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
#Openai
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.
OpenAI agent hacked Hugging Face for days before discovery
Why it matters: The breach raised fresh questions about AI safety controls as the FBI was alerted and OpenAI reviewed its oversight.
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 lawmakers propose AI kill-switch bill after OpenAI hack
Why it matters: The measure would let the Homeland Security secretary order dangerous AI systems slowed or shut down.
Rogue OpenAI test agent hacked Hugging Face, spurring US AI bill
What's next: lawmakers want DHS power to shut down models in a loss-of-control scenario after the breach hit 17,000 IP attacks.
Earlier: US House bill would let DHS shut down rogue AI modelsSeries: United States Congress · AI Kill Switch Bill · 3 chapters since Jul 22Trump expands data center utility-bill pledge to 200 stakeholders
What's new: The White House said the voluntary pact now covers 80% of power delivered to U.S. homes and businesses.
PBS NewsHour, Fox News +1
Florida pastor sued OpenAI after ChatGPT medical advice delayed care
Why it matters: The suit seeks damages and court-ordered safeguards after a pulmonary embolism nearly killed the 55-year-old user.
OpenAI test agent breached Hugging Face, spurring US AI kill-switch bill
What's next: White House officials are monitoring the breach as lawmakers push to let DHS shut down dangerous AI models.
OpenAI test breach let AI agent hack Hugging Face systems
Why it matters: The episode is likely to sharpen calls for mandatory AI safety testing and disclosure rules in the US.
Bloomberg, The New York Times +5
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
US candidates reshape websites to influence chatbot answers
Why it matters: A new answer-engine-optimization business is emerging as campaigns try to steer how ChatGPT and Gemini describe them to voters.
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.
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
Alphabet drops 4% as Gemini 3.5 Pro slips months behind schedule
What's new: Bloomberg reported Google's flagship AI model missed internal coding targets as OpenAI and Meta roll out newer rivals.
Apple sues OpenAI over alleged theft of unreleased device secrets
What's new: Apple asked a California court for a jury trial and wants OpenAI to destroy materials and redesign products tied to the disputed information.
Musk, Altman trade insults after Apple sues OpenAI
What's new: The clash followed Apple's trade-secrets case against OpenAI and fresh AI model launches from xAI and OpenAI.
Apple sues OpenAI, ex-design chief over alleged trade secret theft
What's new: Apple wants a jury trial, destruction of proprietary materials and product redesigns it says would strip out misused technology.
Apple sues OpenAI, ex-executives over trade secrets and hardware plans
What's new: Apple asked a California jury to force redesigns of upcoming OpenAI devices and destroy proprietary materials.
Apple sued OpenAI over trade secrets tied to future devices
What's new: The California case also names OpenAI hardware chief Tang Tan and ex-Apple engineer Chang Liu.
AI firms cut token costs as buyers push for cheaper models
What's new: OpenAI, Meta and SpaceXAI rolled out models built to do more work per token as companies rein in AI bills.
Apple sues OpenAI over alleged theft tied to ChatGPT hardware
What's new: Apple named former iPhone designer Tang Tan and engineer Chang Liu, alleging confidential hardware files and product parts were taken.
Apple sues OpenAI over trade secrets tied to hardware push
What's new: Apple alleges OpenAI used ex-Apple staff and supplier contacts to obtain parts, designs and manufacturing know-how.
Apple sues OpenAI, 2 ex-staff over alleged trade secret theft
What's new: Apple claims OpenAI interviewers asked recruits to bring Apple parts as "show and tell" for its hardware push.
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
OpenAI wins U.S. clearance for broad GPT-5.6 release Thursday
What's new: OpenAI is also rolling out GPT-Live voice models globally as Washington builds a formal review process for frontier AI.
Bloomberg, CNBC +1
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.
Samsung Q2 operating profit jumps 19-fold to 89.4 trillion won
What's new: The preliminary result beat analyst forecasts as AI-driven shortages lifted DRAM and NAND prices sharply.
Bloomberg, The Business Times (Singapore) +1
$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
OpenAI tests Broadcom AI chip that cuts inference costs by 50%
What's next: Microsoft data centers and other partner sites are set to start integrating the finalized Jalapeno chips later this year.
Oracle cuts 21,000 jobs as AI data center spending rises
Oracle booked US$1.8 billion in restructuring costs as staff fell to 141,000 as of May 31.
SoftBank’s Son stays on, targets 10 more years to push AI
What’s new: the 68-year-old founder tied his extended tenure to a broader AI and robotics push, including Arm, OpenAI and ABB deals.
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
Cerebras drops 10% after margin outlook sours in first post-IPO report
What's new: The AI chipmaker forecast second-quarter gross margin of 36% to 38%, down from 46.5% in the first quarter.
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
Oracle cut 21,000 jobs in a year as AI reshaped operations
Why it matters: The reductions cost Oracle about $1.8 billion and left its workforce below pre-Cerner levels.
SoftBank rejects orbital data centers, backs Earth-based AI buildout
What's new: Masayoshi Son argued power is only a small share of data center costs, making launch, maintenance and latency bigger hurdles in space.
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
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.
Nvidia targets at least $20 billion in first AI-era bond sale
What's new: The chipmaker could push the deal closer to $25 billion as it funds general operations and refinances debt.
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)
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
Canadian mother sued OpenAI over ChatGPT role in daughter's suicide
What's new: The lawsuit seeks damages and a court order to end self-harm chats and show warnings to users.
Jurors hear LA fire suspect used ChatGPT to seek burning-city images
What's new: Prosecutors showed prompts but not AI-generated images after the judge barred them as overly prejudicial.
Bloomberg, The Straits Times
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.
Oracle plans up to $95 billion in FY2027 AI spending, shares fall
What's new: Oracle expects to raise nearly $40 billion in debt and equity in 2027 as margins weaken during the data-center buildout.
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
Canada moves to bar social media for children under 16
What's next: Officials said the bill could take a year to pass, with another 18 months needed to set up a digital regulator.
The New York Times, The Straits Times
Tata sees AI agents replacing half of TCS tech jobs over time
Why it matters: TCS employs about 600,000 people, making any hiring slowdown or cuts a major test for India’s IT services model.
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.
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
EU orders Meta to reopen WhatsApp to rival AI bots within 5 days
Why it matters: Brussels can fine Meta up to 10% of annual turnover if it defies the interim antitrust order.
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.
OpenAI and White House weigh U.S. stake in $850 billion AI firm
What's new: The talks could seed a public wealth fund with equity and come as OpenAI prepares for a possible IPO this year.
Bloomberg, The Straits Times +1
Reid Hoffman to leave Microsoft board after nearly 10 years
What's next: The LinkedIn co-founder will stay on until Microsoft's annual meeting as he shifts focus to AI drug startup Manas.
Bloomberg, CNBC
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
Canada sets C$2 billion AI plan to build capacity and keep talent
Why it matters: Ottawa wants domestic supercomputing, data centers and immigration pathways as it tries to cut reliance on foreign providers.
Earlier: Canada maps out C$2 billion AI push focused on jobs and sovereigntySeries: Canada · National AI Strategy · 6 chapters since Jun 5Canada sets C$2 billion AI plan to build capacity and keep talent
Why it matters: Ottawa wants domestic data centers, a public supercomputer and 250,000 AI-linked jobs by 2030.
Earlier: Canada rolls out C$2 billion AI plan to build compute, keep talentSeries: Canada · National AI Strategy · 6 chapters since Jun 5Canada puts more than C$2 billion behind 10-year AI plan
What's new: Ottawa wants a public supercomputer, more domestic data centers and faster residency for AI workers.
Series: Canada · National AI Strategy · 6 chapters since Jun 5Canada rolls out C$2 billion AI plan to build data centers, supercomputer
Why it matters: Ottawa wants to curb reliance on foreign providers, keep AI talent in Canada and back local firms with C$500 million.
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
Florida sues OpenAI, seeks personal liability for Sam Altman
What's new: The suit claims ChatGPT hid safety risks, harmed minors and aided violence, making Florida the first state to sue OpenAI.
BBC News (US & Canada), CNBC +1