# Nvidia — AI news & analysis

Nvidia makes the GPUs that train and run most frontier AI, making it the pivotal hardware supplier of the current boom.

- [Microsoft unveils Azure Maia 300, calls it primary inference system](https://www.parallelquant.com/posts/microsoft-unveils-azure-maia-300-calls-it-primary-inference-system-2278ab) (2026-10-13, Data Center Dynamics): Microsoft announced the Azure Maia 300 chip and said it will be its primary inference system. The earlier Maia 200 is already deployed across multiple regions.
- [Intel's Crescent Island inference accelerator ships as GPU D-Series](https://www.parallelquant.com/posts/intel-s-crescent-island-inference-accelerator-ships-as-gpu-d-series-369533) (2026-10-12, Tom's Hardware): Intel will bring its Crescent Island AI inference accelerator to market as the GPU D-Series family. It offers 1.3 PFLOPS of FP8 compute and up to 480 GB of LPDDR5X memory with 1.5 TB/s of bandwidth.
- [Nvidia reportedly halts GeForce RTX 5090 production for data center GPUs](https://www.parallelquant.com/posts/nvidia-reportedly-halts-geforce-rtx-5090-production-for-data-center-gpus-bd17f2) (2026-10-10, Tom's Hardware): Tom's Hardware reports Nvidia has discontinued GeForce RTX 5090-series cards, leaving the GB202 chip for data center and professional graphics boards. The RTX 5080 24 GB is rumored to become the de facto consumer flagship for a while. The report is unconfirmed.
- [NVIDIA's PivotOPD trains agents to recover from early mistakes](https://www.parallelquant.com/posts/nvidia-s-pivotopd-trains-agents-to-recover-from-early-mistakes-1ccc86) (2026-10-08, MarkTechPost): NVIDIA researchers introduced PivotOPD, an on-policy distillation method for multi-turn LLM agents that targets pivotal early mistakes and teaches recovery from them. It posted the best average result against 13 baselines across 3 agent benchmarks.
- [SpaceX reportedly seeks $40B debt to buy Nvidia Rubin chips](https://www.parallelquant.com/posts/spacex-reportedly-seeks-40b-debt-to-buy-nvidia-rubin-chips-b68315) (2026-10-08, Tom's Hardware): SpaceX is reportedly looking to borrow $40 billion to procure 360,000 Nvidia Rubin AI accelerators and supporting infrastructure. The report comes via Tom's Hardware and is not confirmed by the company.
- [Microsoft launches Nvidia RTX Spark AI PCs and agent-focused Windows 11](https://www.parallelquant.com/posts/microsoft-launches-nvidia-rtx-spark-ai-pcs-and-agent-focused-windows-11-cba796) (2026-10-07, TechCrunch): Microsoft detailed the Surface Laptop Ultra, powered by Nvidia's Arm-based RTX Spark chip, starting at $2,599 and shipping October 16. A $5,999 Surface RTX Spark Dev Box ships in November. Windows gains 'Hybrid Intelligence' features letting Copilot use local files to take actions.
- [Report: China stockpiled 343 immersion DUV chipmaking tools](https://www.parallelquant.com/posts/report-china-stockpiled-343-immersion-duv-chipmaking-tools-39f1e3) (2026-10-05, Tom's Hardware): The Centre for Technology & Statecraft says China has accumulated 343 immersion deep-ultraviolet lithography tools used for advanced chipmaking. It calls for banning exports of all immersion DUV tools to China, echoing the MATCH Act proposed by US legislators.
- [Nvidia launches 64GB DGX Spark from $4,999 for local AI](https://www.parallelquant.com/posts/nvidia-launches-64gb-dgx-spark-from-4-999-for-local-ai-22fb9c) (2026-10-02, Tom's Hardware): Nvidia is adding a 64GB unified-memory version of its DGX Spark local AI workstation, starting at $4,999. It is otherwise identical to the original 128GB model and GB10 siblings. It targets newer compact but capable local models.
- [Amazon and Synopsys sign billion-dollar deal on AI chip design](https://www.parallelquant.com/posts/amazon-and-synopsys-sign-billion-dollar-deal-on-ai-chip-design-4e2639) (2026-10-02, Tom's Hardware): Amazon will license Synopsys chip designs and design tools to create and optimize new AI chips in a multi-year partnership worth over a billion dollars. Synopsys will adopt Amazon Bedrock to build and deploy AI agents, use AWS compute and storage, and optimize its tools for Amazon hardware.
- [US arrests tech CEO over alleged $300M Nvidia chip smuggling to China](https://www.parallelquant.com/posts/us-arrests-tech-ceo-over-alleged-300m-nvidia-chip-smuggling-to-china-0ca77c) (2026-10-02, Ars Technica): US authorities arrested a tech chief executive accused of smuggling about $300 million in Nvidia chips into China. Ars Technica notes that arrests over chip smuggling keep occurring.
- [OpenAI pairs its Jalapeño AI chips with AMD EPYC Turin hosts](https://www.parallelquant.com/posts/openai-pairs-its-jalapeno-ai-chips-with-amd-epyc-turin-hosts-a8d567) (2026-10-02, Tom's Hardware): OpenAI is deploying rack-scale systems of its in-house Jalapeño ASIC (application-specific integrated circuit) with AMD EPYC 'Turin' CPUs as host processors. It did not choose newer agent-oriented chips such as Arm's AGI or Nvidia's Vera.
- [Nvidia CEO dismisses Anthropic's AI existential-risk warnings](https://www.parallelquant.com/posts/nvidia-ceo-dismisses-anthropic-s-ai-existential-risk-warnings-0dbe4a) (2026-09-20, Tom's Hardware): Nvidia CEO Jensen Huang said there is a '0% chance' AI destroys the world by 2030, arguing existing safety mechanisms are sufficient. He said the industry 'should go as fast as we can, irrespective of anyone else' and rejected calls for new AI regulation, directly pushing back on Anthropic's public safety warnings.
- [Dario Amodei proposes plan to 'pace' frontier AI development](https://www.parallelquant.com/posts/dario-amodei-proposes-plan-to-pace-frontier-ai-development-1013c2) (2026-09-18, TechCrunch): Anthropic CEO Dario Amodei outlined a proposal to "pace the frontier" of AI development, relying on independent safety evaluators and coordination between AI labs in democratic countries. The plan has picked up some industry support alongside pushback from Nvidia's Jensen Huang, and follows a doomsday warning from an Anthropic researcher that rattled the AI industry the prior week.
- [AMD claims EPYC 'Venice' CPUs sharply beat Nvidia's Vera](https://www.parallelquant.com/posts/amd-claims-epyc-venice-cpus-sharply-beat-nvidia-s-vera-a64bb6) (2026-09-18, Tom's Hardware): AMD released first official benchmarks for its upcoming EPYC 'Venice' server CPUs, claiming its 256-core chip is more than twice as fast as Nvidia's Vera and its 96-core model is 20% faster per-core. These are AMD's own figures, ahead of independent verification.
- [Huawei pulls forward next-gen Ascend AI chip roadmap by quarters](https://www.parallelquant.com/posts/huawei-pulls-forward-next-gen-ascend-ai-chip-roadmap-by-quarters-364403) (2026-09-18, Tom's Hardware): Huawei detailed its AI accelerator roadmap, moving up next-generation Ascend NPUs and Kunpeng CPUs by several quarters. The company said FP4 performance on the upcoming Ascend 960PR doubles prior expectations, and laid out scale-up and scale-out connectivity plans modeled on Nvidia's AI-factory approach.
- [Crusoe raises $3.9B to build AI data centers, hits $30.9B valuation](https://www.parallelquant.com/posts/crusoe-raises-3-9b-to-build-ai-data-centers-hits-30-9b-valuation-d6bec5) (2026-09-17, TechCrunch): Crusoe, a data center and AI infrastructure company, raised $3.9 billion in new funding, valuing it at $30.9 billion. The money will fund large-scale data centers and smaller modular "AI factories."
- [Google, Nvidia found alliance for AI data center demand response](https://www.parallelquant.com/posts/google-nvidia-found-alliance-for-ai-data-center-demand-response-75b8ec) (2026-09-17, Data Center Dynamics): Google, Nvidia, and Emerald AI have founded the AI Energy Management Alliance to support demand-response capabilities in data centers, with launch partners including Anthropic and utility RWE. Demand response lets data centers flex their power draw to ease strain on electrical grids.
- [Report details how China obtains billions in banned Nvidia chips](https://www.parallelquant.com/posts/report-details-how-china-obtains-billions-in-banned-nvidia-chips-3c7ac8) (2026-09-17, Tom's Hardware): A report from C4ADS, a US-government-funded monitoring nonprofit, documents how Chinese firms circumvent US export restrictions to acquire billions of dollars worth of restricted Nvidia AI accelerators. The investigation lays out specific routes and methods used to move the chips around the controls.
- [Apple reportedly building AI server with M-series Ultra chips](https://www.parallelquant.com/posts/apple-reportedly-building-ai-server-with-m-series-ultra-chips-942348) (2026-09-16, Ars Technica): Apple is said to be developing an enterprise server built around multiple M-series Ultra chips for AI inference, according to The Information. A launch isn't expected before 2029, and Apple is reportedly considering Nvidia's NVLink Fusion to connect the chips.
- [Nvidia's Earth-2 platform used to forecast UK air pollution](https://www.parallelquant.com/posts/nvidia-s-earth-2-platform-used-to-forecast-uk-air-pollution-532b5f) (2026-09-16, NVIDIA): The University of Manchester is using Nvidia's Earth-2 AI platform to forecast air pollution across the UK, which contributes to an estimated 30,000 deaths there annually. Earth-2 uses AI models instead of traditional chemistry-based simulations, making finer-grained, more frequent forecasts computationally feasible.
- [Jensen Huang: China will master advanced chip lithography by 2030](https://www.parallelquant.com/posts/jensen-huang-china-will-master-advanced-chip-lithography-by-2030-085983) (2026-09-16, Tom's Hardware): Nvidia CEO Jensen Huang said he expects China to develop its own advanced lithography tools for chipmaking within three to four years, calling it "just a matter of time." He suggested China's trajectory means it is effectively already on that path.
- [Axelera AI's Europa inference chip ships in Dell and Supermicro servers](https://www.parallelquant.com/posts/axelera-ai-s-europa-inference-chip-ships-in-dell-and-supermicro-servers-7cf709) (2026-09-16, Data Center Dynamics): Axelera AI launched its Europa inference chip, now shipping inside Dell XE5 and Supermicro 111AD servers. The chip targets AI inference workloads as an alternative to Nvidia hardware.
- [Data-retention changes are costing AI labs enterprise trust](https://www.parallelquant.com/posts/data-retention-changes-are-costing-ai-labs-enterprise-trust-fb1e85) (2026-09-15, The Decoder): OpenAI and Anthropic tell corporate customers their data won't be used for training, but when Anthropic said it would store usage logs from its Fable model for 30 days, Palantir, Nvidia, and Booz Allen Hamilton pulled back from using it for sensitive work. It illustrates a gap between AI labs' privacy policies and what security-conscious enterprise customers will actually accept.
- [Perplexity brings local AI agent to Windows RTX PCs](https://www.parallelquant.com/posts/perplexity-brings-local-ai-agent-to-windows-rtx-pcs-54f916) (2026-09-15, Tom's Hardware): Perplexity and Nvidia released 'Portable Computer' for Windows on September 14, a local-first AI agent that can perform multistep tasks. It requires GeForce RTX or RTX PRO GPUs with at least 24GB of VRAM.
- [Salesforce launches Koa, a reasoning model for sales tasks](https://www.parallelquant.com/posts/salesforce-launches-koa-a-reasoning-model-for-sales-tasks-043837) (2026-09-15, TechCrunch): Salesforce released Koa, a reasoning-focused AI model built on Nvidia's open-weight Nemotron model. It's trained specifically for sales, marketing, and customer-support tasks rather than general-purpose use.
- [Nvidia launches RTX Pro 5500 workstation GPU for AI workloads](https://www.parallelquant.com/posts/nvidia-launches-rtx-pro-5500-workstation-gpu-for-ai-workloads-169f4e) (2026-09-14, Tom's Hardware): Nvidia introduced the RTX Pro 5500 Blackwell Workstation Edition, a professional graphics card aimed at agentic and generative AI workloads. The new card offers 2.6 times the VRAM of comparable gaming specs, with Nvidia reportedly deprioritizing gaming performance on this line.
- [Cornelis raises $205M for networking tech to challenge Nvidia](https://www.parallelquant.com/posts/cornelis-raises-205m-for-networking-tech-to-challenge-nvidia-c5f728) (2026-09-14, TechCrunch Startups): AI infrastructure company Cornelis raised $205 million and announced Active Compute Fabric, a networking technology aimed at reducing the time GPUs spend idle while waiting for data. The funding targets a specific bottleneck in Nvidia's dominance over AI data center interconnects.
- [Nvidia open-sources OSMO orchestrator for robot AI training](https://www.parallelquant.com/posts/nvidia-open-sources-osmo-orchestrator-for-robot-ai-training-775161) (2026-09-14, MarkTechPost): Nvidia open-sourced OSMO, the Kubernetes-native workflow orchestrator it uses internally for Project GR00T, Isaac Lab, and Isaac Sim. OSMO lets robotics teams define training, simulation, and hardware-in-the-loop testing in a single YAML file and routes jobs across compute tiers, from GB200 clusters down to Jetson AGX Thor edge devices, under an Apache 2.0 license at version 6.3.1.
- [Nvidia partners with Australian firms for 2GW of AI capacity by 2027](https://www.parallelquant.com/posts/nvidia-partners-with-australian-firms-for-2gw-of-ai-capacity-by-2027-a0507d) (2026-09-10, Data Center Dynamics): Nvidia is partnering with cloud providers and data center operators in Australia to bring 2 gigawatts of AI computing capacity online by 2027.
- [Universal Music launches AI remix platform with ElevenLabs](https://www.parallelquant.com/posts/universal-music-launches-ai-remix-platform-with-elevenlabs-8a38a5) (2026-09-10, The Verge): Universal Music Group is launching an AI-powered platform, built with ElevenLabs under a multiyear licensing agreement, that lets users create remixes and mashups from its licensed catalog. Artists can choose whether to opt in. UMG is separately developing a similar platform with Udio and already has AI licensing deals with Spotify and Nvidia.
- [Sakana AI launches Fugu models for multi-agent orchestration](https://www.parallelquant.com/posts/sakana-ai-launches-fugu-models-for-multi-agent-orchestration-cabf49) (2026-09-11, MarkTechPost): Sakana AI released Fugu Max and Fugu Ultra v2, two models sharing a learned orchestration architecture that routes tasks across lean open and specialized models, including Nvidia Nemotron. Fugu Max costs $2/$6 per million tokens, while Fugu Ultra v2 targets peak capability, scoring 48.3 on Chartography and 74.3 on DeepSWE.
- [Nvidia in talks to invest $10B in Anthropic's IPO](https://www.parallelquant.com/posts/nvidia-in-talks-to-invest-10b-in-anthropic-s-ipo-852d32) (2026-09-12, The Decoder): Reuters reports Nvidia is discussing investing up to $10 billion in Anthropic's planned IPO, which could value the company at $2 trillion and be the largest IPO in history.
- [Altman rules out an OpenAI IPO in 2026](https://www.parallelquant.com/posts/altman-rules-out-an-openai-ipo-in-2026-376b5f) (2026-09-12, TechCrunch): OpenAI CEO Sam Altman said taking OpenAI public in 2026 would be "ill-advised," even though the company has filed confidentially for an IPO. He cited ongoing safety considerations, including the possibility of pausing training if needed.
- [Modded Nvidia RTX 5090 with 96GB memory appears on Alibaba for under $4,000](https://www.parallelquant.com/posts/modded-nvidia-rtx-5090-with-96gb-memory-appears-on-alibaba-for-under-4-0-4aa06a) (2026-09-11, Tom's Hardware): A China-modified Nvidia GeForce RTX 5090 with 96GB of VRAM, three times the original card's memory, has surfaced for sale on Alibaba for about $3,888, roughly 65% of the original card's price, according to Tom's Hardware.
- [DOJ reportedly probing Nvidia's Groq deal over acqui-hire concerns](https://www.parallelquant.com/posts/doj-reportedly-probing-nvidia-s-groq-deal-over-acqui-hire-concerns-db0b26) (2026-09-10, Data Center Dynamics): The Department of Justice has reportedly opened an investigation into Nvidia's deal involving AI chip startup Groq, adding to an existing FTC inquiry. Regulators are examining whether the arrangement functions as a de facto acquisition structured to avoid merger review.
- [Skild AI's robot model learns new tasks from a single video](https://www.parallelquant.com/posts/skild-ai-s-robot-model-learns-new-tasks-from-a-single-video-f5ae01) (2026-09-10, NVIDIA): Skild AI's S1 robot foundation model, built with Nvidia's Physical AI tools, can learn previously unseen, long-horizon tasks from a single video demonstration rather than extensive reprogramming. It targets manufacturing and warehouse settings where layouts and tasks change frequently.
- [Nvidia's new runtime nearly triples protein-folding throughput](https://www.parallelquant.com/posts/nvidia-s-new-runtime-nearly-triples-protein-folding-throughput-534a8a) (2026-09-10, MarkTechPost): Nvidia detailed BioNeMo Inference Runtime (BioIR), a library that speeds up biomolecular structure-prediction models while staying in standard PyTorch. In an 8-GPU benchmark it delivered 58,500 folded residues per GPU-hour with the Boltz-2 model, a 2.9x gain over a torch-compiled baseline, and it already powered an AlphaFold Database expansion generating about 31 million candidate protein complexes.
- [Nvidia and Palantir team up to run supply chains with AI](https://www.parallelquant.com/posts/nvidia-and-palantir-team-up-to-run-supply-chains-with-ai-214af9) (2026-09-10, The Decoder): Nvidia and Palantir are partnering to apply AI to supply chain management, starting with Nvidia's own operation spanning roughly a million parts. The companies plan to extend the approach to other clients' supply chains afterward.
- [China's Biren posts 2,000% revenue growth as Nvidia, AMD exit market](https://www.parallelquant.com/posts/china-s-biren-posts-2-000-revenue-growth-as-nvidia-amd-exit-market-483f68) (2026-09-10, Tom's Hardware): Chinese AI accelerator maker Biren Technology reported roughly 2,000% year-over-year revenue growth in the first half of 2026. The surge coincides with Nvidia and AMD effectively exiting the Chinese market under US export controls.
- [d-Matrix joins Nvidia's NVLink Fusion for its next-gen AI chips](https://www.parallelquant.com/posts/d-matrix-joins-nvidia-s-nvlink-fusion-for-its-next-gen-ai-chips-2caf00) (2026-09-10, NVIDIA): AI inference chipmaker d-Matrix will use Nvidia's NVLink Fusion to connect its upcoming Raptor XPUs into Nvidia's rack-scale AI infrastructure. This ties Raptor into Nvidia's scale-up and scale-out networking within the MGX rack architecture.
- [Nvidia brings Rust to GPU kernel programming with CUDA Rust](https://www.parallelquant.com/posts/nvidia-brings-rust-to-gpu-kernel-programming-with-cuda-rust-3e447a) (2026-09-08, MarkTechPost): Nvidia announced CUDA Rust, backed by two open-source NVlabs projects: cuda-oxide, which compiles SIMT kernels from Rust MIR through Pliron and LLVM to PTX, and cutile-rs, which JIT-compiles Tile kernels through CUDA Tile IR on stable Rust 1.89+.
- [Qualcomm and AWS to co-develop custom silicon for AI data centers](https://www.parallelquant.com/posts/qualcomm-and-aws-to-co-develop-custom-silicon-for-ai-data-centers-f04b06) (2026-09-08, Data Center Dynamics): Qualcomm and AWS announced a partnership to build custom silicon for large-scale AI data centers, alongside joint work on an optical connectivity solution supporting up to 1.6 terabits per second.
- [AMD unveils workstation with dual MI350P AI accelerators](https://www.parallelquant.com/posts/amd-unveils-workstation-with-dual-mi350p-ai-accelerators-38d28d) (2026-09-04, Tom's Hardware): AMD's new Threadripper Halo Station pairs a 96-core Zen 5 Threadripper CPU with dual liquid-cooled MI350P accelerators (expandable to four) and 2TB of DDR5 memory. AMD says the system can run trillion-parameter models locally.
- [DeepSeek plans 160,000-chip Huawei cluster in China](https://www.parallelquant.com/posts/deepseek-plans-160-000-chip-huawei-cluster-in-china-cfb1a2) (2026-09-04, The Decoder): DeepSeek intends to deploy 160,000 Huawei Ascend-950DT chips at a data center in Inner Mongolia, dedicated to inference rather than training. It would be the largest known Huawei chip cluster, though production constraints mean Huawei likely can't deliver the full order for over a year.
- [Broadcom says AI chip demand drove a strong quarter, expects more growth](https://www.parallelquant.com/posts/broadcom-says-ai-chip-demand-drove-a-strong-quarter-expects-more-growth-a924c8) (2026-09-04, Data Center Dynamics): Broadcom reported a strong third quarter driven by AI silicon demand, including its custom AI accelerator business, and said it expects that growth to keep 'double-doubling.' Its VMware software unit continued to perform steadily alongside the chip business.
- [Nvidia's PAIR spreads local AI compute across home devices](https://www.parallelquant.com/posts/nvidia-s-pair-spreads-local-ai-compute-across-home-devices-b58b73) (2026-09-04, The Decoder): Nvidia's Personal AI Router (PAIR) automatically distributes local AI requests across every available device on a home network. Nvidia says this cuts wait times for parallel agent tasks.
- [Anthropic signs $35 billion cloud deal with Lambda](https://www.parallelquant.com/posts/anthropic-signs-35-billion-cloud-deal-with-lambda-e9d308) (2026-09-03, The Decoder): Anthropic has signed a $35 billion cloud computing deal with Lambda, an Nvidia-backed cloud provider, to expand infrastructure for Claude. The agreement significantly scales up Anthropic's committed compute capacity.
- [TSMC's fab equipment demand nearly doubles as AI drives expansion](https://www.parallelquant.com/posts/tsmc-s-fab-equipment-demand-nearly-doubles-as-ai-drives-expansion-8049fc) (2026-09-03, Tom's Hardware): TSMC's fab equipment requirements have nearly doubled over the past eight months as AI chip demand drives an unprecedented expansion in manufacturing capacity, pushing 2026 capital expenditure toward roughly $64 billion. That capex figure rose only about 15% year-over-year despite the surge in equipment needs, pointing to tool shortages across the supply chain.
- [Nvidia to acquire Hugging Face for $12.9 billion](https://www.parallelquant.com/posts/nvidia-to-acquire-hugging-face-for-12-9-billion-1d3534) (2026-09-03, TechCrunch): Nvidia has agreed to acquire Hugging Face, the platform hosting over 3 million AI models and used by more than 18 million developers, for $12.93 billion. The deal brings one of the most widely used open-source AI model and dataset hosting platforms under the ownership of the world's largest AI chipmaker.
- [Lambda raises $1 billion in debt to buy Nvidia GPUs for Microsoft](https://www.parallelquant.com/posts/lambda-raises-1-billion-in-debt-to-buy-nvidia-gpus-for-microsoft-329e0b) (2026-09-01, Data Center Dynamics): Cloud provider Lambda reportedly secured $1 billion in private debt financing to purchase Nvidia GPUs. The GPUs will reportedly be leased to Microsoft.
- [Nvidia invests $3.5 billion in MediaTek for custom AI chip partnership](https://www.parallelquant.com/posts/nvidia-invests-3-5-billion-in-mediatek-for-custom-ai-chip-partnership-f44f77) (2026-09-01, Tom's Hardware): Nvidia is investing $3.5 billion in MediaTek, with MediaTek adopting Nvidia's NVLink Fusion interconnect for its own custom AI accelerators. The partnership also extends to local AI computing and automotive platforms.
- [Nvidia and CrowdStrike launch agentic cybersecurity system SafeMind](https://www.parallelquant.com/posts/nvidia-and-crowdstrike-launch-agentic-cybersecurity-system-safemind-17d844) (2026-09-01, NVIDIA): Nvidia and CrowdStrike announced SafeMind, an agentic cybersecurity system, at CrowdStrike's Fal.Con 2026 conference in Las Vegas. Nvidia CEO Jensen Huang and CrowdStrike CEO George Kurtz framed it as automated defense for automated attacks.
- [Nvidia to raise AI server prices about 15% on memory shortage](https://www.parallelquant.com/posts/nvidia-to-raise-ai-server-prices-about-15-on-memory-shortage-7fd7fe) (2026-08-23, Tom's Hardware): Nvidia has told major customers it will raise prices roughly 15% on Grace Blackwell and Vera Rubin AI server systems shipping in early 2027, driven by a DRAM shortage from Samsung, SK Hynix, and Micron. The increases will hit large cloud buyers including Microsoft, Google, and Meta, which rely on these systems for AI data center buildouts.
- [Nvidia research: agent harness matters more than model quality](https://www.parallelquant.com/posts/nvidia-research-agent-harness-matters-more-than-model-quality-7dae88) (2026-08-21, TechCrunch): Nvidia research shows that AI agents can perform reliably on tasks, without going off track, primarily through fine-tuning of the surrounding harness rather than improvements to the underlying model itself. The finding held even when the base model used wasn't particularly strong at the task.
- [LG enters chip packaging with new laser imaging machine](https://www.parallelquant.com/posts/lg-enters-chip-packaging-with-new-laser-imaging-machine-4b3b96) (2026-08-21, Tom's Hardware): LG has rolled out a laser direct imaging (LDI) lithography machine for chip packaging and high-density PCBs, a maskless process that trades some resolution for higher throughput. It arrives as TSMC's CoWoS advanced-packaging capacity, used to package AI accelerators like Nvidia's GPUs, remains constrained.
- [Waymo builds its own chip, cutting reliance on Nvidia](https://www.parallelquant.com/posts/waymo-builds-its-own-chip-cutting-reliance-on-nvidia-569234) (2026-08-21, The Decoder): Waymo has developed its own custom chip for its robotaxi fleet, reducing its dependence on Nvidia hardware. The move follows a broader industry pattern of AI-heavy companies building in-house silicon rather than relying solely on third-party GPU suppliers.
- [Nvidia H200 GPUs begin reaching China under new export licenses](https://www.parallelquant.com/posts/nvidia-h200-gpus-begin-reaching-china-under-new-export-licenses-a0fbee) (2026-08-21, Tom's Hardware): Nvidia's H200 AI GPUs are now entering China under case-by-case US export licenses, with each company's allowance reportedly capped around 100,000 units, most of which must stay outside mainland China. Domestically made Chinese AI chips have already gained significant ground in the local market despite the opening.
- [Supermicro fires staff after $2.5B AI chip smuggling probe](https://www.parallelquant.com/posts/supermicro-fires-staff-after-2-5b-ai-chip-smuggling-probe-b1ac46) (2026-08-21, Tom's Hardware): An independent investigation into Supermicro's alleged diversion of restricted AI chips to China found no wrongdoing by senior management, but the company fired employees in sales, technical support, and business development for policy violations. The probe covered roughly $2.5 billion in transactions and concluded Supermicro's financial statements remain reliable.
- [Nvidia to pay $6B for Poolside's AI model-building tech](https://www.parallelquant.com/posts/nvidia-to-pay-6b-for-poolside-s-ai-model-building-tech-93932d) (2026-08-21, The Decoder): Nvidia is acquiring Poolside's "Model Factory" software for building AI models, paying $6 billion. The deal brings 109 Poolside employees to Nvidia.
- [China develops its own circular AI financing scheme, report finds](https://www.parallelquant.com/posts/china-develops-its-own-circular-ai-financing-scheme-report-finds-eaff66) (2026-08-20, The Decoder): A Financial Times report found much of the demand for Unitree Robotics' humanoid robots comes from state-backed training centers that buy the machines and then sell the resulting training data back to Unitree. Unitree's shares rose 460% in its Shanghai IPO, valuing the company at around $50 billion.
- [China's SMIC posts record $3B quarter, hikes wafer prices](https://www.parallelquant.com/posts/china-s-smic-posts-record-3b-quarter-hikes-wafer-prices-68a86e) (2026-08-20, Tom's Hardware): Chinese foundry SMIC reported its first-ever $3 billion quarterly revenue, up 36.1% year-over-year, with net profit nearly tripling to $479.2 million. The company also raised wafer prices as US export sanctions push more domestic chip demand its way.
- [Cerebras unveils CS-4 system built on a 4-trillion-transistor chip](https://www.parallelquant.com/posts/cerebras-unveils-cs-4-system-built-on-a-4-trillion-transistor-chip-41d035) (2026-08-19, Data Center Dynamics): Cerebras announced its CS-4 rack-scale system powered by the WSE-3T wafer-scale chip, which the company says delivers 750 petaflops of AI compute. It extends Cerebras's wafer-scale approach as an alternative to GPU clusters.
- [Ukraine says it found Nvidia Jetson hardware in a Russian missile](https://www.parallelquant.com/posts/ukraine-says-it-found-nvidia-jetson-hardware-in-a-russian-missile-050a8a) (2026-08-19, Data Center Dynamics): Ukrainian officials say components from Nvidia's Jetson edge-AI hardware were recovered from a Russian S-71M missile. They are calling for stronger sanctions enforcement to keep such parts from reaching Russia.
- [First Nvidia H200 chips reach China as Beijing eases import block](https://www.parallelquant.com/posts/first-nvidia-h200-chips-reach-china-as-beijing-eases-import-block-dfba63) (2026-08-19, Tom's Hardware): ByteDance and Tencent have received their first shipments of Nvidia H200 chips as Beijing loosens its prior block on importing US-licensed AI chips. However, Beijing reportedly wants most of each company's licensed allowance, up to 100,000 units apiece, kept in Hong Kong, which lacks the power infrastructure to run them.
- [Nvidia leads $500B plan to make compute an asset class](https://www.parallelquant.com/posts/nvidia-leads-500b-plan-to-make-compute-an-asset-class-6a3e0d) (2026-08-19, The Verge): Nvidia is working with Apollo, BlackRock, Blackstone, Brookfield, Goldman Sachs, and KKR to structure roughly $500 billion in financing that treats GPU compute as a tradeable, revenue-generating asset class. CEO Jensen Huang framed the chips as long-lived, fungible, productive assets rather than depreciating IT hardware.

---
Published by Parallel Quant — https://www.parallelquant.com