Company hub
Nvidia
Every release, deal and policy move from Nvidia — updated daily.
-
NVIDIA Study: AI Agents Rarely Need Long Context Windows
NVIDIA researchers find Claude Opus 4.5 and GPT 5.6 Sol lose little accuracy with 16K contexts and frequent compaction, while long-context serving costs up to 7x more per token.
By
Model chronicle · Nvidia
Nemotron 3.5 Super & Lightning
research.nvidia.com (source, opens in a new tab)Nemotron 3 Ultra
550B-parameter (55B active) hybrid Mamba-Transformer MoE with open weights.
developer.nvidia.com (source, opens in a new tab)Nemotron 3 (Nano)
nvidianews.nvidia.com (source, opens in a new tab)Llama Nemotron (reasoning)
nvidianews.nvidia.com (source, opens in a new tab)Nemotron-4 340B
blogs.nvidia.com (source, opens in a new tab)
Sources: company announcements & release notes.
Fact box
- HQ
- Santa Clara, USA
- Founded
- 1993
- CEO
- Jensen Huang
- Core products
- Blackwell & Rubin GPUs · CUDA · DGX Cloud · Nemotron
- Stories
- 133
Sources: nvidia.com
Latest on X
NVIDIA · @nvidia
NVIDIA Alpamayo 2 Edge is here, bringing adaptive AI reasoning in real-time to the vehicle. 🚘 From yielding to pedestrians to making room for an oncoming bus, watch how it plans driving paths and explains its decisions when needed. Developers can build on this open model to
The robots were told there’d be candy. 🍬🎃 Happy Halloween!
Sit back and enjoy the show. You’re watching computer graphics and simulation powered by NVIDIA CUDA-X, the software helping scientists and engineers put their ideas to the test. From #NVIDIAGTC Berlin:
After a breast cancer diagnosis, one of the hardest questions is whether it will come back. @SimBioSys is helping doctors and patients better understand that risk. Powered by NVIDIA technology, its FDA-cleared TumorSight Risk analyzes an existing breast MRI alongside pathology
Europe is building its AI future, and we’re all in. From AI factories and AI agents to robotics and healthcare, Jensen took the stage at #NVIDIAGTC Berlin to show how AI is taking shape across the continent. Here's a look back at the highlights:
Editor's picks
Deals & funding
Nvidia · rounds, valuations, partners, acquisitions · latest 8 of 20
Acquisition
NVIDIA to acquire Hugging Face for ~$12.9B
Agreement to acquire Hugging Face for about $12.9B (incl. up to $1B retention equity); expected to close in H1 2027 subject to approvals.
blogs.nvidia.com (source, opens in a new tab)Partnership
NVIDIA guarantees up to $105B for OpenAI’s Ohio data-center campus
Residual-value guarantee for OpenAI's 20-year lease of SB Energy's PORTS-Pike campus in Pike County, Ohio: 4.25 GW initially plus a 3.75 GW option, online from 2028. NVIDIA pays only on default; OpenAI must reimburse.
investor.nvidia.com (source, opens in a new tab)Partnership
NVIDIA invests $2B in Marvell
cnbc.com (source, opens in a new tab)Partnership
NVIDIA invests $2B in Nebius
investor.nvidia.com (source, opens in a new tab)Partnership
NVIDIA invests $2B each in Lumentum and Coherent for optics
reuters.com (source, opens in a new tab)Partnership
NVIDIA commits $30B to OpenAI’s $110B funding round
NVIDIA and SoftBank each commit $30B and Amazon $50B to a $110B round at $730B pre-money; the round later closed at $122B ($852B post-money).
openai.com (source, opens in a new tab)Partnership
NVIDIA invests $2B in CoreWeave
reuters.com (source, opens in a new tab)Partnership
NVIDIA licenses Groq inference tech in ~$20B deal
Non-exclusive licence to Groq's inference technology plus key hires (incl. founder Jonathan Ross); reported ~$20B. Under DOJ scrutiny.
cnbc.com (source, opens in a new tab)
All 20 rounds & deals as a table · 0 funding, 15 partnerships, 5 acquisitions
| Date | Event | Type | Counterparty / round | Amount | Valuation | Source |
|---|---|---|---|---|---|---|
| Sep 3, 2026 | NVIDIA to acquire Hugging Face for ~$12.9B Agreement to acquire Hugging Face for about $12.9B (incl. up to $1B retention equity); expected to close in H1 2027 subject to approvals. | Acquisition | to acquire Hugging Face | $12.9B | — | blogs.nvidia.com |
| Aug 17, 2026 | NVIDIA guarantees up to $105B for OpenAI’s Ohio data-center campus Residual-value guarantee for OpenAI's 20-year lease of SB Energy's PORTS-Pike campus in Pike County, Ohio: 4.25 GW initially plus a 3.75 GW option, online from 2028. NVIDIA pays only on default; OpenAI must reimburse. | Partnership · compute | with SB Energy, OpenAI | $105B | — | investor.nvidia.com |
| Mar 2026 | NVIDIA invests $2B in Marvell $2B investment via Series A convertible preferred shares, part of NVIDIA's push into optics and networking (CNBC, Marvell filing). | Partnership · investment | with Marvell | $2B | — | cnbc.com |
| Mar 11, 2026 | NVIDIA invests $2B in Nebius Partnership to scale Nebius's full-stack AI cloud on NVIDIA platforms plus a $2B equity investment (NVIDIA, Reuters). | Partnership · investment | with Nebius | $2B | — | investor.nvidia.com |
| Mar 2, 2026 | NVIDIA invests $2B each in Lumentum and Coherent for optics Strategic partnerships to develop optical/photonics technology for next-generation data centers; announced by NVIDIA, Lumentum and Coherent. | Partnership · investment | with Lumentum, Coherent | $4B | — | reuters.com |
| Feb 27, 2026 | NVIDIA commits $30B to OpenAI’s $110B funding round NVIDIA and SoftBank each commit $30B and Amazon $50B to a $110B round at $730B pre-money; the round later closed at $122B ($852B post-money). | Partnership · investment | with OpenAI | $30B | — | openai.com |
| Jan 26, 2026 | NVIDIA invests $2B in CoreWeave Equity investment to accelerate CoreWeave's data-center build-out and expand the partnership (Reuters, CNBC). | Partnership · investment | with CoreWeave | $2B | — | reuters.com |
| Dec 24, 2025 | NVIDIA licenses Groq inference tech in ~$20B deal Non-exclusive licence to Groq's inference technology plus key hires (incl. founder Jonathan Ross); reported ~$20B. Under DOJ scrutiny. | Partnership · licensing | with Groq | $20B | — | cnbc.com |
| Dec 1, 2025 | NVIDIA takes $2B stake in Synopsys Multi-year partnership to accelerate chip-design and simulation software on NVIDIA GPUs. | Partnership · investment | with Synopsys | $2B | — | reuters.com |
| Nov 19, 2025 | xAI and Saudi Arabia’s HUMAIN to build 500 MW+ data center and deploy Grok nationwide Framework agreement at the US-Saudi Investment Forum: hyperscale GPU data centers in Saudi Arabia anchored by a 500 MW+ flagship site (NVIDIA GB300), xAI's first large deployment outside the US; Grok integrated into HUMAIN ONE. Value not disclosed. | Partnership · compute | with HUMAIN | — | — | x.ai |
| Nov 18, 2025 | Microsoft and NVIDIA to invest up to $15B in Anthropic Microsoft (up to $5B) and NVIDIA (up to $10B) to invest in Anthropic; Anthropic commits to $30B of Azure compute. | Partnership · compute | with Microsoft, Anthropic | $15B | — | anthropic.com |
| Oct 28, 2025 | NVIDIA invests $1B in Nokia for AI-native 6G networks NVIDIA buys ~2.9% of Nokia at $6.01/share; partnership on AI-RAN for 5G-Advanced and 6G. | Partnership · investment | with Nokia | $1B | — | nvidianews.nvidia.com |
| Sep 22, 2025 | NVIDIA and OpenAI sign 10 GW systems deal with up to $100B investment Letter of intent: at least 10 GW of NVIDIA systems; NVIDIA intends to invest up to $100B progressively as capacity is deployed. | Partnership · compute | with OpenAI | $100B | — | openai.com |
| Sep 18, 2025 | NVIDIA pays $900M+ to hire Enfabrica CEO and license its networking tech Acqui-hire: CEO Rochan Sankar and staff join NVIDIA, which licenses Enfabrica's GPU-networking technology (CNBC, Reuters). | Acquisition | acquires Enfabrica (acqui-hire) | $900M | — | cnbc.com |
| Sep 18, 2025 | NVIDIA to invest $5B in Intel, co-develop chips NVIDIA to invest $5B in Intel common stock; joint data-center and PC products. | Partnership · compute | with Intel | $5B | — | nvidianews.nvidia.com |
| Jun 11, 2025 | Mistral Compute: European AI cloud with NVIDIA (18,000 Grace Blackwell chips) Announced at VivaTech with NVIDIA; sovereign GPU cloud in Essonne, France, initially 18,000 NVIDIA Grace Blackwell superchips. | Partnership · compute | with NVIDIA | — | — | mistral.ai |
| May 13, 2025 | NVIDIA to supply HUMAIN with 18,000 GB300 chips, scaling to hundreds of thousands Saudi PIF company HUMAIN plans AI factories of up to 500 MW over five years; first phase 18,000 GB300 Grace Blackwell chips. Expanded on Nov 19, 2025 to up to 600,000 NVIDIA GPUs over three years. | Partnership · chips | with HUMAIN | — | — | nvidianews.nvidia.com |
| Dec 30, 2024 | NVIDIA completes Run:ai acquisition Completed after EU approval; Run:ai software to be open-sourced. | Acquisition | acquires Run:ai | $700M | — | techcrunch.com |
| Feb 8, 2022 | NVIDIA and SoftBank terminate $40B Arm deal Deal announced September 2020 abandoned over regulatory challenges in the US, UK, EU and China; NVIDIA keeps a 20-year Arm licence. | Acquisition | to acquire Arm | $40B | — | nvidianews.nvidia.com |
| Apr 27, 2020 | NVIDIA completes $7B Mellanox acquisition Israeli networking company (InfiniBand, Ethernet); the basis of NVIDIA's data-center networking business. Announced March 2019 at $6.9B. | Acquisition | acquires Mellanox Technologies | $7B | — | nvidianews.nvidia.com |
Sources: company announcements & release notes, press reports.
-
One in 17 AI Agent Skills Tells Agents to Skip Human Oversight, Study Finds
A study of 1.53 million public agent skills finds 5.8% tell agents to act without asking or skip…
-
Subtly Wrong Training Data Misaligns AI Models More Than Obvious Errors
A replication across 12 open-weight models finds that fine-tuning on plausible but wrong advice causes more broad misalignment…
-
VLMs Still Can’t Tell Which Attribute Goes With Which Object, Until Text-Only Finetuning
SABER binding benchmark: an 8B model finetuned on synthetic text alone jumps from 52.5% to 84.2% mean accuracy,…
More stories
Models & specs
Nvidia · context window, max output, API price, modalities, knowledge cutoff, license
| Model | Released | Context | Max output | Price in / out (per 1M) | Modalities | Knowledge cutoff | License / access | Spec source |
|---|---|---|---|---|---|---|---|---|
| Nemotron 3.5 Super & Lightning Base, post-trained BF16 and NVFP4 checkpoints | Oct 14, 2026 | 1M | — | open weights | Text, image, video in → text out | — | Open weights, OpenMDW 1.1 Super 120B / 12B active; Lightning 30B / 3B active | huggingface.co |
| Nemotron 3 Ultra | Jun 4, 2026 | 1M | — | open weights | Text in → text out | — | Open weights 550B total / 55B active (hybrid Mamba-Transformer MoE) | developer.nvidia.com |
| Nemotron 3 (Nano) | Dec 15, 2025 | — | — | — | — | — | — | nvidianews.nvidia.com |
| Llama Nemotron (reasoning) | Mar 18, 2025 | — | — | — | — | — | — | nvidianews.nvidia.com |
| Nemotron-4 340B | Jun 14, 2024 | 4K | — | open weights | Text in → text out | — | Open weights, NVIDIA Open Model License 340B dense (Base, Instruct, Reward) | blogs.nvidia.com |
Vendor benchmark claims
- Nemotron 3.5 Super & Lightning: SWE-bench Pro: 51.78% (Super)
Benchmarks as published by the vendor at launch; not independently verified by i10x.
Sources: company announcements & release notes. Prices = standard API list price in USD per 1M tokens at time of release or current docs; batch is typically 50% off.