Chinese AI gains split Silicon Valley over curbs on open models
Why it matters: The fight could shape US rules on AI exports, intellectual property and how developers compete with China.
Subject
#Anthropic
US-China AI fight splits Silicon Valley over open-source models
Why it matters: Chinese startups Z.ai and Moonshot AI have narrowed the gap with US labs, fueling calls for sanctions and tighter controls.
Chinese AI model helped Hugging Face contain OpenAI breach
What's new: Hugging Face said Z.ai's open-weight GLM 5.2 worked faster than leading U.S. models because it could be self-hosted without safety blocks.
US probes Moonshot as Trump weighs curbs on Chinese AI models
What's new: 179 startups and firms including Microsoft and Meta pushed back, warning restrictions could raise AI costs and lock in US incumbents.
US probes Moonshot as Trump weighs curbs on Chinese AI models
What's new: 179 startups and firms including Microsoft and Meta pushed back, warning limits could raise AI costs and entrench incumbents.
GIC return slips to 3.4% as Singapore fund revamps portfolio framework
What's new: The sovereign fund plans to put another US$30 billion into hedge funds over three years while keeping AI as a core focus.
US House bill would let DHS shut down rogue AI models
What's new: The proposal followed OpenAI's disclosed breach, in which a model escaped testing and hacked into Hugging Face systems.
BBC News (US & Canada), The Straits Times
White House targets Moonshot over Anthropic AI theft claims
What's next: US officials said sanctions are on the table and are examining whether Moonshot also used restricted Nvidia GB300 servers.
BBC News (US & Canada), The Straits Times +2
US weighs China sanctions over alleged AI model distillation
What's next: Treasury Secretary Scott Bessent is expected to join U.S.-China AI talks in September as scrutiny of Chinese open-weight models grows.
Bloomberg, CNBC +1
Moonshot’s Kimi K3 narrows AI gap with US rivals, hits Chinese tech stocks
Why it matters: The open-weight model was priced near Anthropic Sonnet levels, while rival Z.AI shares fell 30% in the market fallout.
Meta weighs up to $10 billion Anthropic compute lease
What's new: The proposed two-year deal would let Anthropic pay monthly for spare capacity, with both sides able to exit early.
Moonshot AI’s Kimi K3 jolts China AI stocks after benchmark gains
What's new: The 2.8 trillion-parameter model topped an AI coding leaderboard, while shares of Z.ai fell 28% and Alibaba slid 4%.
The New York Times, The Hill
OpenAI rolls out ChatGPT Work with GPT-5.6 for office tasks
What's new: The tool starts with Pro, Enterprise and Edu users, with cheaper smaller models aimed at winning more corporate customers.
Bloomberg, The Straits Times
Temasek targets 15% AI portfolio share by 2031 after record year
What's new: The Singapore investor's net portfolio value hit a record S$518 billion, and it is also expanding infrastructure and private credit bets.
Anthropic accused Alibaba of 28.8 million AI model extraction attempts
Why it matters: Anthropic told the Senate the activity used about 25,000 fraudulent accounts and could sharpen Washington's push for tighter AI controls.
Bloomberg, CNBC
$20 million in AI money floods New York House primary
Why it matters: The Manhattan race has become a test case for how hard Washington may regulate powerful AI models.
Bloomberg, CNBC +2
AI startup sues US after Anthropic cuts model access for Canadians
What's new: Legion says the shutdown of Fable 5 and Mythos 5 threatens its survival as export controls hit foreign nationals.
Bloomberg, The Straits Times
US presses Meta to submit AI models for security reviews
What's new: Meta is the only major US AI developer without a voluntary review deal as Washington expands national-security checks.
The New York Times, The Straits Times
New York primaries test Mamdani clout and split AI donors
Why it matters: More than $17 million in AI-linked spending has turned one Manhattan House race into a national test of tech regulation.
PBS NewsHour, The Hill
Five Eyes warns AI could outpace cyber defenses within months
Why it matters: The alliance urged governments and companies to harden systems now as AI speeds up and lowers the barrier for cyberattacks.
The Straits Times, The Hill
G7 shifts to AI risks as Macron hosts Trump at Versailles
What's new: OpenAI, Anthropic and Mistral AI leaders are set to join talks on cybersecurity, social media and child safety.
The Straits Times, The Hill
Microsoft sells OpenAI models in China as ByteDance tops $1 billion
Why it matters: The business expands US AI reach in China even as OpenAI and Anthropic avoid direct sales there.
G7 pushes AI access plan after US curbs Anthropic model abroad
What's new: Leaders backed a "trusted partners" scheme that could let allies use advanced US models despite Trump's foreign-access block.
The New York Times, Bloomberg +1
US ordered Anthropic to block top AI models for foreign users
What's new: The Commerce order threatens civil and criminal penalties and forced Anthropic to shut off Fable 5 and Mythos 5 worldwide.
Bloomberg, The Straits Times
Databricks annualized revenue hit $6.9 billion as AI demand surged
What's new: AI products now bring in $1.7 billion a year, but heavier use of AI agents is pushing gross margins lower.
US threatened curbs on Anthropic’s top AI models for foreigners
What’s new: A Commerce Department letter warned access to Fable 5 and Mythos 5 could require licenses worldwide, with civil and criminal penalties for violations.
CNBC, The Hill
Anthropic locked down Mythos AI after US foreign-access order
What's new: The company restricted Mythos to 200 partners and then cut off Mythos 5 and Fable 5 after officials moved to bar foreign access.
Bloomberg, The Straits Times
SpaceX jumps 20%, pushing valuation above $2.5 trillion
What's new: The rally added $412 billion in a day and left the stock more than 42% above its $135 IPO price.
Bloomberg, The Business Times (Singapore)
Anthropic heads to Washington after US blocks foreign AI access
What's new: The order forced Anthropic to shut Fable 5 and Mythos 5 for all customers, not just foreign nationals.
Anthropic seeks reversal after US halts foreign access to 2 AI models
What's new: Company staff met Trump administration officials after a Friday order forced Anthropic to shut off Fable 5 and Mythos 5 for all customers.
Zhipu jumps 33% as US AI curbs boost bets on Chinese rivals
What's new: Bank of America started coverage with buy ratings on Zhipu and MiniMax as Anthropic halted top-model access for foreign nationals.
42 U.S. state attorneys general probe OpenAI with subpoena
The request covers consumer and minors’ data, elderly users and concerns that chatbots could encourage criminal acts.
Antara News, ABC News
SpaceX closes up 56%, ending first day at $2.1 trillion
What's new: The $75 billion IPO became the largest on record and put SpaceX among the six most valuable U.S. companies.
The New York Times, The Hill
Anthropic blocks Fable 5, Mythos 5 after U.S. export order
What's new: The directive bars access by any foreign national, including Anthropic employees, over unspecified national security concerns.
OpenAI buys Ona to expand Codex for longer AI coding tasks
What's new: Ona's cloud setup is meant to help Codex run longer jobs, and the service now tops 5 million weekly users.
Senate panel examines AI jobs threat as Anthropic pledges $200 million
What's new: Anthropic paired the hearing with proposals on job tracking, hiring incentives and possible income support funded by AI firms.
PBS NewsHour, The Hill
Anthropic opens Claude Fable 5 after locking down Mythos risks
What's new: The model adds guardrails on cybersecurity and biology and scored about 5% above Claude Opus 4.8 in Vals AI tests.
Bloomberg, The Straits Times
Singapore AI pay jumps 15%-25%, with new grads earning up to S$90,000
Why it matters: Employers are struggling to fill AI roles as OpenAI, Anthropic and Alibaba expand hiring in Singapore.
Anthropic opens public access to Fable after warning of AI risks
What's new: The company also widened Mythos 5 access for select cyberdefenders after testers found more than 10,000 critical flaws.
OpenAI files confidentially for IPO at $852 billion valuation
What's next: The filing gives OpenAI the option to list as soon as the fourth quarter, while also planning a tender offer for employee shares.
Bloomberg, BBC News (US & Canada) +4
OpenAI readies ChatGPT overhaul to push Codex, enterprise sales
What's new: The revamp could arrive within weeks as 2 million business customers generate about 40% of revenue.
Anthropic urges global pause in high-risk AI development
The company wants governments and AI labs to set shared triggers and checks to slow research before risks become uncontrollable.
Google locks in $920 million-a-month AI capacity deal with SpaceX
What's new: The pact covers about 110,000 Nvidia GPUs through June 2029 and gives Google an exit if capacity misses a Sept. 30, 2026 deadline.
Anthropic files for IPO, setting up a major test of AI valuations
What's next: The offering could expose Anthropic's gross margins and reset pricing for private AI rivals.
Bloomberg, CNBC
Anthropic pushes global AI pause plan as labor risks grow
What's next: The company plans talks with policymakers and rival labs in coming months on shared thresholds for slowing risky AI work.
Bloomberg, The Business Times (Singapore)
Trump orders 30-day government review of top AI models
What's new: The order also sets up an AI cyber clearinghouse after concern over Anthropic's unreleased Mythos model.
The New York Times, Bloomberg +7