse, open models and accelerated software libraries, aiming to create science-validated Industry World Models for real-time physics-grounded simulations.
The reciprocal deal will see Dassault deploy NVIDIA-powered “AI factories” via its OUTSCALE cloud while NVIDIA adopts Dassault’s systems-engineering tools for its Rubin platform, positioning NVIDIA’s infrastructure as the backbone for scaled industrial AI.
At its Feb. 2 SOLIDWORKS and 3DEXPERIENCE summit, Dassault unveiled three AI assistants built on these World Models, with the CEO projecting potential tenfold productivity gains for adopters.
The partners emphasized training on decades of validated scientific and engineering data to produce physics-grounded AI, targeting a market opportunity as high as $9 trillion.

3 days ago

Morningstar

NVIDIA joins collaboration on micro data centers for AI inference

NVDA

New Projects/Investments

EPRI is partnering with Prologis, NVIDIA, and InfraPartners to develop micro data centers (5–20 MW) for distributed inference, aiming for at least five U.S. pilot sites by end-2026.
The project will co-locate inference compute at or near utility substations to leverage underused grid capacity, reduce transmission congestion, and enhance grid reliability.
NVIDIA will provide its GPU-accelerated computing platform and architectural guidance, while Prologis sources sites and manages development and InfraPartners supplies high-density AI data center modules.

3 days ago

Press Release

NVIDIA unveils co-packaged silicon photonics switches for AI factories

NVDA

Product Launch

NVIDIA’s AI supercomputer architecture combines scale-up NVLink, scale-out Spectrum-X Ethernet, context memory storage via BlueField DPUs, and Scale Across for multi-data center connectivity.
Spectrum-X Ethernet delivers 3× higher expert dispatch performance for inference and 1.4× faster, fully synchronous training by eliminating jitter through RDMA, fine-grain adaptive routing, and SuperNIC injection control.
Co-packaged optics shifts the optical engine into the switch package, cutting scale-out network power by 5×, boosting signal integrity by 64×, and improving data-center resilience by encapsulating lasers in liquid-cooled modules.
NVIDIA products include a 102 Tb/s Spectrum-X switch (120× 800 GbE or 512× 200 GbE), a 409 Tb/s Spectrum-X variant, and a 115 Tb/s Quantum-X InfiniBand switch (144× 800 Gb/s), all fully liquid-cooled for maximum energy efficiency and million-GPU scalability.

4 days ago

Transcript

NVIDIA details co-packaged photonics switches for AI data centers

NVDA

Product Launch

NVIDIA outlined a four-tier AI supercomputer infrastructure comprising NVLink for rack-scale GPUs, Spectrum-X Ethernet for scale-out, ConnectX with BlueField for storage, and Spectrum-X scale-across for multi-site deployments.
Introduced co-packaged optics in Spectrum-X Ethernet and Quantum-X InfiniBand switches, embedding the optical engine with the switch ASIC to reduce network power by up to 5×, boost signal integrity 64×, and increase reliability 13×.
Unveiled Spectrum-X Photonics switches: a 102 Tbps model (120 × 800 Gbps or 512 × 200 Gbps) and a 409 Tbps variant (512 × 800 Gbps or 2000 × 200 Gbps); plus Quantum-X InfiniBand photonics: 115 Tbps with 144 × 800 Gbps ports.
Announced early 2026 deployments of Quantum-2 InfiniBand CPO with CoreWeave, Lambda and TACC, with Spectrum-X Ethernet CPO shipping later in the year.