TMTPost reports that NVIDIA announced on November 12 that Jensen Huang had stepped down as president, CEO and a board member, with immediate effect. CFO Colette Kress became interim CEO while the board began a search for a permanent CEO. The commentary says NVIDIA has no established second-in-command, with roughly 60 executives reporting directly to the CEO and no publicly disclosed succession plan. It uses the transition to compare key-person succession at technology companies with the challenges facing venture-capital firms, while noting that the permanent CEO, the eventual succession structure and the independent investigation’s outcome were not established in the report.
Nvidia's stock continued to slide on Monday following CEO Jensen Huang's resignation. The decline reflected investors pricing in the fallout from the leadership change.
VicOne says AI-assisted research identified an initial set of 28 zero-day vulnerabilities in less than two days across software supporting robotics development, simulation, deployment, and operation. The set includes projects developed or maintained by Google, NVIDIA, and Meta, but the count may change after vendor and vulnerability-authority review, and some findings lack CVE numbers. In a separate demonstration, VicOne researchers exploited a vulnerability that prevented a robot’s remote-stop command from working.
Shares of Chinese AI firms MiniMax and Z.ai plunged on Monday amid fears that the U.S. will tighten Chinese companies’ access to export-controlled chips following Nvidia CEO Jensen Huang’s reported resignation. In Hong Kong trading, MiniMax fell 18% and Z.ai fell 13%.
IMF Managing Director Kristalina Georgieva told CNBC at the ASEAN summit in the Philippines that the global economy is "standing on one leg" — AI — and that a sudden drop in investor confidence and capital outflow from AI could trigger financial stability problems. The IMF estimates circular financing among AI companies has grown to just over $1 trillion, about two-thirds of it debt, where investor money returns to investors through procurement orders; Nvidia is identified as a core participant, and UBS analyst Tim Arcuri estimated Nvidia pulling back could remove at least $50 billion in funding. Georgieva called recent market moves "orderly" and said some correction in AI valuations is necessary, and noted Nvidia's earnings and Anthropic's IPO are due this week.
Cassava Technologies and Graphiant reportedly plan to offer quantum-safe networks across Africa, beginning with South African customers in the first half of 2027. The post-quantum service would encrypt data, identify the countries it traverses and restrict where AI agents can send data, according to Liquid C2 chief commercial officer Vinay Hiralall. Neither company has publicly confirmed the plan, and The Next Web did not independently verify the timeline or capabilities.
TMTPost reported that Anthropic announced the acquisition of one-employee biotechnology company Mithril Medicine, which studies treatments for two rare muscular dystrophies. OpenAI also announced the acquisition of Orchestra Bio and plans to integrate its drug-development workflow coordination technology into Rosalind. Other initiatives include Isomorphic Labs’ drug pipeline, Nvidia and Eli Lilly’s planned AI laboratory, Amazon and Novo Nordisk’s strategic partnership, and Microsoft Discovery. Mithril Medicine’s reported mouse results are preclinical, and its potential entry into human trials by the end of 2028 remains conditional on further research.
Technology stocks plunged ahead of Nvidia's earnings and an expected Anthropic listing, Bloomberg reported. The two upcoming events were described as likely to influence the near-term course of an increasingly fragile AI trade.
Analyst Ming-Chi Kuo said OpenAI and Anthropic orders have extended Google TPU shipment visibility through 2029. He attributes v8-series orders, including 8t and 8i variants, to the two customers and expects v9 to become Google’s main option for its internal compute from 2028. Kuo estimated OpenAI would need about 12 million MediaTek-designed 8t TPUs and Anthropic about 14 million Broadcom-designed 8i TPUs in 2028–2029. Accounting for supply-chain constraints, he projected Google would ship 16 million to 18 million TPUs in 2029.
Taiwan’s weighted index closed at 41,040.08, down 2,635.04 points, or 6.03%—its second-largest point decline on record. TechNews linked the broad decline in major stocks to concerns about AI’s outlook following Jensen Huang’s resignation as NVIDIA’s chief executive. TSMC fell 5.74%, MediaTek 7.89% and Foxconn 9.34%. Barclays said forced selling by retail investors was a main driver of the steep selloff.
IT之家 reported that HSBC analyst Frank Lee downgraded Nvidia to Hold and cut its target price from $365 to $170, a reduction of about 54%. Lee cited uncertainty over the CEO succession plan following Jensen Huang’s resignation and the risk that stricter chip export rules could affect overseas revenue. The report said Nvidia’s fiscal 2027 second-quarter revenue was $96.22 billion, up 105.8% year over year, and that management said customer demand exceeded its supply chain’s capacity.
NVIDIA's board announced on November 12 that Jensen Huang resigned as president, CEO and director, with CFO Colette Kress serving as interim CEO while retaining her finance role. Fubon Investment & Trust said Taiwan suppliers may face near-term pressure from disrupted China sales and uncertainty about NVIDIA's new leadership, while also emphasizing the industry's advanced-process and packaging advantages. The firm also projected tighter U.S. controls on lower-end AI compute chips sold to China, which could impede NVIDIA's China-specific Hopper chip, downgraded RTX 4090 and H200 export plans.
TechNews reported that Nvidia CEO Jensen Huang resigned, prompting concern about Taiwan’s AI supply chain. Central Bank Governor Thomas Yang told lawmakers that the matter was a “short-term disruption,” saying Taiwan plays an important role not only in Nvidia’s supply chain but also in those of AMD and Google. Yang said the situation remains unsettled until Nvidia selects a permanent CEO and its operating strategy becomes clearer. He also said the central bank has relevant data and is closely monitoring AI financing concentration.
36Kr, citing Sina Finance, reports that Nvidia’s new financial leadership is expected to further tighten financial support for AI startups that have depended on the company. The report says Huang began adjusting Nvidia’s approach in the final months before stepping down as CEO. At a recent Nvidia AI conference in Berlin, Huang said the era of Nvidia acting as an “unlimited ATM” had ended. The report also notes that several emerging cloud companies buy Nvidia chips and rent computing capacity to customers.
InfoQ reports that AWS will host the “What If Lab” AI hardware hackathon in Shanghai on November 21, sponsored by NVIDIA and supported by Tuya Smart as technical partner. The event is recruiting teams of one to three developers by November 18, with 20 teams selected to demonstrate functioning AI-powered hardware on site. AWS, Tuya Smart and NVIDIA say they will provide cloud, open-source device connectivity, inference services and models for participants.
Wang Qiangyu has become CEO of Baiwang as the company works to convert its tax and finance experience into professional models and agent products. Baiwang has launched an AI tax-and-finance expert team for invoicing, bookkeeping and tax filing, using natural-language interaction and standardized task execution while retaining rule validation and human confirmation, according to Lei Feng Wang AI. The described capabilities and business benefits remain vendor claims without independent performance measurements or customer results.
TMTPost reports that Sereact, Zalando and CEVA deployed dual-arm robots in October at two return-processing sites in Germany and Poland. The robots directly grasp, identify and sort returned apparel, while employees handle supervision, exceptions and quality control. The companies did not disclose the number of robots, daily throughput, task success rate, human-intervention rate or per-item processing cost.
Calls for Jensen Huang to return as Nvidia CEO are growing, with technology executives and major shareholders backing a potential comeback following his resignation.
TechNews reports that Jensen Huang’s resignation as NVIDIA’s chief executive has renewed scrutiny of the company’s overseas headquarters project in Taipei’s Beitou-Shilin technology district. Property-market analyst Li Tong-rong said the project’s land-rights contract and permit work remain on schedule, but advised monitoring NVIDIA’s Taiwan investment policy, construction plans and surrounding supply-chain development. He expects the district’s property market to remain at high levels with lower transaction volume and little immediate price movement over the next six months.
TMTPost reported that NVIDIA announced Jensen Huang’s immediate resignation as president, CEO and director on November 12, 2026. The board appointed CFO Colette Kress as interim CEO while retaining her CFO role, named Harvey C. Jones as its first board chair and began searching for a permanent CEO. TMTPost cited NVIDIA’s 2026 annual report for data-center revenue of about $194 billion and nearly 9 gigawatts of deployed Blackwell AI-factory capacity; those figures were not independently verified here.
Three former Fortune 500 CEOs and a consultant told 36Kr that NVIDIA had not prepared a mature succession plan or a ready successor for Jensen Huang. They argued that the gap, noted as Huang approaches his 64th birthday in February, exposes a significant weakness in the company’s governance.
More than 2,000 people attended the two-day AI Day Seoul at COEX, according to NVIDIA Blog, where industry representatives, startups and developer communities discussed physical AI and AI agents. The event included a Hangul Day hackathon whose five finalists built Korean-language agents with the OpenShell runtime, plus developer workshops, certification exams, startup pitches and an AI for Science forum. Sessions were expected to become available soon on NVIDIA On-Demand.
Together AI introduced Together Compiler, an agent-driven framework for continual AI inference optimization that replaces hand-written megakernels with composable kernels, CUDA Graphs, PDL, and parallel agent workflows. The company said its internal experiments recovered 98% of megakernel performance with default Torch Inductor kernels plus CUDA Graphs and PDL, while later configurations achieved up to 1.7 times the throughput of its megakernel baseline. Together AI is offering early access and says the system produced 1,111 TPS on MiniMax M3.1-Flash and 1,205 TPS on DeepSeek v4.1-Flash.
A commentary reports that Nvidia’s board announced co-founder Jensen Huang’s resignation. It argues that Huang drove three major transformations: launching CUDA in 2006, directing Nvidia toward deep learning after researchers used its chips to win the 2012 ImageNet contest, and expanding the company from GPU boards to integrated AI-datacenter racks through the 2020 Mellanox acquisition. The commentary says the next CEO must execute the existing business while continuing to reinvent it, warning that an execution-focused leader may reduce investment in innovation as Nvidia’s major customers develop competing AI chips. The report does not identify a permanent successor or establish how Nvidia’s strategy will change.
36氪, citing Securities Times, says the controversy surrounding NVIDIA founder Jensen Huang’s departure has weakened the company’s perceived certainty premium rather than long-term global AI demand. Funds and related institutions described by the report see the shift as encouraging greater diversification in AI compute and creating a possible revaluation opportunity for domestic alternatives. These institutional assessments are summarized without underlying statements or supporting market data.
Nvidia was scheduled to report fiscal third-quarter results after Tuesday’s market close, covering a quarter that ended in late October. Bloomberg Technology said interim CEO Colette Kress was preparing to lead her first earnings appearance since Jensen Huang’s exit, with investors watching her comments for clues about the company’s strategy. The Information also reported that Huang had agreed to testify before Congress, potentially that week.
TechSpot reports that the Stand with Jensen campaign has collected more than 150,000 signatures urging Nvidia’s board to reinstate co-founder Jensen Huang as CEO during an independent investigation. Supporters include Elon Musk, Tim Sweeney, Amjad Masad and Garry Tan, along with self-identified Nvidia employees and shareholders. Organizers seek a fair hearing and a public explanation for Huang’s departure, while saying the campaign is not endorsed by Nvidia or Huang.
HP introduced the ZGX Fury G1n workstation with Nvidia's GB300 Grace Blackwell Ultra processor, combining 496GB of LPDDR5X system memory and 252GB of HBM3e graphics memory. The standard configuration starts at $132,000, while selected graphics options can raise available video memory to 276GB. The system ships with Ubuntu 24.04 LTS, liquid cooling, a 1,600W power supply and 400Gbps networking.
A Reddit poster claims OpenAI generated about $13 billion in the previous year and committed roughly $1.15 trillion to compute. The post describes a 2026 funding round involving $50 billion from Amazon, $30 billion from Nvidia and $30 billion from SoftBank, saying it closed at $122 billion and a $852 billion post-money valuation. It also says OpenAI committed to Nvidia Vera Rubin systems for 3 GW of inference and 2 GW of training, alongside about $300 billion in commitments for Oracle data centers running Nvidia chips. These figures are presented as the poster’s claims rather than independently confirmed findings.
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…