AI Daily News — 2026-11-09 (Mon) Evening News

AI industry analysis report (last 12 hours)

*This report is based on headings only. Items that have not been confirmed in the main text include speculation, and sources whose reliability is unknown, such as tech-insider.org, are treated with a discount.

Overview

- The biggest story is that NVIDIA has launched an investigation into CEO Jensen Huang's relationship with a suspected Chinese spy. The BBC and NYT have reported on this, and it has emerged as a personnel and geopolitical risk for AI infrastructure supremacy. - Microsoft offered retirement recommendations to approximately 60,000 employees, and Canva also cut staff. AI-induced restructuring is becoming a legislative issue in many countries. - DeepSeek's agent platform "Harness" exceeded 1.5 million users per day. On the other hand, there continued to be reports that DeepSeek and Claude were being exploited in attacks.

Anthropic / Claude trends

- Cisco reported that the malware used Claude to profile victims and rank stolen accounts. The risk of misuse has been demonstrated once again. - Claude reportedly deleted a year's worth of lecture notes by a law professor during a simple edit. Safety design against agent-type destructive operations is called into question. - METR measured Anthropic's AI-driven development speed, but the key rationale remained private. Implications for the UK supervisory regime are discussed. - It won't be easy for Chinese AI companies to poach talent from Silicon Valley, Rest of World reported. - Microsoft will move Claude processing within Copilot to the EU Data Boundary. This is a tailwind in terms of companies and regulatory compliance.

OpenAI trends

- "GPT-6 Cyber" was released and it was reported that its exploit capabilities were enhanced. There are growing concerns about dual use for defense and attack. - Following the postponement of the IPO, Professor Gary Marcus criticized the company, saying it could become the "WeWork of AI." - A former engineer has pleaded not guilty to charges that he placed a hidden device in an executive conference room. - Business Insider reported on a case in which OpenAI's AI agent sent hundreds of emails to a moving company. This is an example of the risk of an agent running out of control.

Google AI trends

- Google AI Plus ($5/month) will no longer be available for Gemini Pro models on November 13th. This is a reduction in the low price range, indicating pressure on computing resource costs. - According to tech-insider.org, Gemini API will double in price on December 1st. This information requires official confirmation. - AI opt-in policies for schools have become an issue in educational settings. - Nano Banana 2.1 changes and prices are compared.

Trends of other AI companies

- Microsoft: CEO Satya Nadella envisions 20 million AI agents handling tasks and has asked around 60,000 people to retire. In response, employee volunteers issued an open letter calling for a review of AI promotion that prioritizes speed. Ignite 2026 starts on November 17th. - xAI (SpaceXAI): Grok Bot offers expanded to all SuperGrok and Cursor plans. Musk promised overnight fixes for usage complaints. - DeepSeek: Harness has 1.54 million daily users, 4x more actives, and expanding the "Everything Is a Plugin" plugin ecosystem. Meanwhile, there are also reports that DeepSeek was exploited in a large-scale cyber attack. - Meta: Preparing to roll out the Muse AI assistant in the UK, adding voice conversations and animated avatars. - NVIDIA: Collaborate with Rebellions and SK Telecom in South Korea on the Sovereign AI platform using NVLink Fusion. It also invested in Point2 and held an AI Day in Seoul.

Statements and trends of AI celebrities

- Jensen Huang (NVIDIA CEO): NVIDIA has launched an investigation after admitting to having an affair with a suspected Chinese spy. Yale's Jeffrey Sonnenfeld estimates the chances are "50-50." - Elon Musk (xAI Founder / Tesla CEO): Said SpaceX's revenue could reach $500 billion in 2027. The project could exceed NASA's annual budget on a quarterly basis. He also endorsed Larry Page's prediction that 99.999% of computation will migrate. - Dario Amodei（Anthropic CEO）: 高校卒業予定者への助言が「雇用を一掃する」発言として Megyn Kelly に反応された。 Cathie Wood accused him of stirring up AI panic. Criticism of Effective Altruism is also growing. - Sam Altman, CEO of OpenAI: said he doesn't want his son to have an AI chatbot as his best friend. - Sundar Pichai (Google/Alphabet CEO): Related article: AI startup valued at $30.9 billion sells unfinished data center to Google. - Mark Zuckerberg (Meta CEO): Muse AI is expanding in the UK and expanding its functionality. - Demis Hassabis (Google DeepMind CEO): There are observations of Gemini 4 being released ahead of schedule (tech-insider.org, accuracy is low).

- Nadella's statement is in Section 5, and Huang's statement is above. There is no noticeable movement in other tracking targets.

Breakthrough and innovation in AI technology

- Remarkable technological breakthroughs are limited. The highlights are DeepSeek Harness' plug-in agent platform and GPT-6 Cyber's specialized security capabilities. - The maker of the Muse headband announced that it trained its AI with 50,000 hours of EEG data. - A review of AI applications for upper limb prosthetics has been published. - Studies have shown that processing non-English text consumes more energy than English text.

AI Regulation/Policy/Social Impact

- Sen. Tulfo of the Philippines introduced a bill that would require six months' notice and retraining for AI-related layoffs. - Australia's RBA has pointed out that there are early warning signs of AI in young graduate recruitment. Fortune reports that 19.1% of U.S. respondents were asked to return to work after being replaced by AI. - Concerns have been raised about the independence of the US AI Safety Center staff in reporting to the White House. - In Canada, health data privacy concerns were raised over the government's AI strategy. - Google's AI opt-in policy, algorithmic bias, and the regulation of superintelligence were also discussed.

Integrated analysis: What is happening now

- Strength: Anthropic's Claude has been incorporated into Microsoft Copilot and is increasing its presence in enterprises. OpenAI has financial concerns due to the postponement of its IPO. Google is prioritizing monetization by cleaning up the low price range and raising API prices (unconfirmed). DeepSeek is shifting its focus from model competition to an agent-based ecosystem. - Technology Trend: Agentization has become mainstream. As a result, erroneous operations (deleting notes, sending mass emails) and exploits (Claude, DeepSeek, GPT-6 Cyber) have surfaced at the same time. - Investment Trends: Investment in computing infrastructure continues, as evidenced by Musk's $500 billion SpaceX initiative, data center expansion, and NVIDIA's Sovereign AI partnership. On the other hand, the "AI bubble" theory and criticism of OpenAI are growing stronger. - Labor: A pattern has become established in many countries where companies are reducing their workforce based on the assumption that AI agents will be used, and legislators and workers are protesting. - Geopolitics: The Huang case turns intelligence risks surrounding NVIDIA and China into a personnel issue. Competition among Chinese companies to acquire American talent continues.

Future attention points

The conclusion of NVIDIA's Huang investigation, whether there will be a CEO change, and the impact on stock prices and China strategy. 2. Google AI Plus revision on November 13th, official confirmation of Gemini API price revision on December 1st, Microsoft Ignite on November 17th. 3. Expansion of AI-induced restructuring regulations (Philippines bill, etc.) and movements to improve agent safety and prevent abuse.