SIGNALHuang combined technical direction, customer relationships and capital coordination; choosing a finance, engineering or external operating leader will also determine how Nvidia distributes those responsibilities in the future.
Read the full issue →SIGNALDeepSeek is extending competition beyond model answers into application infrastructure; adoption will increasingly depend on the combined cost and convenience of its search, file and voice services.
SIGNALNative low-precision training puts practical operating costs at the center of open-weight deployment, while the hundreds-of-gigabytes memory requirement keeps much of the deployment advantage with cluster operators and cloud providers.
SIGNALBetter memory organization can reduce personal AI's operating costs: locating relevant experiences before processing evidence avoids repeatedly rereading life recordings and makes answers easier to trace.
SIGNALAviation AI can optimize routes, but the data and communications systems beneath it determine whether advice reaches users; its value depends on both scheduling performance and infrastructure resilience.
According to IT之家, Cursor said that after a two-week investigation it found a malicious, unauthorized copy of the Anthropic Claude Code extension in its VS Code extension marketplace on October 29. The company said most customers were unaffected, with at most 5,000 users impacted.
According to IT之家, Cloudflare said on Nov. 13 that it was open-sourcing Kitesurf, a browser built for AI agents. Kitesurf runs on Cloudflare Workers and can parse HTML, apply CSS, execute page JavaScript and render screenshots without relying on a traditional browser container.
According to IT之家 on November 14, Huawei has continued to reveal more specifications for the Kirin 9050 series τ chip, focusing on its new-generation Malu GPU and Kirin AI ISP. The company says the GPU uses a fourth-generation architecture, lifts performance by 40% and cuts power use by 45%.
According to IT之家, citing The Elec, Intel announced the K-AI Full-Stack Alliance on November 12 at its 2026 AI Summit in Seoul, bringing together 23 Korean companies. The group is led by Intel and spans hardware, software, vertical AI solutions, and services, with the goal of building a sovereign and open AI infrastructure ecosystem for Korea.
According to IT之家 on Nov. 14, Elon Musk responded on X to criticism over Cybercab’s $299 monthly Autonomy Subscription, saying the price should be closer to $115.
According to IT之家, a research team from Tsinghua University’s Department of Automation has unveiled ProtCell, a foundation model for single-cell proteomics, alongside the paper “ProtCell: a foundation model for single-cell proteomics.” The model was trained and evaluated on about 380 million cells from 98 public datasets.
According to IT之家, Google said Gemini’s Personal Intelligence now supports Google Wallet. After users grant permission, they can ask about personal finances, spending, and budgeting by typing @Google Wallet in chat.
IT之家 reports that Google is expanding its Antigravity multi-agent system with Teamwork, a set of Gemini-powered specialist agents aimed at validating hard mathematical proofs and improving software performance. Google says the system formalized the Weil conjectures, reproduced a proof related to Navier–Stokes finite-time blowup, and helped solve the Courtade–Kumar conjecture.
ITHome reported on November 14 that Ammaar Reshi, Google AI Studio’s product and design lead, posted a video on X on November 12 showing Windows 11 running on Apple’s first foldable phone, the iPhone Duo. He also open-sourced the WindowsiOS project on GitHub.
According to IT之家, Microsoft AI scientist John Langford, a co-founder of the AI Frontiers lab, has left Microsoft after more than 14 years. He said the key reason was that his research no longer had a clear path to practical application, and new ideas were not taken seriously enough.
According to IT之家 on Nov. 14, BYD said its assisted-driving vehicle fleet has exceeded 4.19 million units, while its Tianyan system now generates more than 260 million km of data each day.
In an interview with TechRadar on November 13, Apple hardware engineering vice president Matthew Costello said HomeView, Apple’s first smart-home hub, can dim to just 0.3 nits. He described it as the darkest product Apple has ever designed and said he was proud of that.
According to MacRumors, Apple has pushed tvOS 27.2.1 and HomePod Software 27.2.1, both identified by internal build 24K157. The report says second- and third-generation Apple TV 4K models are getting Siri AI support without new hardware.
IT之家 reported on November 14 that Perplexity has invited paid Max subscribers to try its Teammates AI agent feature in the Mac version of the Perplexity app. The feature had previously been available only on iOS and the web, and now extends to desktop workflows.
IT之家 reported on Nov. 14 that developers found quota-management logic in the chat.deepseek.com front end, including a message saying usage limits have been reached and users should try again after quotas recover. The code also appears to support a 5-minute polling cycle while the page is visible to sync model quotas, mute status and server-side copy.
IT Home reported that Georgi Gerganov announced on X that the open-source AI inference project llama.cpp and the underlying ggml engine have reached library version 1.0.0. The project says future major-version bumps will signal API or ABI breaking changes, helping developers judge upgrade risk.
According to Reuters, Amazon has an internal plan for a “World Product Graph” that would collect data on products sold anywhere online, including items not sold on Amazon. The goal is to keep shoppers and AI agents inside Amazon by making its product index more useful than rivals such as Google or Reddit.
In a case study published by AWS Machine Learning, GS Caltex describes a workflow that hands validated models from process engineers in SageMaker Canvas to ML engineers running SageMaker Pipelines. The company automated the transfer of preprocessing flows, model configuration and approval steps into production jobs.
According to IT之家, citing Fortune, OpenAI has dissolved its independent robotics team and moved the staff into other parts of the company. OpenAI confirmed the reorganization to the media, but said it is not leaving robotics and wants to speed up development.
Axios reports that U.K. neocloud startup Nscale disclosed in a regulatory filing that it will need additional capital to meet projected 2027 obligations, and warned there is no assurance it can raise the money on acceptable terms. The filing showed debt jumping from $137 million at the end of Q2 to $5.5 billion at the end of Q3.
AWS Machine Learning describes how OpenFold3-preview can be deployed on Amazon SageMaker JumpStart. OpenFold3, developed by the OpenFold Consortium and the AlQuraishi Lab at Columbia University, predicts 3D structures for proteins, DNA, RNA, small-molecule ligands, and their complexes.
According to the GitHub Copilot blog, GitHub will deprecate several models on Dec. 14, 2026 across Copilot Chat, inline edits, ask and agent modes, and code completions. The list includes Claude Fable 5, Claude Haiku 4.5, Gemini 3.8 Flash, Gemini 4 Argon, Grok 4.6 and Kimi K3.
According to TechCrunch, AI agent startup LangChain said on Friday that it raised a $300 million Series C at a $3 billion valuation, with CapitalG leading the round and GIC, DFJ Growth, IVP and Sapphire Ventures participating. The company’s valuation is more than double the $1.25 billion it reached about a year ago, when it raised $125 million led by IVP.
AWS Machine Learning described how AssemblyAI runs its speech-to-text inference pipeline on Amazon ECS, EC2 Spot Instances and Amazon SQS. The architecture uses a flexible fleet of NVIDIA GPUs with ECS capacity providers, mixing Spot and On-Demand capacity across multiple instance families and regions.
Investor identity can determine whether willing cross-border capital reaches frontier AI companies; underwriting compliance screening directly affects which institutions can obtain shares in an initial offering.
Huang combined technical direction, customer relationships and capital coordination; choosing a finance, engineering or external operating leader will also determine how Nvidia distributes those responsibilities in the future.
If general models handle more robot cognition, specialist teams' advantage will depend increasingly on real-world operating data, hardware adaptation and safe execution, shaping their relationship with foundation-model companies.
DeepSeek is extending competition beyond model answers into application infrastructure; adoption will increasingly depend on the combined cost and convenience of its search, file and voice services.
AI data centers have become AMD’s main growth engine, extending competition into rack-scale delivery and system deployment.
Find transaction amounts, reporting stages and coverage. Announcements, reports and talks stay distinct; download CSV or JSON.
Compare input, cache and output rates and estimate usage costs. Check the verification date, currency and conditions.
Follow model releases, products and enterprise adoption, with company statements and media reports linked to our coverage.
SIGNALWith board oversight and executive leadership separated, NVIDIA must turn founder-driven customer coordination and technical decisions into durable team processes; delivery continuity will be tested before a successor’s reputation.
SIGNALHuang's concentration of technical and commercial authority makes the board's decision a test of both Nvidia's succession arrangements and its ability to operate independently of its founder.
SIGNALThe offering lets Anthropic raise money from public investors and key compute suppliers while its voting structure gives the founders a way to retain control as the shareholder base expands.
SIGNALThe board’s practical test is whether it can keep chip road maps, systems deliveries, and major customer relationships moving during an independent review when those decisions so often converge on Huang.
SIGNALHuang's reputation is closely tied to Nvidia's technology decisions and customer relationships, making management trust an important issue for customers committing to long-term compute investment.
SIGNALBy pairing open models with a paid training platform, Thinking Machines can let customers move their models elsewhere while competing to keep their business through easier training and operation.