Daily · Nvidia Leadership Crisis and AI Market Shifts · November 13, 2026
Key points
- Jensen Huang resigned as Nvidia CEO on November 12, triggering an 11% weekly stock drop and a leadership vacuum at the world's leading AI chip supplier.
- A global Microsoft 365 outage disrupted Outlook, Teams, and Copilot in 74 countries for over 22 hours, exposing critical infrastructure dependencies.
- Anthropic is scheduled to price its IPO on November 18 at a $1.26–$1.41 trillion valuation, the largest AI company listing to date.
- Google launched Gemini 4 Hydrogen, a 1-million-token agentic model priced at $1.50 per million input tokens, challenging existing cost-performance benchmarks.
Nvidia Leadership Vacuum and Export Control Scrutiny
Jensen Huang resigned as Nvidia's President, CEO, and board member on November 12, ending his 33-year tenure. The departure followed revelations of an affair with Alice Huang, a former executive at Singapore-based Megaspeed, whom US intelligence agencies suspect is a spy for China's Ministry of State Security. A second reported relationship with Nvidia Vice President Kimberly Powell, a direct report, accelerated the board's decision to remove him. CFO Colette Kress has been named interim CEO while the company searches for a permanent successor.
The scandal has intensified scrutiny over US export controls. The FBI and Justice Department are investigating whether Megaspeed diverted advanced Nvidia chips to China in breach of restrictions. US Treasury Secretary Scott Bessent stated that multiple agencies favored placing Megaspeed on the entity list for months, but Commerce Secretary Howard Lutnick delayed the process. Lawmakers from both parties have called for testimony from Huang and Lutnick, with Senate Banking Chair Tim Scott requesting a public hearing on December 3.
Market reaction was immediate. Nvidia shares fell 11% this week, dropping below $4 trillion in market value for the first time since July 2025. Rivals AMD and Broadcom gained 30% and 24% respectively, while TSMC rallied 11%. Analysts warn that Huang's departure may weaken Nvidia's ability to secure supply chain capacity and negotiate multi-party investments, with Goldman Sachs noting the company has $36 billion in off-balance-sheet AI cloud agreements. The company is scheduled to report Q3 earnings on November 17, its first update since the leadership change.
Global Microsoft Outage Exposes Infrastructure Fragility
A major outage affecting Microsoft 365 services began on Thursday, November 12, and continued into Friday, impacting Outlook, Teams, SharePoint, OneDrive, and Copilot in at least 74 countries. Downdetector recorded nearly 140,000 reports of Outlook issues at peak. Microsoft attributed the failure to a number of servers becoming unexpectedly unavailable, which strained remaining infrastructure, and ruled out a cyberattack.
The disruption affected government agencies, healthcare systems, and financial institutions. In the UK, police forces and NHS practices reported delays in processing online forms and prescriptions. In Japan, government employees resorted to paper documents. AI companies OpenAI and Anthropic reported issues with their Microsoft 365 connectors, affecting tools like ChatGPT and Claude. Electric truck maker Workhorse Group postponed its earnings call to November 16 due to business disruptions.
The incident has renewed calls for digital sovereignty in Europe. Mistral CEO Arthur Mensch stated that only 20% of European digital services run on European infrastructure, arguing that spending on foreign technology should be reinvested in local capabilities. The outage highlighted the risks of single-vendor dependency in critical public and private sector operations.
Anthropic IPO and AI Market Valuations
Anthropic is scheduled to price its initial public offering on Nasdaq on November 18, targeting a valuation between $1.26 trillion and $1.41 trillion. The company aims to raise approximately $64.4 billion, backed by cornerstone investors including Alphabet, Nvidia, and Microsoft. Goldman Sachs, the underwriter, projects Anthropic will achieve annual revenue of roughly $685 billion by 2030.
The IPO comes amid market volatility following the Nvidia leadership crisis and concerns over an AI bubble. The valuation is lower than the $2 trillion mark Anthropic previously targeted. The listing will serve as a key indicator of investor appetite for AI infrastructure spending, with potential knock-on effects for downstream AI developers and data center investors in Europe and the Gulf.
New AI Models and Developer Tools
Google released Gemini 4 Hydrogen on November 12, a fast agentic model with a 1-million-token context window. Priced at $1.50 per million input tokens and $7.50 per million output tokens, Hydrogen is positioned as a cost-effective alternative to higher-tier models. It scores 51 on the Artificial Analysis Intelligence Index at its default setting, surpassing Gemini 4 Argon's 50. The model is available across Google's developer, enterprise, and consumer surfaces, including a new Orchestrator feature that dynamically switches between models within a session.
Apple released the Foundation Models framework, a native Swift API for integrating any large language model into apps. The framework supports on-device Apple Foundation Models and server models like Claude and Gemini, with full tool-calling support. Developers in the App Store Small Business Program can access Apple's cloud models at no cost if they have fewer than 2 million first-time downloads.
IBM released two experimental preview models based on Granite 4.1, trained to use LLMON, a markup language designed to mitigate prompt injection attacks. The models avoid 47% more attacks relative to the best frontier model tested. Additionally, Hugging Face's llama.cpp team released version 1.0.0, establishing stable API and ABI versioning for the local AI inference ecosystem.
Security Threats and Regulatory Actions
The FBI attributed the $6.3 billion hack of Binance to North Korean group TraderTraitor. Binance has committed over $6.34 billion in company funds to reimburse users. The US Justice Department obtained court authorization to seize more than $1.2 billion in cryptocurrency linked to the theft. Separately, the FBI issued an advisory on a North Korean hacking cluster called 'Sapphire Sleet' that used AI deepfakes and compromised video calls to steal over $400 million from 900 victims.
CrowdStrike reported that agentic AI attacks are compressing security response times, with average breakout times dropping to 29 minutes. The company profiled four attack campaigns where AI agents, including Claude Code, were used to switch tactics and access credentials within minutes. CISA issued Binding Operational Directive 26-04, giving federal agencies three days to fix the most dangerous flaws, a significant reduction from previous two-week deadlines.
In other regulatory news, the EU's first dedicated AI summit in Dublin saw leaders discuss coordination on financing and incident response. Meanwhile, the UK civil service plans to increase in-person job interviews to counter a surge in AI-generated applications.
Our read
The resignation of Jensen Huang marks a pivotal moment for the AI infrastructure sector, shifting the narrative from hyper-growth to governance and security scrutiny. For decision-makers, the immediate risk is not just leadership instability at Nvidia but the potential tightening of export controls that could disrupt global supply chains. The Microsoft outage serves as a stark reminder of the fragility of centralized cloud dependencies, likely accelerating European efforts toward digital sovereignty and multi-vendor strategies. As Anthropic approaches its record-breaking IPO, investors should weigh the company's ambitious revenue projections against the broader market skepticism surrounding AI valuations. Finally, the emergence of agentic AI in cyberattacks demands that security teams update their detection and response protocols, as traditional breakout-time metrics are becoming obsolete.
This material was produced automatically by a large-language-model system from the public sources listed below; it is AI-generated content and may contain inaccuracies — verify facts against the original sources.
Sources
- Nvidia shares plunged 11% this week following Jensen Huang's resignation amid revelations about his affairs including with a woman believed by the US intelligence community to be a Chinese spy, while rivals AMD and Broadcom gained 30% and 24% respectively — cnbc.com, asia.nikkei.com, wired.com (+34)
- A global Microsoft 365 outage affecting Outlook, Teams and Copilot began Thursday and entered its second day, disrupting services in at least 74 countries — en.mercopress.com, bbc.co.uk, aljazeera.com (+24)
- Google introduced Gemini 4 Hydrogen, a fast agentic model with 1 million-token context, available globally across its developer, enterprise, and consumer surfaces — venturebeat.com, arstechnica.com, antigravity.google (+3)
- FBI attributed the $6.3 billion Binance hack to North Korean group TraderTraitor; Binance promised to reimburse users from its own funds — en.yna.co.kr, rbc.ru (+1)
- Anthropic is scheduled to price its initial public offering on Nasdaq on November 18, targeting a valuation between $1.26 trillion and $1.41 trillion — dw.com, france24.com
- Mistral CEO Arthur Mensch and Nobel laureate Geoffrey Hinton urge EU leaders to act now on AI strategy at Dublin summit — scmp.com, politico.eu, france24.com (+1)
- FBI issued an advisory on a North Korean hacking cluster called 'Sapphire Sleet' (UNC1069) that stole more than $400 million using a social engineering scheme dubbed 'FaceStealer' involving compromised video calls and AI deepfakes — dailymaverick.co.za, bbc.co.uk
- Will DePue of Unemployment Labs published a forensic analysis finding that six Chinese frontier AI models carry initialization fingerprints of export-controlled NVIDIA Blackwell GPUs — unemploymentlabs.com, scmp.com
- CrowdStrike published a report profiling four previously unreported agentic AI attack campaigns traced between May and August 2026, naming Claude Code in all four cases — venturebeat.com
- Cerebras Systems reported Q3 core revenue of $230.9 million, beating LSEG consensus, and announced a new deal with Microsoft — cnbc.com
- PostgreSQL released security fixes for CVE-2026-6460, a 22-year-old composite-type confusion vulnerability enabling remote code execution — mehmetince.net
- Bloomberg reported that SpaceX and Tesla may announce a merger in December, which would become the largest deal in history — tass.com
- Court-appointed administrators of Dutch chipmaker Nexperia announced they would sound out investors on a majority or minority stake after Wingtech rejected a proposal to hand control of non-China operations to foreign investors — scmp.com
- Apple releases Foundation Models framework with Swift API, Python SDK, Dynamic Profiles, and Evaluations for integrating any LLM into apps — developer.apple.com
- Nvidia confirmed its commitment to deploying up to 260,000 GPUs in South Korea in partnership with Samsung, SK Group, Hyundai Motor Group, and Naver Cloud, despite Jensen Huang's resignation — en.yna.co.kr
- OpenAI announced ChatGPT Wallet, a payment feature enabling its 'dots' AI agents to make purchases on behalf of users in partnership with American Express, Mastercard and Visa — cnbc.com
- The Bank of England, Monetary Authority of Singapore, and Bank of Thailand published findings from Project Meridian FX APAC, demonstrating that synchronised payment-versus-payment FX settlement can work across central bank systems using different technologies including RTGS and DLT-based platforms — bankofengland.co.uk
- The Global Initiative for Information Integrity on Climate Change launched the Priorities for Action on AI and Climate Information Integrity at COP31, calling on AI companies to assess and disclose their impact on climate information and environmental footprints — news.un.org, scmp.com, arstechnica.com (+5)
- South Korean government says Google has complied with security requirements for exporting high-precision map data — en.yna.co.kr (+1)
- IBM releases LLMON, an LLM-native markup language, with two experimental preview models based on Granite 4.1 3B and 8B — huggingface.co
- New research published in Science finds the cryosphere has fallen outside its 'safe operating space' due to fossil-fuel-driven warming — wired.com
- Attackers used two zero-days to reach root on the Zammad ticketing system of the Dutch Institute for Vulnerability Disclosure on September 21, after which AI agents ran the rest of the attack — venturebeat.com
- The Communication Workers Union and Unite the Union filed an application to the UK's Central Arbitration Committee asking it to confirm sufficient worker support to trigger a unionization ballot at Google DeepMind's UK operations — wired.com
- The US Federal Aviation Administration expanded its $875 million SMART AI system to manage air traffic congestion across all 22 air traffic control centers nationwide — arstechnica.com
- Nintendo sent four DMCA takedown notices to GitHub targeting Switch emulators Yuzu and Citron Neo and 144 mod chip firmware repositories — ign.com