Nvidia

NVIDIA contributed Vera Rubin rack and compute-tray designs to the Open Compute Project through its open NVIDIA MGX architecture. NVIDIA said the contribution gives system builders a common foundation for next-generation AI-factory infrastructure.

Samsung Electronics' preliminary third-quarter 2026 guidance put consolidated operating profit at about 107.4 trillion won, up 782.5 percent year on year, and revenue at about 195 trillion won, up 126.6 percent. TrendForce reported it as a record quarter in which operating profit crossed 100 trillion won for the first time and beat the LSEG SmartEstimate of 106.1 trillion won. It said HBM4 sales began making a meaningful contribution during the quarter, driven by demand for Nvidia's Vera Rubin accelerator, and cited Reuters for Samsung's HBM bit shipments jumping nearly 50 percent quarter on quarter.

STAR WARS: Galactic Racer became available with DLSS 4.5. NVIDIA published the availability announcement.

NVIDIA said Gears of War: E-Day was available worldwide. Tom’s Hardware reported that the game was among the first titles to ship with NVIDIA RTX Mega Geometry technology.

NVIDIA released GeForce Game Ready Driver 617.42 WHQL. NVIDIA said the driver could be downloaded through the NVIDIA app or GeForce.com.

Reuters, confirming an earlier Financial Times report, said SpaceX was in talks with banks and asset managers to raise $40 billion for an order of Nvidia AI chips, with PIMCO among the managers in talks. The Reuters copy puts the structure at about $10 billion in bank loans and $30 billion in investment-grade debt, with Apollo expected to lead the deal and place the paper. Bloomberg described the discussions as early and said they could end without a deal; none of the companies confirmed it.

A Groq spokesperson called the lawsuit challenging the company's transaction with NVIDIA meritless. Groq said it would vigorously defend itself against the case.

Bloomberg News reported that filings with Beijing regulators show Xinxin Finance Leasing, created in 2015 by China's national integrated-circuit investment fund and now majority-held by the Shenzhen, Beijing and other provincial governments, had over the past year advanced Hongjing Technology and its subsidiaries more than 3 billion yuan to buy more than 700 servers. One contract covered 32 Asus servers fitted with Nvidia B300 chips, bought from a Chinese company about which little is public; the B300 cannot be sold to China without a licence. Nvidia said it was investigating the report with its OEM customers, and Asus said it abides by all applicable export-control laws.

Nvidia said a GeForce RTX GPU with at least 12GB of VRAM is required to enable RTX Mega Geometry, in the article setting out DLSS 4.5 and ray-tracing support for Gears of War: E-Day. Tom's Hardware's launch testing found DLSS Multi Frame Generation present and working as expected in the game.

The Justice Department said Greg Lui, owner of the City of Industry technology company Earthmade Computer Inc., was arrested on a three-count federal indictment charging conspiracy to violate the Export Control Reform Act, outbound smuggling and conspiracy to commit money laundering. The indictment, returned on September 29, alleges that from 2023 to 2024 Lui bought export-controlled servers containing U.S.-manufactured GPUs, gave manufacturers false documentation about their end users, shipped them to Malaysia and Singapore, and had them re-exported to China; it says Earthmade received more than $176 million from two Malaysia-based shipping companies between January and October 2024. Tom's Hardware identified the GPUs as Nvidia parts.

Micron said it was working with NVIDIA on NV-HBM. Micron described NV-HBM as a custom HBM4E implementation for next-generation GPUs and NVLink Fusion platforms.

CoreWeave reported that Cognition measured a 4.8 times increase in total token throughput for SWE-2 inference workloads on Vera Rubin NVL72 compared with GB200 NVL72. NVIDIA reported that Cognition observed up to a 4.8 times increase in early tests.

CoreWeave announced the availability of NVIDIA Vera Rubin NVL72 on its cloud, with Cognition as the first customer anywhere running production workloads on the system.

AION 2's global version was scheduled for release on September 30. NVIDIA said it would include DLSS 4.5 Super Resolution and Multi Frame Generation on GeForce RTX GPUs.

ACE COMBAT 8: WINGS OF THEVE began Advance Access on September 29. NVIDIA said the game included DLSS 4.5 Super Resolution and DLAA at the start of Advance Access.

The Witcher 3: Wild Hunt — Remastered added Linear-Swept Spheres path-traced hair rendering for HairWorks characters on GeForce RTX 50 Series GPUs. Dynamic Multi Frame Generation supports up to 6X Frame Generation on GeForce RTX 50 Series GPUs.

Benzinga, citing Reuters and The Information, reported that Anthropic committed to pay SpaceX up to $84.5 billion for Nvidia-powered computing infrastructure through 2029. The report said the agreements generally allow either party to terminate with 90 days' notice.

Benzinga and Newsquawk reported that AMD and Nvidia were lobbying the Trump administration over proposed China-related chip-export restrictions in an annual defense policy bill. The reports attributed the account to Politico and sources.

Nscale announced $3.36 billion in pre-IPO convertible financing. The financing includes a $2.36 billion tranche at closing and a $1 billion NVIDIA commitment expected to fund in mid-November 2026.

SK hynix displayed NVIDIA's Vera Rubin superchip. The superchip was equipped with SK hynix's 12-layer 36GB HBM4 and 96GB SOCAMM2.

Supermicro said it was shipping NVIDIA Vera Rubin NVL72 racks. It said the racks were integrated with its Data Center Building Block Solutions and DLC-2 direct-liquid-cooling stack.

NVIDIA published a technical post describing DLSS 5 for game developers. NVIDIA said DLSS 5 was built for real-time 3D rendering and produces deterministic, temporally stable output.

NVIDIA announced GeForce Game Ready Driver 617.14 for AION 2, CONTROL Resonant, Gears of War: E-Day, and The Witcher 3: Wild Hunt - Remastered. NVIDIA said the driver included optimizations and enhancements for the four games.

A leaker with a record on Nvidia's plans said the GR20x family - the Rubin-based gaming GPUs expected to become the GeForce RTX 60 series - is now a 2028 product, moved from the second half of 2027. Nvidia has not announced the consumer Rubin lineup, its product names or a launch date.

VideoCardz reported a statement that would place NVIDIA’s Rubin-based GeForce generation into 2028. The item presented the timing as a report rather than an NVIDIA product announcement.

CoreWeave announced the bring-up of a multi-rack NVIDIA Vera Rubin NVL72 system on CoreWeave Cloud. The cluster connected hundreds of NVIDIA Rubin GPUs for agentic AI.

MLCommons released MLPerf Inference v6.1 results listing NVIDIA Rubin and NVIDIA Vera Rubin NVL72 as preview platforms. NVIDIA said its first MLPerf Inference preview submission for Vera Rubin NVL72 delivered up to 3.7 times higher throughput than GB300 NVL72.

Nscale entered a $3.1 billion minimum subscription agreement that included $1 billion of securities to be issued to NVIDIA. Nscale’s filing said the NVIDIA sale was expected to close on or around November 16, 2026.

NVIDIA said RuneScape: Dragonwilds would exit Early Access later that day. NVIDIA said the game would upgrade to DLSS 4.5.

NVIDIA reported that No Man’s Sky’s DLSS 4.5 update was available. The announcement also reported that Halloween: The Game, featuring DLSS Ray Reconstruction, was available.

NVIDIA reported that 007 First Light’s Path Tracing and DLSS Ray Reconstruction update had gone live. VideoCardz reported that IO Interactive released the game’s path-tracing update.

Tom’s Hardware reported that Micro Center still had GeForce RTX 5090 cards available around the average selling price. VideoCardz reported that NVIDIA’s flagship was becoming difficult to buy through normal online channels.

VideoCardz reported that HKEPC posted pictures reportedly taken at companies assembling AI servers with GeForce RTX 5090 graphics cards. The report described the cards as retail GeForce RTX 5090 products.

Reuters reported that Nvidia was in talks to become an anchor investor in Anthropic's prospective initial public offering. The discussions could involve an investment of up to $10 billion, according to the report.

An Nvidia spokesperson answered the reporting by calling the Groq deal an example of the American system working as designed.

The New York Times reported that Aivres, a California server maker, had exported at least US$5.6 billion of advanced technology to Southeast Asia between April 2024 and February 2026, including more than US$3 billion of computers carrying Nvidia's Blackwell chips, and that those servers ultimately served Chinese customers including ByteDance and Alibaba. Aivres is Inspur Group's US subsidiary and is not itself on the Entity List, though Inspur was added in 2023. Nvidia spokesman John Rizzo said the company does not support product diversion and sells only to well-known partners that work with Nvidia to comply with export control rules. Federal officials had begun looking into the subsidiary's business.

Bloomberg reported that DeepSeek intended to deploy at least 160,000 of Huawei's next-generation Ascend 950DT accelerators at a data centre it was building in Inner Mongolia, using them to operate its models, which would be one of the largest known clusters of Huawei AI chips. People familiar with the matter said the timeline depended on Huawei's production capacity and that DeepSeek did not currently plan to use the chips for training. DeepSeek did not respond to requests for comment and Huawei had no immediate comment.

Electronic Times reported that Micron began mass production of 12-layer HBM4 products in the second quarter of 2026. The report identified the products as 12-layer HBM4. The same report said Micron plans to add up to 60,000 wafers a month of HBM capacity by the end of 2026.

The driver that actually switches DLSS 5 on shipped on September 3, alongside Game Ready support for Onimusha: Way of the Sword and a frame limiter and V-Sync support for DLSS 4.5 Dynamic Multi Frame Generation.

Nvidia signed the definitive agreement on September 2 and announced it the next morning: about $11.9 billion to Hugging Face stockholders plus up to $1 billion in retention equity, closing expected in the first half of 2027 subject to US and EU regulatory approvals. Huang committed that Nvidia compute would not be required to build on or deploy through Hugging Face and that it would keep supporting other silicon vendors, the openness question regulators are expected to focus on. CEO Clément Delangue said he had approached Huang weeks earlier, after Hugging Face was hacked.

Where earlier DLSS reconstructed data the engine had already produced, higher resolutions, extra frames, cleaner ray-traced lighting, DLSS 5 generates lighting and material detail that real-time budgets had forced developers to strip out. The rendered frame, with its artist-designed geometry, textures and lighting buffers, is the fixed foundation rather than a prompt.

Nvidia's announcement article, bylined September 1, said DLSS 5 was available starting now in NBA 2K27, from developer Visual Concepts and publisher 2K, for every GeForce RTX 50 desktop and laptop GPU and on GeForce NOW. RTX 40 and RTX 30 cards were left out. The Game Ready driver that actually enables it did not ship until two days later.

Politico reported that NVIDIA and AMD urged the Trump administration to discourage lawmakers from including China-related chip-export restrictions in the annual defense policy bill. The reported outreach sought to prevent the proposed restrictions from being added to that bill.

Reuters reported that China's Ministry of Industry and Information Technology asked ByteDance, Alibaba and other companies to report plans to purchase NVIDIA RTX PRO 5500 chips. The report described the request as recent and did not establish a precise date.

NVIDIA invested $3.5 billion in convertible bonds issued by MediaTek. NVIDIA and MediaTek expanded their collaboration on AI infrastructure, local computing and automotive systems. MediaTek agreed to adopt NVIDIA NVLink Fusion as an option for customers developing custom XPUs for NVIDIA-connected rack-scale AI factories.

NVIDIA stated that Vera Rubin began production shipments in the third quarter of fiscal 2027. The filing described Vera Rubin as its next-generation Data Center architecture.

Days after Business Insider reported that Hugging Face was fielding takeover interest, The Information said Nvidia had agreed to pay $12.9 billion for the open-model platform, though talks had not produced a signed agreement. Unlike the Groq licence, an outright purchase would require a Hart-Scott-Rodino filing and likely an EU merger review.

In its second-quarter fiscal 2027 Form 10-Q, Nvidia said it took a US$0.4 billion charge in the first half of the year for excess H200 inventory and purchase obligations as demand diminished, that the Chinese government had restricted sales of H200 units it was licensed to ship, and that it had made only a fraction of the allowed shipments.

Nvidia said it was not assuming Data Center compute revenue from China in its outlook. The company issued the outlook with its second-quarter fiscal 2027 results.

Second-quarter fiscal 2027 results showed revenue of $96.2 billion, up 106% year on year, with data center revenue of $89.0 billion. Jensen Huang said Vera Rubin was in full production, and the release noted racks running at partners.

A second-generation Ray Reconstruction transformer model arrived through an Nvidia app update, with a denoiser Nvidia said had 35 percent more compute capability and processed 20 percent more parameters at similar performance.

Bloomberg reported that NVIDIA agreed to pay $6 billion for a Poolside technology licence and make a further $1 billion investment. Bloomberg reported that Poolside would continue operating independently.

The Financial Times reported, citing two people with knowledge of the matter, that ByteDance and Tencent had each taken delivery of roughly 10,000 H200 accelerators, the first meaningful movement of the chips into mainland China since the December 2025 approval. Beijing wanted most of each company's licensed allowance, which the FT put at up to 100,000 units apiece, kept off the mainland, with regulators directing buyers to Hong Kong instead. Reuters said it could not immediately verify the report.

NVIDIA announced it would invest $1.5 billion in SB Energy. The investment was announced with plans for NVIDIA to secure capacity at SB Energy’s PORTS-Pike campus, where OpenAI would be the tenant.

Nvidia and Safe Superintelligence Inc. announced a long-term partnership. Nvidia also made an investment in Safe Superintelligence Inc.

Nvidia reported that Data Center Hopper shipments to China were less than 1% of Data Center revenue in the second quarter of fiscal 2027. The quarter ended July 26, 2026.

NVIDIA said it planned to invest $1 billion in NAVER Corp., and NAVER said NVIDIA had decided to participate in a third-party private placement of new NAVER shares worth about $1 billion. The investment is subject to customary closing conditions and to NAVER finalising at least $9 billion of committed project financing, separate from NVIDIA's stake; Brookfield had signed a nonbinding term sheet to fund up to $9 billion. Neither source confirmed the share issuance had closed.

BIS Under Secretary Jeffrey Kessler told Congress that H200 chips had been shipped to China or Hong Kong. He said the number of chips shipped to date was very few.

NVIDIA said, “Our roadmap is intact.” Tom’s Hardware published the statement in an article about the Rubin Ultra Kyber NVL144 rack.

With memory makers prioritising AI customers, the RTX 5070 Ti and the 16GB 5060 Ti were squeezed out of supply rather than formally cancelled, and the Super refresh that depended on 3GB GDDR7 modules slid towards CES 2027. Australian retailers had already reported in January that they could not source 5070 Ti stock at all.

Fortune reported that Nvidia had quietly acquired Kumo AI, a four-year-old Sequoia-backed company whose foundation models make business predictions such as churn and credit risk from relational data, with its three co-founders moving to Nvidia in May; the price was later reported above $400 million.

NVIDIA said DLSS 4.5 Ray Reconstruction would be available in August. NVIDIA said the feature would be available for all GeForce RTX GPUs.

At GTC Taipei ahead of Computex, Nvidia said hundreds of supply chain partners across more than 350 factories in 30 countries were ramping Vera Rubin, with production shipments to begin in the fall.

NVIDIA reported $75.2 billion in Data Center revenue for the first quarter of fiscal 2027. It said the revenue was up 21% from the prior quarter and 92% from a year earlier.

IREN granted NVIDIA a five-year right to invest up to $2.1 billion in IREN shares. The partnership was intended to combine NVIDIA’s AI-factory architecture with IREN’s infrastructure operations.

Reuters reported that the United States had cleared roughly 10 Chinese companies to buy H200 chips. The report said deals had stalled and deliveries had not begun.

OpenAI said it had closed its financing round with $122 billion in committed capital at a post-money valuation of $852 billion, naming NVIDIA, Amazon and SoftBank as anchor investors with continued participation from Microsoft. CNBC reported the total was up from the $110 billion OpenAI had announced in February and called it the largest private funding round in Silicon Valley history.

NVIDIA disclosed that it had invested $2 billion in Marvell Technology. The announcement linked the investment to an NVLink Fusion partnership and planned silicon-photonics collaboration.

NVIDIA released an NVIDIA App update with the complete set of DLSS 4.5 features. The update included DLSS 4.5 Dynamic Multi Frame Generation and Multi Frame Generation 6X.

The U.S. Attorney's Office for the Southern District of New York announced the unsealing of an indictment charging Yih-Shyan Liaw, Ruei-Tsang Chang, and Ting-Wei Sun with conspiring to divert U.S.-assembled high-performance computer servers to China in violation of export-control laws. Reuters later reported that prosecutors charged Sun and two others associated with Super Micro with conspiring to divert about US$2.5 billion of U.S. AI technology to China.

Speaking to press at GTC 2026, Jensen Huang said Nvidia had been licensed for many customers in China for H200, had received purchase orders from many of them, and was restarting H200 manufacturing.

DLSS 5 was announced as a real-time neural rendering model that infuses pixels with photoreal lighting and materials, which Nvidia billed as its biggest graphics breakthrough since real-time ray tracing in 2018. Jensen Huang framed it as blending handcrafted rendering with generative AI. The company committed only to a autumn launch window.

At Nvidia's GTC 2026 Samsung made its sixth-generation HBM4, in mass production and designed for the Vera Rubin platform, the centerpiece of its showcase and unveiled HBM4E, positioning itself as a Nvidia supplier after missing the first HBM3E wave.

Micron said it had begun volume shipment of HBM4 36GB 12-high in the first quarter of calendar 2026, designed for Nvidia's Vera Rubin platform, with a 2048-bit interface above 11 Gbps and more than 2.8 TB/s per stack.

Nvidia announced that the seven chips of the Vera Rubin platform were in full production and described the platform as five racks and one supercomputer. This kept to the second-half-2026 timing set a year earlier.

NVIDIA announced a $2 billion strategic investment in AI-cloud provider Nebius. The companies paired the investment with a full-stack cloud engineering partnership.

Nvidia announced a multiyear partnership with the Thinking Machines Lab, founded by former OpenAI CTO Mira Murati, to deploy at least one gigawatt of Vera Rubin systems from early 2027, alongside an undisclosed equity investment reported at up to $2.5 billion in a round valuing the lab near $40 billion.

Lumentum completed the sale of 2,876,415 shares of a newly created Series A Convertible Preferred Stock to Nvidia for $2 billion in cash, at $695.31 a share. The preferred converts one-for-one into common stock, at the holder's option only once the Hart-Scott-Rodino waiting period has expired, and carries no vote on the election of directors. Lumentum supplies the optical components that carry traffic between AI servers.

NVIDIA completed a $2 billion private placement in Coherent common stock. Coherent issued NVIDIA 7,788,161 shares at $256.80 per share and said the investment would support research, capacity, and operations.

OpenAI announced a $110 billion financing round that included a $30 billion investment from NVIDIA. OpenAI put the round's pre-money valuation at $730 billion and identified Amazon and SoftBank as other major investors. It said NVIDIA would provide 3 gigawatts of dedicated inference capacity and 2 gigawatts of Vera Rubin training capacity.

NVIDIA reported $62.3 billion in Data Center revenue for the fourth quarter of fiscal 2026. It said the revenue was up 22% from the prior quarter and 75% from a year earlier.

NVIDIA unveiled the Rubin platform, which it said comprised six chips. It said AWS, Google Cloud, Microsoft Azure and Oracle Cloud Infrastructure would be among the first to deploy Vera Rubin-based instances.

Calcalist reported that Nvidia had bought Illumex, whose Generative Semantic Fabric turns enterprise data into ontologies for AI agents, for an estimated $60 million, a small deal aimed at the enterprise agentic-AI stack.

Anthropic said its $30 billion Series G round included part of Nvidia's previously announced investment. Anthropic reported a $380 billion post-money valuation for the round.

The U.S. government granted Nvidia a license to ship small amounts of H200 products to specified China-based customers. Nvidia said the license required U.S. inspection before shipment.

CoreWeave completed the sale of 22,935,780 Class A shares to Nvidia for $2 billion in cash. The purchase price was $87.20 per share.

NVIDIA said demand for GeForce RTX GPUs was strong. NVIDIA also said memory supply was constrained.

Reuters reported that Chinese customs authorities had told agents the H200 was not permitted to enter China, and that officials had summoned domestic technology companies on January 13 and instructed them not to buy the chip unless necessary. One person said the wording was severe enough to amount to a ban for now. The directive came days after the Trump administration formally approved H200 exports to China with conditions, and the authorities gave no reasons and no indication whether it was formal or temporary.

NVIDIA rolled out DLSS 4.5 Super Resolution to all NVIDIA App users. The rollout was announced on January 14, 2026.

Following Trump's December 8 decision to allow H200-class sales to approved Chinese customers, BIS revised its licensing policy so that applications for the H200, the AMD MI325X and similar chips are reviewed case by case rather than under a presumption of denial. Applicants must show exports will not shrink capacity available to US customers, that the buyer screens its own customers, and that the chips pass independent third-party testing in the United States.

BIS revised its license review policy so that applications to export Nvidia H200, AMD MI325X and similar chips to China are reviewed case by case if the exporter shows no reduction in capacity for U.S. customers, the Chinese buyer has export-compliance screening, and the product passed independent third-party testing in the United States. The rule took effect on publication.

TrendForce said HBM4 mass production was now expected no earlier than the end of the first quarter of 2026, because Nvidia had revised the HBM4 specification for Rubin in the third quarter of 2025 to a per-pin speed above 11 Gbps, forcing all three suppliers to modify designs, and because strong Blackwell demand led Nvidia to adjust the Rubin production timeline.

Nvidia was a strategic investor in xAI's $20 billion Series E financing. xAI did not disclose Nvidia's individual investment amount.

A year on, the announcements were all software and displays: a second-generation Super Resolution transformer model live in over 400 games and apps, and 6X Dynamic Multi Frame Generation due in the spring. No new desktop GPUs were revealed.

Reuters reported that ByteDance, Alibaba, and Tencent had been approved to purchase more than 400,000 H200 chips in total. The report described the approvals as conditional and cited anonymous sources.

On Christmas Eve Nvidia agreed to pay roughly $20 billion in cash for a perpetual, non-exclusive licence to Groq's inference-chip patents and software and to hire founder Jonathan Ross, president Sunny Madra and other staff, while Groq and GroqCloud nominally continued as an independent company. Structured as a licence rather than a merger, the deal avoided a Hart-Scott-Rodino filing; analysts said it kept 'the fiction of competition alive'.

Reports citing people familiar with the talks said Alibaba was considering buying 40,000 to 50,000 MI308 accelerators, priced around $12,000 each and roughly 15 percent below Nvidia's H20, in what would be AMD's largest China AI deal since the controls began. AMD had excluded MI308 from its guidance, so any order would be upside.

Nvidia bought SchedMD, developer of the open-source Slurm workload manager that runs most supercomputers and many AI clusters, for an undisclosed sum, promising to keep Slurm open and vendor-neutral. With Run:ai and Slurm it now owned the two dominant schedulers for GPU clusters, drawing questions about neutrality for mixed-vendor sites.

Oracle reported that remaining performance obligations increased by $68 billion in the second quarter to $523 billion. It said the increase included new commitments from Meta, NVIDIA, and others.

The Trump administration reversed course and said Nvidia could sell its H200 to approved commercial customers in China vetted by the Commerce Department, with 25 percent of the revenue going to the U.S. government. Nvidia welcomed the decision.

Nvidia bought 4,821,717 Synopsys common shares at $414.79 apiece, an aggregate $2 billion in cash, in a private placement under Section 4(a)(2). The two companies announced the placement in a joint press release the same day. Synopsys makes the EDA software chip designers use, including for Nvidia's own parts.

The Justice Department opened an inquiry into Nvidia's Groq licence-plus-hire arrangement shortly after it was announced in December 2025, and has since sent Nvidia a formal request for information. Nvidia took a non-exclusive licence to Groq's technology and hired several of its executives rather than buying the company, a structure the department is examining.

The Information reported that Meta was negotiating to deploy TPUs in its own data centers in 2027 and to rent them from Google Cloud in 2026, and that Google was now promoting TPUs for installation in customers' facilities rather than only in its cloud. Alphabet rose 4 percent and Nvidia fell as much as 7 percent on the report; a Google spokesperson said demand for both TPUs and Nvidia GPUs was accelerating.

NVIDIA reported $51.2 billion in Data Center revenue for the third quarter of fiscal 2026. It said the revenue was up 25% from the prior quarter and 66% from a year earlier.

NVIDIA reported that H20 sales were insignificant in the third quarter of fiscal 2026. The company disclosed the result in its CFO commentary.

NVIDIA reported Blackwell sales. It said cloud GPUs were sold out.

Nvidia and Microsoft announced investments of up to $10 billion and $5 billion in Anthropic, with Anthropic committing to buy $30 billion of Azure capacity and up to a gigawatt of Grace Blackwell and Vera Rubin systems, and to optimise Claude for Nvidia hardware. It gave Nvidia equity in both of the leading frontier labs.

Reuters reported that Beijing had issued guidance requiring any new data centre project that had taken state funds to use only domestically developed AI chips, and that projects less than 30 percent complete were told to remove foreign chips already installed or cancel planned purchases. More advanced projects would be assessed case by case. The rule reached Nvidia's H20, the part built specifically to satisfy the U.S. rules, and applied to a sector that had taken more than US$100 billion of state money.

NVIDIA announced a planned $1 billion equity investment in Nokia at a subscription price of $6.01 per share. NVIDIA and Nokia said they would develop NVIDIA-powered AI-RAN products for Nokia’s radio-access-network portfolio.

The largest investment Nvidia had ever proposed: a non-binding agreement to put up to $100 billion into OpenAI progressively as each gigawatt of Nvidia systems is deployed, starting with the first gigawatt on Vera Rubin in the second half of 2026. Critics called it circular financing; the definitive terms were still being negotiated months later.

In a cash-and-stock deal modelled on the acquihires Meta and Google had used, Nvidia paid over $900 million to bring in Enfabrica founder Rochan Sankar and staff and license its technology for connecting tens of thousands of GPUs, without buying the company, which Nvidia had backed in a 2023 funding round.

Nvidia agreed to buy $5 billion of Intel common stock at $23.28 a share and to co-develop custom x86 CPUs linked to Nvidia GPUs over NVLink for data centers and PCs, a month after the US government took its own 10 percent stake in Intel. The same day CNBC reported the Enfabrica acquihire.

Nvidia agreed to buy US$5 billion of Intel stock at US$23.28 a share. Intel will build Nvidia-custom x86 CPUs for Nvidia's data center platforms and x86 system-on-chips with Nvidia RTX GPU chiplets for PCs, connected with NVLink.

The Financial Times reported, and The Register relayed, that the Cyberspace Administration of China had ordered Alibaba, ByteDance and other companies not to order Nvidia's China-specific RTX Pro 6000D Blackwell GPU.

Five years after clearing the deal, SAMR announced that a preliminary investigation found Nvidia had breached the anti-monopoly law and the 2020 conditions, and said it would investigate further before deciding penalties. The finding landed during US-China trade talks in Madrid and a week before Beijing told companies to stop buying Nvidia's China chips.

The State Administration for Market Regulation said its preliminary investigation found Nvidia had violated China's Anti-Monopoly Law and the conditions attached to its 2020 approval of the Mellanox acquisition, and escalated the case to a formal investigation.

NVIDIA reported $41.1 billion in Data Center revenue for the second quarter of fiscal 2026. It said the revenue was up 5% from the prior quarter and 56% from a year earlier.

Second-quarter fiscal 2026 results showed no H20 sales to China-based customers, a US$180 million release of previously reserved H20 inventory tied to about US$650 million of sales to a customer outside China, and a third-quarter outlook that assumed no H20 shipments to China.

NVIDIA said production of Blackwell Ultra was ramping at full speed. It identified Blackwell as its AI platform.

NVIDIA added NVIDIA Smooth Motion support for GeForce RTX 40 Series GPUs. The release also added support for Global DLSS Overrides.

Six U.S. senators wrote to President Donald Trump asking him to reverse the arrangement allowing AMD and Nvidia to sell advanced AI chips to the People's Republic of China for a fee. The letter said the sales conflicted with U.S. national-security interests and questioned the fee's legal basis.

Chinese authorities sent letters to companies discouraging use of Nvidia's H20 for AI work, particularly by firms involved in government or national security-related projects, days after the U.S. cleared the chip for export again.

After the Financial Times reported the arrangement, a White House official confirmed to CBS News that Nvidia and AMD had agreed to pay the US government 15 percent of revenue from H20 and MI308 sales in China as a condition of the export licenses granted the previous week. No regulation was published, a point AMD would later flag to investors.

Reports, first from the Financial Times, said Nvidia and AMD had agreed to hand the U.S. government 15 percent of revenue from H20 and MI308 sales in China as the condition for export licenses. The Register described the arrangement as revealed that week; President Trump confirmed the terms publicly.

NVIDIA said its chips contain no backdoors, kill switches, or spyware. The statement addressed allegations concerning NVIDIA chips.

The U.S. government granted Nvidia licenses to ship certain H20 products to certain China-based customers. Nvidia later reported about $60 million in H20 revenue under those licenses.

The Cyberspace Administration of China said it had interviewed Nvidia and asked it to explain security risks from backdoor vulnerabilities in H20 chips sold to China and to submit supporting material. Nvidia said its chips have no backdoors that would allow remote access or control.

On the second-quarter 2025 call, CEO C.C. Wei said construction of the second Arizona fab, for 3nm, was complete and its volume production schedule was being sped up by several quarters. Work on the third fab, for 2nm and A16, had begun. Wei said about 30 percent of TSMC's 2nm-and-smaller capacity would eventually be in Arizona.

The day after Nvidia said it had been assured H20 licenses would be granted, AMD said the Commerce Department would move forward with its MI308 license applications and that it planned to resume shipments as approvals came, praising the administration's trade negotiations. The stock rose more than 6 percent.

After Jensen Huang met officials in Washington and travelled to Beijing, Nvidia said the U.S. government had assured it that H20 export licenses would be granted and that it hoped to start deliveries soon. It also announced an RTX Pro GPU aimed at the China market.

NVIDIA released a driver with support for GeForce RTX 5050 desktop and laptop GPUs. The driver release was dated July 1, 2025.

NVIDIA said GeForce RTX 5050 graphics cards would arrive on store shelves in the second half of July. The statement was published on June 24, 2025.

NVIDIA reported $39.1 billion in Data Center revenue for the first quarter of fiscal 2026. It said the revenue was up 10% from the prior quarter and 73% from a year earlier.

NVIDIA reported US$4.6 billion in H20 product sales in the first quarter of fiscal 2026 before the new export licensing requirements. The company disclosed the sales figure in its quarterly filing.

In first-quarter fiscal 2026 results Nvidia recorded a US$4.5 billion charge for H20 excess inventory and purchase obligations, said it could not ship a further US$2.5 billion of H20 revenue, and guided second-quarter revenue of US$45 billion reflecting about US$8 billion of lost H20 revenue.

NVIDIA said its Blackwell NVL72 AI supercomputer was in full-scale production. It said production covered system makers and cloud service providers.

NVIDIA announced that the GeForce RTX 5060 was available. NVIDIA said the card started at $299.

NVIDIA launched the GeForce RTX 5060 Ti as the first GeForce RTX 5060 desktop-family product. NVIDIA said the card started at $379.

NVIDIA announced the GeForce RTX 5060 family of GPUs. NVIDIA said the family started at $299.

A day after Nvidia disclosed the H20 license requirement and a $5.5 billion charge, AMD filed an 8-K saying the same new rule covered exports of its MI308 to China, Hong Kong, Macau and other D:5 countries. It expected up to $800 million of inventory, purchase-commitment and reserve charges, said it would apply for licenses with no assurance of approval, and its shares fell more than 7 percent.

NVIDIA announced that Blackwell chips had started production at TSMC chip plants in Phoenix, Arizona. The statement identified the production location as TSMC's Phoenix plants.

The Information reported that Nvidia had closed the purchase of Lepton AI, a two-year-old startup led by former Alibaba vice president Yangqing Jia that rented Nvidia GPUs across clouds. Lepton became the basis of DGX Cloud Lepton, Nvidia's compute marketplace launched at Computex in May 2025.

Nvidia disclosed that on April 9 the government required a license for exports of H20 and any chip with its memory or interconnect bandwidth to China, Hong Kong, Macau and D:5 countries, and on April 14 said the requirement would apply indefinitely. Nvidia expected up to about US$5.5 billion of first-quarter charges for H20 inventory, purchase commitments and reserves.

Wired reported that Nvidia had bought Gretel, a San Diego company with about 80 employees whose platform generates synthetic training data, for more than its last $320 million valuation, folding it into Nvidia's model-training tools.

In the GTC 2025 keynote, Nvidia dated the next two generations. Vera Rubin NVL144 systems were slated for the second half of 2026 and Rubin Ultra systems for the second half of 2027.

Nvidia announced Blackwell Ultra and the GB300 NVL72 rack, claiming 1.5x the AI performance of the GB200 NVL72. Partner products were expected from the second half of 2025.

NVIDIA said DLSS 4 with Multi Frame Generation was available in over 100 games and applications. The company published the statement on March 13, 2025.

Reuters, via TechCrunch, reported that Meta was testing an in-house chip for training AI systems as part of its strategy to reduce reliance on Nvidia and other hardware makers. It was the first sign that MTIA would move from recommendation inference into training, which Meta confirmed a year later with the MTIA 300 to 500 roadmap.

NVIDIA released a driver that added support for the GeForce RTX 5070. The driver release was dated March 5, 2025.

NVIDIA's GeForce RTX 5070 was scheduled to launch on March 5, 2025. The Founders Edition was not scheduled to be available for purchase until later in March.

Fourth-quarter fiscal 2025 results reported data center revenue of $35.6 billion. Nvidia said Blackwell AI supercomputers achieved billions of dollars in sales in their first quarter, without giving an exact figure in the release.

NVIDIA announced that the GeForce RTX 5070 Ti was available. The announcement was published on February 20, 2025.

The first two Blackwell cards shipped three weeks after the keynote. DLSS 4 with Multi Frame Generation was live in more than 75 titles on day one, including Alan Wake 2, Cyberpunk 2077 and Hogwarts Legacy.

Nvidia brought an action seeking annulment of the European Commission's October 31, 2024 decision on Italy's referral request. The action concerned the Run:ai transaction.

Multi Frame Generation was made exclusive to the RTX 50 series, with 75 games and apps supporting it at launch. The bigger change was architectural: Ray Reconstruction, Super Resolution and DLAA moved to transformer models, which Nvidia called the graphics industry's first real-time application of the architecture behind frontier language models.

The Blackwell consumer line was announced with the RTX 5090 as the halo part at $1,999, the 5080 at $999, the 5070 Ti at $749 and the 5070 at $549. Every announced part shipped with ordinary 2GB GDDR7 modules, a detail that would matter a year later.

Nvidia completed the Run:ai purchase ten days after EU clearance, and Run:ai's founders said the platform, which then supported only Nvidia GPUs, would be open-sourced so it could extend to the whole AI ecosystem, blunting concerns that Nvidia would use it to lock in its hardware.

Although the deal fell below EU turnover thresholds, Italy's competition authority referred it to Brussels under Article 22 and the Commission opened a review in November. It found Run:ai had no significant position in GPU orchestration software and that customers had credible alternatives, and approved the deal without conditions; the US Justice Department had also examined it.

China's State Administration for Market Regulation opened an investigation into Nvidia. The investigation concerned suspected Anti-Monopoly Law violations and conditions attached to approval of Nvidia's Mellanox acquisition.

In its third-quarter fiscal 2025 results, Nvidia described Blackwell as being in full production and reported record data center revenue of $30.8 billion, up 17% from the prior quarter.

OpenAI raised $6.6 billion in new funding at a $157 billion post-money valuation. Axios reported that Nvidia joined the round led by Thrive Capital.

Italy's competition authority submitted an Article 22 referral request concerning Nvidia's proposed acquisition of Run:ai. The request was sent to the European Commission.

OctoAI, spun out of the University of Washington's Apache TVM project and once valued at $850 million, told customers it was winding down its services by October 31 as it was absorbed into Nvidia in a deal reported at roughly $250 million. Earlier in 2024 Nvidia had also picked up Shoreline.io and Brev.dev.

NVIDIA said Blackwell samples were shipping to partners and customers. The company identified the products as Blackwell samples.

On the earnings call, CFO Colette Kress said Nvidia had changed the Blackwell GPU mask to improve production yields and that the production ramp would begin in the fourth quarter. Jensen Huang said Nvidia meant to ship product, not merely start production, in Q4. Nvidia also said Blackwell demand was well above supply.

Italy's competition authority considered Nvidia's proposed Run:ai acquisition to meet applicable conditions of the Italian Competition Act. The authority exercised its call-in power over the transaction.

Citing The Information, The Register reported that Nvidia had told Microsoft of delays to the most advanced Blackwell parts, with volume shipments pushed to the first quarter of 2025 and TSMC's CoWoS packaging suspected as the cause. Nvidia's statement did not confirm a delay: it said sampling had started and production was on track to ramp in the second half.

Nvidia acquired AI-development-platform startup Brev.dev. Nvidia confirmed the acquisition to CRN, and Brev.dev cofounder Nader Khalil announced it separately.

Bloomberg reported that Nvidia had agreed to buy Shoreline.io. The report cited people familiar with the matter.

In his Computex keynote in Taipei, Jensen Huang showed the Rubin platform as Blackwell's successor, together with an Arm-based CPU called Vera and new networking parts. He described a one-year release rhythm for the whole data center stack. No launch year was given in Nvidia's own write-up.

Reuters reported that Nvidia had cut H20 prices in China, where oversupply left the chip selling at more than a 10 percent discount to Huawei's competing part. The first export-compliant chip Nvidia built for China after the October 2023 rules was therefore competing on price against the domestic alternative it was meant to hold off.

NVIDIA said the Blackwell platform was in full production. It said the platform formed the foundation for trillion-parameter-scale generative AI.

A week after Run:ai, Nvidia agreed to acquire Deci, a Tel Aviv company whose neural-architecture-search tools shrink and speed up deep learning models, for an estimated $300 million. The pair took Nvidia's spending on Israeli startups past $1 billion in a week.

Nvidia agreed to buy the Israeli Kubernetes-based scheduling company, reportedly for about $700 million, its largest purchase since Mellanox, to help customers squeeze more out of their GPU clusters. It was the first of at least five 2024 acquisitions, most of them Israeli or AI-infrastructure software firms.

SK hynix said it had begun volume production of HBM3E for supply to a customer from late March 2024. It did not name the customer; the announcement came the day after Nvidia's GTC keynote.

Nvidia announced the Blackwell architecture in San Jose, listing the B200 Tensor Core GPU and the GB200 Grace Blackwell Superchip, which pairs two B200 GPUs with a Grace CPU. The release claims up to 25x lower cost and energy for LLM inference compared with H100.

Micron said it had begun volume production of its 24GB 8-high HBM3E, which would ship inside Nvidia H200 GPUs from the second calendar quarter of 2024, making Micron a first-wave HBM3E supplier alongside SK hynix.

Nvidia reported fourth-quarter data-center revenue of $18.4 billion, up 409 percent from a year earlier, and full-year data-center revenue of $47.5 billion, up 217 percent. In one fiscal year the data-center business went from a large segment to most of the company, on the back of a single product generation.

After Nvidia's quarterly results, Jensen Huang told Reuters the company was sampling two new China-market AI chips with customers and that both complied with the rules without needing a licence. He did not name the chips or the customers. The move was Nvidia's attempt to defend a market position that the October 2023 controls had put at risk.

Reuters reported, citing two people familiar with the matter, that Nvidia had told Chinese customers it was delaying the launch of the export-compliant H20 until the first quarter of 2024, because server makers were having trouble integrating the chip into their systems. Nvidia was also preparing the L20 and L2 for the same market. Its shares closed down about 1.9 percent.

The H200 kept the Hopper architecture but moved to HBM3e memory. Nvidia said systems from server makers and cloud providers would begin shipping in the second quarter of 2024.

The Register, citing Chinese outlet Chinastarmarket and supply chain sources, reported that Nvidia was working on three GPUs for China, the H20, L20 and L2, engineered to stay below the performance thresholds in the October 2023 rules, with an announcement possible as early as November 16.

Nvidia disclosed that the government had told it the October 2023 licensing requirements for products with a total processing performance of 4800 or more, designed for data centers, were effective immediately rather than after the planned 30-day period, affecting A100, A800, H100, H800 and L40S shipments.

The October 2023 update replaced the interconnect-bandwidth parameter with a total processing performance and performance density test, closing the loophole that Nvidia's A800 and H800 had used, and added a notification requirement for chips just under the line. AMD's MI300 family fell above the new threshold, so any China part would have to be cut down much further.

BIS released updated rules on advanced computing chips and chipmaking equipment. Among the changes it removed interconnect bandwidth as a parameter for identifying restricted chips, which brought Nvidia's China-specific A800 and H800 within the controls, and extended license requirements to more countries. Nvidia's same-day filing listed A100, A800, H100, H800, L40, L40S and RTX 4090 as covered.

Second-quarter fiscal 2024 data-center revenue more than doubled from the prior quarter to a record $10.32 billion, up 141 percent sequentially and 171 percent from a year earlier, as H100 systems shipped to cloud providers and consumer-internet companies. It was the quarter the demand shift first showed in Nvidia's numbers.

The Financial Times reported that Baidu, ByteDance, Tencent and Alibaba had ordered about 100,000 A800 processors worth US$1 billion for delivery that year, and a further US$4 billion of GPUs for 2024, rushing to secure supply amid fears of further U.S. export controls. Demand for the compliant part was intact two months before the October 2023 rules removed the test it was designed around.

On the second-quarter 2023 call, as AMD prepared to ramp the MI300, Su said the company would be fully compliant with US export controls but saw an opportunity to develop an accelerator specifically for Chinese customers, following Nvidia's A800 and H800 and Intel's Gaudi variant. That product became the MI308.

With first-quarter fiscal 2024 data-center revenue at a record $4.28 billion, Jensen Huang said the computer industry was going through two transitions at once, accelerated computing and generative AI, and that Nvidia was significantly increasing supply to meet surging demand. The guidance that followed became the clearest public marker of the H100 shortage.

Nvidia said it had modified its flagship H100 into a version legal to export to China, the H800, and that the cloud computing units of Alibaba, Baidu and Tencent were already using it. It was the same move as the A800 applied to the newer generation, and Jensen Huang pointed Chinese startups building large language models at those clouds.

Microsoft said it reached a binding 10-year legal agreement with Nintendo. The agreement concerned Call of Duty availability for Nintendo Switch devices.

NVIDIA said it began shipping the Hopper-based H100 Tensor Core GPU. It said the first H100 systems were available.

Two months after the U.S. cut off the A100 and H100, Nvidia offered the A800 as an alternative for Chinese customers, saying it met the government's test for reduced export control and could not be programmed to exceed it. The part had gone into production in the third quarter. Nvidia had previously said the ban could affect as much as US$400 million of third-quarter sales to China.

The Commerce Department turned the August letters into a published rule: advanced computing chips above a performance and interconnect threshold, the supercomputers they feed and the equipment to make them all needed licenses for China. The thresholds caught the MI250 class and set the ceiling under which AMD, Nvidia and Intel would design China-specific parts.

Days after Nvidia disclosed a license requirement for the A100 and H100, AMD confirmed it had received a similar notice covering the Instinct MI250 and MI250X, while the older MI100 was unaffected. AMD said the impact on its data center business would be marginal, since most of its compute GPUs went into US and European supercomputers, and disclosed the requirement in its third-quarter 10-Q.

Nvidia disclosed that the U.S. government had imposed, effective immediately, a license requirement for exports of its A100 and forthcoming H100 chips to China, including Hong Kong, and Russia, along with systems containing them and any future chip at A100-level performance. Nvidia said it did not sell to Russia and that the rule could affect H100 development and A100 support.

Nvidia introduced Hopper, the successor to Ampere, with the H100 as its first GPU. The press release cites 80 billion transistors on a TSMC 4N process and says the H100 will be available in the third quarter.

Excelero became part of Nvidia. Nvidia said Excelero's NVMesh technology would be integrated into its enterprise software stack.

Seventeen months after signing, the companies terminated the agreement, citing significant regulatory challenges, and Nvidia wrote off the $1.36 billion it had prepaid. Arm listed on Nasdaq in September 2023 at a $54 billion valuation; Nvidia kept a 20-year Arm architecture licence and went on to build its own Arm-based Grace CPUs.

Bright Computing became part of Nvidia. Nvidia said Bright Cluster Manager would join its accelerated-computing software stack.

The Federal Trade Commission voted 4-0 to challenge the deal, arguing that owning Arm would let Nvidia stifle rivals that depend on Arm designs in data center CPUs, automotive and DPUs, and give it access to competitors' confidential information. The UK's CMA had already opened an in-depth probe on national-security and competition grounds and the European Commission a Phase 2 review.

The UK secretary of state referred Nvidia's proposed acquisition of Arm to the CMA for a Phase 2 investigation. The referral cited competition and national-security concerns.

The UK Competition and Markets Authority found that Nvidia's proposed purchase of Arm should proceed to a Phase 2 investigation. The CMA cited competition grounds for the in-depth investigation.

Nvidia announced an agreement to acquire high-definition mapping startup DeepMap. The contemporaneous report described the transaction as pending.

The largest chip deal ever proposed would have put the architecture behind nearly every smartphone under the company that dominated AI accelerators, with SoftBank keeping under 10 percent of Nvidia. Huang promised to keep Arm's open licensing model and its Cambridge headquarters; customers including Qualcomm, Google and Microsoft objected within months.

Nvidia said the first Ampere GPU, the A100, was in full production and shipping worldwide, with up to 20 times the performance of its predecessors and adoption by the top cloud providers and server makers. The A100 was the chip the large language models of 2020 to 2022 were trained on, which is why the H100 launched into an installed base that already ran on Nvidia.

Nvidia announced plans to acquire Cumulus Networks. The companies did not announce a purchase price.

The acquisition completed at a transaction value of about $7 billion, adding Mellanox's roughly 3,000 employees and the InfiniBand, Spectrum Ethernet and BlueField DPU lines that became the networking half of Nvidia's data center segment. Within five years networking was a multi-billion-dollar quarterly business inside Nvidia.

Thirteen months after signing, China's State Administration for Market Regulation became the last authority to approve the deal, attaching behavioural conditions for six years: no bundling of GPUs with networking gear, supply on fair and non-discriminatory terms, and no restricting Chinese customers' access. The EU, Mexico and the US waiting period had already cleared.

Nvidia acquired data-storage and management company SwiftStack. The transaction value was not disclosed.

Nvidia's first big acquisition, at $125 a share in cash, brought in the Israeli maker of InfiniBand and Ethernet interconnects that already sat inside most of the world's top supercomputers, after reports that Intel and Xilinx had also bid. Jensen Huang framed it as uniting compute and networking for data-center-scale computing; the deal needed approvals in the US, EU, Mexico and China.

At GTC 2017 Nvidia launched the Volta platform and the Tesla V100, with 640 Tensor Cores delivering 120 teraflops of deep-learning performance, which it equated to 100 CPUs. Tensor Cores made matrix math a first-class hardware feature, the design choice every later generation has extended.

Krizhevsky, Sutskever and Hinton reported a 60-million-parameter convolutional network that cut ImageNet's top-5 error to 18.9 percent, far below the previous state of the art, and credited "a very efficient GPU implementation" for making the training feasible. The result is the moment deep learning and Nvidia GPUs became tied together in research practice.

Nvidia introduced CUDA as a new architecture for computing on its graphics processors and the first C-compiler development environment for the GPU, pitching hundreds of on-chip cores working in parallel at up to 100 times the speed of conventional approaches. It is the software layer that later let researchers train neural networks on gaming cards, and the ecosystem rivals still have to work around.