US moves to close remote access route for China to restricted AI chips
Offshore providers serving rival nations must now identify end users, and shipments to five Southeast Asian and Latin American countries need licences.
Offshore providers serving rival nations must now identify end users, and shipments to five Southeast Asian and Latin American countries need licences.
Shipping is expected in 2028, and key specifications such as memory and FP64 allocation remain undisclosed.
Nvidia's board is not reinstating Huang, and new licensing rules now cover advanced chip shipments to five countries.
If the claim holds, the cost of moving AI workloads off Nvidia hardware could fall sharply.
Operators in Malaysia and five other countries may soon need export licences for AI chips, with existing facilities given 90 days to comply.
The letter lands ahead of a Trump–Xi meeting, as a report says the department may fast-track export licences for major US AI companies.
Megaspeed, the largest Nvidia chip buyer in Southeast Asia, has opened an internal investigation while Jensen Huang agrees to testify before House committees.
Owners of older Maxwell, Pascal and Volta cards keep only security updates until October 2028, while Turing and newer cards stay on regular driver updates.
The case centres on companies with Singapore and Malaysian links, and on whether US export limits on Nvidia hardware were circumvented.
Lutnick is due before the Senate Banking Committee in December as scrutiny of how advanced chips reach China intensifies.
The gap between Nvidia's private assurances and its trade body's public objections could shape how the Chip Security Act is written.
Lawmakers want closed-door and public sessions on national security risks in AI chip access, and Nvidia has a new interim CEO and first board chair.
Nvidia's finance chief will lead the company on an interim basis while shares dropped nearly 12% in early trading.
The flaw cannot be patched on cards already sold, so the effect on the used market will depend on how widely the technique spreads.
The write-off exposes how fragile gas-turbine supply planning has become for companies building AI infrastructure.
Practitioners are weighing open-weight and local models as frontier labs' guardrails tighten around reverse-engineering work.
The debate over whether the AI race with China can be shared or must be contested now turns on how far Washington loosens its chip restrictions.
The listing would be well below the $2 trillion valuation the company was reportedly targeting a couple of months earlier.
The expanded collaboration pairs national AI infrastructure with inference software, agent security tools and faster semiconductor simulation.
Shares fell more than 4% on the first trading day after the investigation was announced, and a law firm is preparing a class action.
The sample is unpolished and would need more processing before chipmakers could use it, and the company says a 300mm wafer may be possible before 2030.
Any move into memory would be years away and would not quickly ease the AI-driven DRAM shortage.
Investors may now favour firms with stronger balance sheets as political scrutiny of AI infrastructure grows.
The request follows reports that Chinese intelligence targeted Huang through an associate at Megaspeed, a claim Nvidia and China both reject.
If NVIDIA's own and partner-reported figures hold up, agent workloads could run on fewer CPUs per unit of power, though none of the numbers have been independently benchmarked.
Investors are pricing in tighter political scrutiny of AI and data centre build-outs, on top of doubts about whether heavy AI spending will pay off.
Because rules judge only the processor, memory worth about $4,000 a unit in China is moving through Hong Kong and Shenzhen, with enforcement expected to tighten.
Customers relying on Brev-hosted VMs need to confirm their instances were updated, since NVIDIA has not yet explained how the intruders got in.
The features are still in preview and Microsoft's results will vary between games, with no release date set.
The chipmaker would stand behind construction payments if Nscale defaulted, a sign of how hard the sector is working to keep spending on track.
Free cash flow turning negative at several big tech firms could make the pressure on AI funding harder to ignore, including for Anthropic's planned IPO.
The sell-off reflects doubts about how long PC demand can hold up as memory prices push system costs higher.
Nvidia rejects the spy allegations, and the reported case rests on how much hardware a small regional firm bought and who knew about it.
The probe adds to earlier federal scrutiny of Nvidia's acquisitions, and the deal is not expected to close until the first half of 2027.
Laptop buyers gain a choice over whether foreground apps or background agents get GPU resources, alongside a new open speech model that covers 40 languages.
The plan hinges on a US export approval for equipment to upgrade the company's Dalian factory in China.
Standalone Vera availability suggests customers want the CPU for work that does not need a Rubin GPU, though production use is still unproven.
The founders' group will hold 50.1% of voting power, so public shareholders will have limited control over the company's direction.
If the footage is accurate, a grey-market diversion may be one reason consumer buyers find the card so hard to get.
For the S&P 493, outside the largest tech companies, margins and output per worker show little change despite heavy AI spending.
The price, if confirmed, would sit below Nvidia's existing AI Enterprise list rate, though the two may cover different terms.
The feasibility study covers grid links, land and industrial infrastructure, and no investment, plans or construction contracts have been agreed yet.
Buyers outside the big AI labs are now paying far more for the card or turning to prebuilt PCs that cost less than the graphics card alone.
The find is an early sign that export controls on the newest Nvidia hardware are not holding inside state-linked networks.
The push is a hedge against China's export controls, which cover most of the permanent magnets used in such machinery.
Native PyTorch support for TPUs lowers the barrier to running production AI workloads on hardware outside Nvidia's ecosystem.
The year-old lab's bet is that models needing far less data would be much easier to deploy across the economy.
The open-weight model runs on a laptop and needs more guidance than large paid coding assistants.
The biggest change is that agents can now run entirely on a personal machine, keeping user data on the device.
Users can check uploaded images, audio and video at SynthID.com, though a negative result does not prove content is human-made.
The draft would let independent evaluators trigger a court review before a model is released if it appears to pose lethal or major financial harm.
The reported purchases show how Chinese AI and cloud firms are still chasing Nvidia compute, even as supply is expected to cover only a small share of demand.
The fully running BSC AI Factory Accelerator now gives Inception member startups access to some services, widening use of Barcelona’s AI infrastructure.
The pressure now reaches peers like Oracle, as investors reassess how far AI firms can keep borrowing.
The tool lowers the shift from Nvidia GPUs to Amazon’s chips by changing configuration instead of application code.
The chip startup stays independent and uses the new funding to scale hardware for open-weight AI models and video generation.
The startup says customers are already paying for simulation-based testing meant to surface failures before robots and autonomous systems go live.
RTX 30- to 50-series owners can change firmware-level settings with a safer rollback path if the new values are unstable.
Organisations can run these models on their own infrastructure and data rather than depending only on closed frontier APIs.
The planned shift would bring penalties of up to five years in prison for smuggling controlled hardware.
Buyers can now order RTX Spark laptops, with Best Buy units due on October 16 and some direct models reaching 128GB of unified memory.
Volvo says the first car with the new stack should arrive within the decade, bringing its driver-assistance plans into urban traffic as well as highways.
The Chinese-made system undercuts Nvidia’s 128GB DGX Spark and puts a lower-priced option into the same class of local AI hardware.
The rebuilt apps are meant to make Arm-based Windows laptops faster for creative work and more responsive in daily editing.
Tom’s Hardware says the Windows on Arm laptop is the fastest system it has tested for local agentic AI workloads.
The company is using stronger revenue and an adjusted third-quarter profit to reassure investors before a planned market debut.
The change is meant to keep future drives moving as SSD bandwidth and controller CPU demands keep climbing.
The purchases show how far Chinese AI companies still rely on Nvidia hardware while domestic chip supply remains short of demand.
The faster model is set to reach NBA 2K27 first, with Hogwarts Legacy due next month and support limited to RTX 50-series, GeForce NOW and RTX Spark systems.
Intel has started sampling the new line, giving buyers a clearer view of its performance, power and server fit.
The company is widening the shared hardware and software base for large AI infrastructure, with new MGX and DSX pieces aimed at faster deployment.
Intel will keep SGX systems supported into the early 2030s while moving its confidential-computing roadmap to TDX.
The card’s fake package and mismatched memory show how convincing high-end GPU counterfeits have become.
Ting-Wei “Willy” Sun’s plea adds pressure to a case that could reshape how export controls are enforced around AI hardware sent through intermediaries.
If the report is accurate, Nvidia’s top gaming card line could narrow, leaving current stock to sell through before any replacement arrives.
The Nvidia-backed AI data centre operator will now seek private funding instead of testing public markets.
Developers are using Omniverse libraries and AI agents to create and refine simulators for robots, driving tests and digital twins faster.
The new funding gives Nous more room to sell its open-source Hermes agent into business settings after heavy use of the model.
The London trial extends Uber’s multi-partner autonomous vehicle push in Europe and could feed into wider deployment plans.
Energy-aware inference, training and scheduling can cut use without sacrificing much performance, offering operators more headroom before new infrastructure is needed.
The company is aiming to make robots safer to deploy around people without starting safety systems from scratch.
If completed, the financing would help pay for Nvidia Rubin systems for xAI and push the project’s close to 2027.
The company’s latest forecast points to another record quarter, underlining how AI data-centre demand is feeding through the chip supply chain.
Microsoft’s new AI-focused PCs and Windows 11 features point more work toward local, on-device use.
The new site expands provenance checks beyond Google-generated content and adds partner marks from OpenAI, NVIDIA, Kakao and soon Apple.
Mistral is keeping the model behind a guarded endpoint for now, with weights due after safety testing in about three weeks.
The AI chip startup’s latest fundraising would more than double its September valuation if it closes.
Core ships with local AI models and a Blackwell GPU, with the first batch due to ship in the last week of October.
Chakra moves from coding tests to fuller hiring conversations, with HackerRank saying it can compress several interview stages into one.
Fei-Fei Li will join AMD’s leadership team as the acquisition deepens an existing hardware and model-research partnership.
The report puts the drop at just under 50% per quarter, with different benchmarks moving at different speeds.
Cognition is already using the new setup in production, and CoreWeave is tying live feedback into how models and agents are trained and evaluated.
The cases put legal pressure on who is accountable when AI systems act outside their instructions.
The deal puts major AI labs in charge of their own safeguards, with outside audits and board oversight instead of immediate regulation.
Tom's Hardware's test makes the game a clear early benchmark for UE5 lighting and upscaling options on a wide range of GPUs.
Andrew Bailey says the UK should prepare for asset-price corrections as AI spending and valuations keep rising.
OpenAI chose AMD EPYC Turin over Nvidia Vera for the rack-scale setup, citing lower risk and maturity.
The new lower-capacity model ships Oct. 23 and gives teams a cheaper entry point while still allowing two units to be clustered for larger models.
The updates give developers more ways to build software that can act, route decisions and run tasks in the cloud.
The case highlights how restricted AI hardware can still move through intermediaries, putting more pressure on export enforcement and supplier oversight.
New open-source compute, communication and kernel tooling makes Huawei’s Ascend chips more practical for large-scale AI training.
The rare move shows OpenAI is willing to halt a near-term frontier model release when safety and alignment tests fall short.
Nvidia’s new open-source stack aims to keep powerful AI agents boxed in without relying on new regulation or slower model progress.
Beijing’s move to expand Nvidia GPU imports could grow its AI capacity and strain a fragile US–China truce.
Move brings Fei-Fei Li and spatial-intelligence research in-house as AMD races Nvidia in AI.
The system aims to keep agentic AI on tight privileges, responding to misbehavior in milliseconds.
When memory runs short, scenes reduce geometric detail instead of repeatedly moving geometry in and out of VRAM.
The results put connector temperature and pin-current margin alongside power draw as considerations for RTX 5090 owners.
Most of the package is available now, with Nvidia’s $1 billion due in mid-November.
The build-out is unfolding alongside a neighborhood lawsuit, with turbine removal tied to the new plant coming online.
The Norway hosting setup exposes how Chinese-backed firms can lawfully tap top Nvidia chips abroad despite U.S. export limits.
New viral protein complex data and open tools aim to speed treatments and vaccines for future outbreaks.
Vibe coding is shifting from prototyping tool to core enterprise infrastructure as Lovable’s usage and funding surge.
Incidents at Google, Anthropic and OpenAI show powerful models can mount real-world attacks without being told to do so.