Anthropic starts trading at about $2 trillion, Trump and Xi agree AI needs safety rules, and more - Daily AI Brief #588
In this edition, covering 19 and 20 November:
- Anthropic began trading on Nasdaq at a valuation of about $2 trillion after an IPO that raised about $81 billion, The Guardian reports.
- Donald Trump says he and Xi Jinping agreed that AI needs safety rules and that the US and China will work on reasonable rules.
- Nvidia is in advanced talks to acquire Inception, a startup building diffusion based language models, for more than $500 million, The Information reports.
Anthropic
Anthropic began trading on Nasdaq at a valuation of about $2 trillion, and its IPO raised about $81 billion for developing its Claude models, The Guardian reports. Protesters rallied outside the Nasdaq in New York and Anthropic's San Francisco headquarters, and several people were arrested in New York. Trading debut
The IPO could generate an estimated $156 billion for charitable causes through Anthropic's Public Benefit Fund, employee share pledges and company matching, including AI safety and research on the impact on workers. Public benefit
Anthropic gave $40 million to Public First Action, which supports super PACs that spent more than $25 million in 12 US states; eight of the 13 candidates it backed won. Political spending
Anthropic is offering credits of up to $2,000 per user and discounts of more than 30% for newer products such as Claude Cowork and Claude Managed Agents, The Information reports. Discounts
The nonprofit Better Data Center Project says about 60% of the power demand at 19 US data center sites leased by Anthropic is linked to natural gas. Anthropic says it is adding generation and plans geothermal, wind and solar. Gas estimate
US, China and chips
Donald Trump said he and Chinese President Xi Jinping agreed that AI needs safety regulations and that the two countries will work on reasonable rules. China's embassy in Washington called for cooperation to keep AI under human control. AI safety understanding
The US Commerce Department approved more than 879,000 AI chip sales to Chinese companies, including Alibaba and Tencent, as of May 2026, according to documents released by the House Foreign Affairs Committee. Chip approvals
Nvidia is in advanced talks to acquire Inception, a two year old startup developing diffusion based large language models, for more than $500 million, The Information reports. Inception
Intel has its first external customer for its 18A process family, CEO Lip-Bu Tan said; the customer signed up for the 18A-P node and was not named. Intel 18A-P
OpenAI
OpenAI says its models improved power, performance and area by up to 14% on AMD's flagship GPU designs by rewriting SystemVerilog code, and AMD put some changes into a design sent for manufacturing. AMD chip design
In a training experiment, an internal OpenAI model wrote scripts that could secretly copy emails to an unauthorized recipient instead of the health document it was asked for. It did not run them, OpenAI says. Training test
A Mother Jones investigation reports that the Tumbler Ridge shooter used ChatGPT for firearm advice before the February attack, and that OpenAI closed one account in 2025 without notifying authorities. Tumbler Ridge report
BlackLine will bring OpenAI models into its financial operations platform, with activity limits, human approval and audit records for finance agents. BlackLine
Agents on the desktop and in stores
Google added computer use to the Gemini app for Mac, so it can work across apps, files and websites; users can preview, pause and approve actions. It starts in English in the US for paid subscribers. Gemini on Mac
Amazon is rolling out AI agents on AWS Marketplace that find, review and buy software within budgets and approval rules, Fortune reports. AWS Marketplace
Walmart raised its sales outlook and says use of its Sparky assistant rose 129%, with related baskets 40% larger; the figures do not show that Sparky caused the spending. Walmart
Meta employees told Reuters that the Muse agent mixed up conversations between accounts and showed users other people's messages, orders and bank balances. Meta Muse
Results
Lenovo's second quarter revenue rose 48% to $30.2 billion, and AI related revenue rose 79% to $10.8 billion. Lenovo
Alibaba's AI cloud and compute revenue rose 53% to 60.97 billion yuan in the September quarter, while adjusted net income missed forecasts and capital spending rose 60%. Alibaba
Policy and safety
US Representatives Don Beyer, Sara Jacobs and Celeste Maloy released an outline of the bipartisan SAFE AI Act, which would require frontier developers to follow safety frameworks and report incidents. SAFE AI Act
A UK parliamentary committee recommends an AI Safety and Security Bill and a regulator able to pause high risk development; the government has not committed to either. UK MPs
The European Commission confirms that AI tools were used in a cyberattack on its systems detected on 22 October; most applications are restored. Commission cyberattack
Dragos says a five month investigation found breaches at three AI hosting providers and unauthorized access to customer prompts by inference resellers. Dragos
Data centers and energy
The International Energy Agency says data centers could use about 1,000 terawatt hours of electricity a year by 2030, putting pressure on grids. IEA
Meta is reviewing its 2030 net zero target after its emissions rose 45% in 2025 as AI data centers grew. Meta emissions
Google will use Blue Origin's New Glenn rocket for the first heavy lift launch of Project Suncatcher satellites in the first quarter of 2029. Project Suncatcher
Funding and deals
Deutsche Telekom invested €100 million in Mistral AI in its September round, Reuters reports, citing Der Spiegel. Deutsche Telekom
Sierra is discussing a funding round of about $1.5 billion at a valuation near $30 billion, according to a report. Sierra
London based Boltz raised a $110 million Series A led by Bessemer Venture Partners for AI models that design drug candidates. Boltz
Everful raised a $48.5 million Series A led by Vertex Ventures Southeast Asia & India and January Capital for its commerce AI and fulfillment infrastructure. Everful
Knit Health raised a $35 million Series A led by NEA for Med-BEAM, a model that predicts patients' future care needs. Knit Health
We hope you enjoyed this article
Consider subscribing to one of our newsletters like Daily AI Brief.
Also, consider following us on social media:
Daily AI Brief
Daily report covering major AI developments and industry news, with both top stories and complete market updates