Station F Director Roxanne Varza told Bloomberg Technology that French founders remain broadly reassured following Jensen Huang’s departure from Nvidia. She said there is no sign Nvidia is withdrawing its recent investment commitments in France, although the immediate fallout appears more reputational than operational. Varza also said the upheaval could eventually create opportunities for Europe to develop more sovereign alternatives to Nvidia’s technology stack.
Callosum co-founder Dan Akarca told Bloomberg Technology that Nvidia will remain dominant in the near term but is likely to lose AI-chip market share as more specialized processors emerge. He also said Callosum’s software can route AI tasks to suitable processors, potentially lowering costs and energy use while improving performance; those benefits remain vendor claims.
A Reuters review of Chinese contracts, disclosures and other documents identified more than $3.2 billion in compute projects built or planned around Nvidia GPUs restricted by the United States. Reuters contrasted the findings with Nvidia’s statement that U.S. export controls had effectively shut it out of China’s AI data-centre market.
Wccftech, citing Reuters, reported that more than 150 procurement notices, investment plans, construction contracts and corporate disclosures indicate NVIDIA AI-computing projects in China worth over $3.2 billion. The projects reportedly involve H100, H200 and H800 processors and users including technology companies, universities, government-backed ventures and regional computing centers. Analysts cited by Wccftech said there is no confirmation that NVIDIA GPUs have reached the premises. A facility in Gansu was also described only as reportedly being repurposed to house thousands of processors.
Huxiu AI reported that Nvidia shares briefly fell 10% on November 12, with its market capitalization moving toward the $4 trillion level. The article title also claimed that $1 trillion was erased and that Jensen Huang resigned, but the supplied report provides no evidence for either claim.
NVIDIA released the Apache 2.0-licensed SkillEvaluator for checking AI agent skills through quality validation, semantic overlap detection, model-backed rubric evaluation, synthetic evaluation dataset generation and live agent comparisons. It offers CLI and CI integration, and its three tiers can run independently. Tier 1 performs validation and security scanning, Tier 2 identifies overlapping guidance, and Tier 3 compares agent task results with and without a skill using provider access, a supported agent and a running sandbox, with Docker as the default. The bundled text-analyzer example receives a static quality score of 86.2 out of 100. NVIDIA describes the project as experimental and community-supported, with best-effort support and no SLA or enterprise support entitlement.
36氪, citing 字母榜, reports that Lens Technology, Lingyi iTech, and Luxshare Precision are expanding their Apple-supply-chain manufacturing expertise into Nvidia AI servers. Lens Technology agreed to acquire about 63% of Zhejiang Yuanshi through PMG, which reportedly holds Nvidia RVL rack certification, while Lingyi iTech acquired Liminda for reported AVL and RVL liquid-cooling certifications; Luxshare supplies high-speed copper cabling for AI rack wiring. The article also frames Nvidia’s leadership transition under Jensen Huang’s resignation as a new risk for this increasingly crowded supply chain.
TechNews reported that Nvidia shareholder Leo KoGuan backed founder Jensen Huang and said he would make no further Nvidia investment without Huang. KoGuan, who said he had invested about US$1.3 billion in Nvidia, plans to redirect capital toward other super-intelligence companies while urging Huang to remain. The report also said Huang had stepped down as CEO under board pressure after a scandal, while KoGuan defended him as a major driver of super-intelligence chips.
NVIDIA announced that co-founder Jensen Huang resigned as president, CEO and director with immediate effect, according to TechNews. Colette Kress, the executive vice president and CFO, will serve as interim CEO while retaining responsibility for finance, and Harvey C. Jones was appointed the company’s first board chair; the board is considering internal and external permanent successors.
The ashhart/TensorFold project describes itself as a native LLM inference engine for Metal, CUDA and Vulkan. Its documentation says a Zig binary serves language models on Apple Silicon and NVIDIA GPUs by reading checkpoints, tokenizing requests and running Metal or CUDA kernels directly, without requiring Python or MLX. The supplied material contains no independent performance benchmarks, and the Vulkan implementation is not detailed beyond the project’s headline claim.
Lex Fridman publicly defended Nvidia founder Jensen Huang after Huang resigned as CEO amid a personal controversy, according to TechNews. Fridman called Nvidia’s decision a major mistake for America and humanity, compared it to Apple firing Steve Jobs, and said Huang remains humble and caring despite making mistakes.
NVIDIA released Alpamayo 2 Edge on November 10, an end-to-end autonomous-driving VLA that combines causal reasoning, adaptive inference, and a diffusion action expert. The approximately 2.45B-parameter VLM backbone and 0.2B-parameter action expert run on DRIVE AGX Thor, with NVIDIA reporting a 48.9% closed-loop driving gain after reinforcement learning and a 160–190 ms Pixel-to-Trajectory latency in its reference configuration. The performance figures are vendor-reported and primarily reflect simulation-based evaluation.
LMSYS Org launched the SGLang Consortium during the opening keynote of the inaugural SGLang Summit in San Francisco, with NVIDIA, AMD, Intel, Google, Meta, Cloudflare, Modal and RadixArk as founding members. Members will assign technical representatives and attend two roadmap meetings annually, contributing engineering expertise, deployment feedback, testing resources or joint development, while SGLang maintainers retain authority over technical decisions and code acceptance. The stated priorities include frontier-model support, reasoning and multimodal workloads, distributed inference and hardware optimization.
Nvidia’s leadership shake-up raises questions for Asian chipmakers, while analysts cited in the report expect limited disruption in the near term. The report does not identify the analysts or provide company-specific evidence, and its longer-term implications are not established.
Bloomberg News reported, citing documents it reviewed, that a Chinese company founded by Alice Huang procured smuggled Nvidia AI chips installed at a Shanghai data center with links to Megaspeed International Pte. A follow-up report said some servers were originally sold to Aperia Cloud Services, an official Nvidia Cloud Partner headquartered in Singapore, according to a person familiar with the matter.
TechNews reported that NVIDIA Korea's former representative Yu Ying-chun expects Jensen Huang's reported departure to have limited impact in the short term. He cited the CUDA ecosystem, the production status of the Vera Rubin AI platform, and contracts covering next year with Samsung Electronics and SK hynix as reasons for that view. He said the new leadership's strategy could introduce variables from 2028 onward, but this is a personal forecast rather than an independently verified assessment.
cnBeta reports that NVIDIA’s updated management page lists interim CEO Colette Kress first among corporate officers, while Jensen Huang appears only as a founder. Huang’s profile now says he founded NVIDIA in 1993 and previously served as president, CEO and a board member, following his immediate resignation from those roles. Kress remains executive vice president and CFO while the board considers internal and external candidates for a permanent CEO.
cnBeta reported that some retail investors planned to buy or add to Nvidia shares after Jensen Huang’s resignation, viewing the decline as a buying opportunity because of the company’s products and AI demand. Other investors said they would wait for management’s comments and the next CEO appointment, while one investor said he would no longer buy the stock, showing sharply divided sentiment.
CNBC Technology reports that Nvidia entered its longest weekly stock losing streak since 2012, with its market value falling by $1.7 trillion. The stock had been sliding since early October, and Jensen Huang’s resignation added to Nvidia’s market woes. The report does not provide the exact weekly closing dates or a breakdown of the market-value decline.
Nvidia closed down more than 4%, erasing nearly $182 billion in market capitalization. A report dated November 13, 2026, also said the company’s market capitalization fell below $4 trillion. The supplied reports do not identify the cause of the decline.
Bloomberg reporter Mark Gurman called speculation that Tim Cook could replace Jensen Huang as NVIDIA CEO “ridiculous.” He made the comment while forwarding Cook’s post promoting the MacBook Neo; the post did not mention the CEO position or a willingness to take it. The speculation followed President Donald Trump’s claim that Cook was the only suitable successor after Huang’s resignation, but Trump’s statement did not represent an NVIDIA board decision. As of the report, Cook had not publicly said he intended to become CEO, and neither Apple nor NVIDIA had announced him as a candidate. NVIDIA’s board was reportedly seeking a permanent successor, with Colette Kress serving as interim CEO.
The Information reports that Nvidia general counsel Tim Teter asked CEO Jensen Huang to confirm his reported intimate relationship with Alice Huang, whom the publication described as a suspected Chinese spy, and Huang said it was true. By Sunday evening, Huang had publicly apologized for the affair, and Nvidia’s board said it would investigate his conduct. Teter reportedly had to urge directors to act because the board, which includes executives who have known Huang for decades, had been reluctant. People familiar with the situation said Huang told senior Nvidia executives that the affair was his only infidelity.
A report citing US media said Jensen Huang described his final moments as Nvidia’s CEO and called himself a bastard. The supplied report does not identify the US outlet or establish the context of the alleged remark.
NVIDIA announced that co-founder Jensen Huang resigned as CEO and director after 33 years, naming CFO Colette Kress as interim CEO while beginning a search for a permanent successor. Four semiconductor supply-chain executives interviewed by Central News Agency said NVIDIA’s published Blackwell, Rubin and Feynman roadmaps, product cadence and 2027 orders were expected to remain unchanged, while ASIC development could accelerate. Suppliers also reported seeing some mass-production orders into 2028, though these figures were not independently disclosed by NVIDIA.
Nvidia’s turmoil is prompting a rotation across chip stocks. The shift comes as the company deals with the fallout from Huang’s resignation following revelations about multiple affairs. The available excerpts do not identify the stocks involved or provide details on the rotation’s timing, size or duration.
IT之家, citing The Wall Street Journal and people familiar with the matter, reports that Nvidia co-founder Jensen Huang publicly criticized then-CFO and now interim CEO Colette Kress during an August executive meeting. Huang favored using financing guarantees to support customers unable to buy chips directly, while Kress questioned project selection, collateral risk and returns; she later reduced Nvidia's proposed support for OpenAI's Ohio data center agreement. Huang also raised concerns about inventory pressure, customers considering alternative chips and financing commitments approaching unreasonable levels.
TechNews reports that Jensen Huang resigned as Nvidia’s president, CEO and director. CFO Colette Kress will serve as interim CEO while retaining her finance role, and Harvey C. Jones will become Nvidia’s first board chair. CTBC Financial Advisory said Huang’s departure may lead cloud service providers to reassess their reliance on Nvidia and accelerate existing efforts to diversify toward custom ASICs. It also said AMD could gain opportunities to compete for AI GPU orders. Nomura Asset Management described the transition as a confidence test for the AI industry, while saying any shift toward a downturn would depend on capital spending and order changes; confidence pressure does not necessarily mean a demand collapse.
Nvidia appointed longtime CFO Colette Kress as interim CEO after Jensen Huang resigned, while the board searches for a permanent successor. Reports citing The Wall Street Journal and people familiar with the negotiations say Kress challenged Nvidia's expanded chip-financing guarantees and renegotiated support for OpenAI's Ohio data-center project. The initial proposal could have included up to $250 billion in financing support, while later reports said Nvidia committed about $100 billion to the first portion of a split arrangement.
Nvidia’s board announced that Jensen Huang resigned as president, CEO and a board member with immediate effect. Executive vice president and CFO Colette Kress will serve as interim CEO while remaining in her CFO role until a permanent successor is selected, and Nvidia updated its leadership pages to reflect the changes. TMTPost, citing Reuters and the Associated Press, reported that Harvey Jones became the company’s first board chair and that Nvidia shares fell more than 10% in premarket trading after the news.
Bloomberg Technology, citing reporter Ian King, says Jensen Huang’s personal affairs—including one involving an alleged Chinese spy—left him no choice but to resign as Nvidia’s CEO. King said Huang tried to repair the damage and accept responsibility during his final days, but that his efforts were insufficient. IT之家 AI also reported that Huang appeared exhausted and expressed remorse, based on unnamed sources.
What problem it solves
AI agents need to choose among models while balancing task success, cost, and latency for their workloads. Applications also need flexibility in where routing runs and who controls model clients, credentials, retries, and service operation.
Key features
Multiple routing algorithms — Routing can be configured with Auto, Task, Execution, or Composite strategies, including mode…
What problem it solves
Serving large language models requires efficient memory management, flexible model execution, and practical integration with existing models and APIs. vLLM addresses these needs with inference and serving capabilities spanning batching, caching, quantization, distributed execution, and multiple hardware platforms.
Key features
Efficient memory and serving — vLLM manages atte…
What problem it solves
AI agents need product-specific instructions to use NVIDIA software reliably across physical AI, robotics, simulation, CUDA-X libraries, RAG, and platform workflows. The catalog centralizes these skills while keeping them maintained in their respective product repositories and synchronized through an automated pipeline.
Key features
Agent-discoverable skill catalog — The NVI…
What problem it solves
AI agent skills need to be checked for quality, security, duplication, and their actual effect on agent behavior. Static checks alone cannot establish whether guidance improves task results, so evaluation must also support live comparison with and without a skill.
Key features
Three-tier evaluation pipeline — Independent tiers validate skill safety and structure, detect over…
What problem it solves
Running large language and vision-language models can require substantial setup and may depend on the available hardware. The project aims to make inference accessible across a wide range of local and cloud hardware with minimal setup.
Key features
Minimal-setup inference — The project provides installation options for common platforms, including scripts, Docker, pre-built b…
What problem it solves
Serving language models requires reading checkpoints, tokenizing requests, and running accelerated kernels while managing a local model server. TensorFold addresses these requirements with a native binary that can run on Apple Silicon and NVIDIA GPUs without requiring Python or MLX for serving.
Key features
Native Metal and CUDA inference — The Zig-based engine executes mode…
What problem it solves
Interactive multimodal and speech systems often require generation to span heterogeneous stages with different compute patterns, dependency structures, and resource needs. Serving these pipelines requires coordinated stage lifecycles, inter-stage transport, model integrations, and a unified API surface.
Key features
Coordinated multi-stage runtime — Models generation as coor…
What problem it solves
Data analysis workflows often require expressive query composition, flexible execution models, and interoperability with columnar data. Polars addresses these needs by allowing the DataFrame API and SQL to share the same optimizer and engine, with support for large workloads and custom extensions.
Key features
Unified DataFrame and SQL Queries — Queries can be written with e…
What problem it solves
Running frontier-scale open-weight models locally requires an inference platform that can make efficient use of the heterogeneous hardware available on a personal device. The project addresses this by treating GPUs, CPUs, host memory, and interconnects as a unified, elastic inference platform.
Key features
Edge-native MoE runtime — The runtime provides MoE serving with bandw…
What problem it solves
Autonomous AI agents often need to read files, install packages, call APIs, and use credentials, but unrestricted access can expose host data, secrets, and networks. OpenShell provides these capabilities within an isolated runtime where declared policies constrain agent access.
Key features
Kernel-level runtime enforcement — Agents run in isolated sandboxes where kernel cont…
What problem it solves
AI agent skills execute with implicit trust and often receive minimal vetting before installation. SkillSpector helps users assess whether a skill contains vulnerabilities, malicious patterns, or other security risks before installing it.
Key features
Multi-format scanning — Scan agent skills supplied as Git repositories, URLs, zip files, directories, or individual files.
So…
What problem it solves
Mosaics obscure portions of videos and images, making it difficult to view the affected content without changing the rest of the picture. Users need a restoration workflow that can identify selected mosaic regions and process them with GPU-accelerated video handling.
Key features
Targeted mosaic restoration — Detects mosaics in videos and restores them while leaving the rest…
What problem it solves
Deep learning models often need to be adapted for efficient deployment across different inference frameworks. Model Optimizer addresses this by providing a unified set of optimization techniques and export workflows for quantized checkpoints.
Key features
Post-training quantization — Compress models with quantization techniques available for Hugging Face, Megatron-Bridge, Di…
What problem it solves
Platform teams need to create and operate isolated Kubernetes clusters for customers on infrastructure they supply and control. They need a management plane that can provision GPU machines, manage cluster lifecycle operations, and establish node identity before issuing join credentials.
Key features
Independent workload clusters — Create workload clusters through the NVKE AP…
What problem it solves
Running AI models locally can require integrating multiple inference engines, model formats, hardware accelerators, and API surfaces. LocalAI addresses this with a small core, on-demand backends, and compatible APIs that work across supported hardware without requiring a GPU.
Key features
Composable inference backends — Backends are packaged separately and pulled only when a…
What problem it solves
LLM and Visual Gen inference requires specialized kernels, efficient runtime orchestration, and flexible deployment across NVIDIA GPU configurations. Developers also need a modular framework that supports customization while integrating with established inference tooling.
Key features
Optimized inference kernels — Custom kernels accelerate common inference operations, includ…
What problem it solves
Rust developers need host-side access to the CUDA runtime and tools for expressing GPU kernels while preserving Rust's ownership and borrowing rules. The platform provides separate programming models for workloads that require direct thread-level control and workloads naturally expressed through tensor operations.
Key features
Shared CUDA runtime — Host-side crates provide d…
What problem it solves
Agents can retrieve documents and query databases, but unreliable results undermine agentic analytics. Teams need a context layer that records where data lives, what its definitions mean, whether it is usable, how it should be queried, its provenance, and related concepts.
Key features
Enterprise ontology generation — Generate semantic ontologies from database metadata, infe…