Topic: #ai
-
Cerebras' 5nm wafer-scale chips skirt HBM and CoWoS supply bottlenecks, analysis says
Cerebras' WSE-3 chips skip HBM, CoWoS packaging and 3nm capacity, according to a Crypto Briefing analysis, which also flags execution risk on the company's $25.4 billion backlog.
-
Cathie Wood rebuts Bill Ackman's warning that AI demand could fuel inflation
ARK Invest's Cathie Wood pushed back on Bill Ackman's warning that the Fed's September rate hike could backfire. She argued AI is driving "benign deflation," not an inflationary spiral, Crypto Briefing reported.
-
Meta, Duke and UC Davis researchers detail branching method for AI agent harnesses
Researchers affiliated with Meta, Duke and UC Davis report in an arXiv preprint that splitting harness optimization into self-improving branches raised Gemini 3 Flash's Olympiad math accuracy from 46.0% to 62.0%.
-
Meta's Muse AI shopping agent adds Walmart, Gap, Sephora, Wayfair, Best Buy
Meta announced retail partnerships for Muse, its standalone AI agent app, with Walmart, Gap, Sephora, Wayfair and Best Buy on September 23, about two weeks after launch, according to Crypto Briefing.
-
Arbitrum Security Council pauses new Stylus activations, citing AI-assisted attacks
Arbitrum's Security Council blocked new Stylus activations on Arbitrum One and Nova on October 2, citing AI-assisted attack risks. Live Stylus apps and Solidity contracts are unaffected.
-
Apple tightens Mac Full Disk Access prompts for AI tools amid Meta Muse dispute
Apple announced macOS updates on October 2, 2026 that make AI requests for Mac data more visible and require explicit user action for Full Disk Access, after complaints about Meta's Muse agent that Meta disputes.
-
Eric Schmidt calls a mutual US-China pause in AI development unworkable
Former Google CEO Eric Schmidt told The Economist that a mutual US-China pause in AI development is unworkable. He cited US incentives to build faster and the lack of any reliable way to verify that the other side has stopped.
-
Trillium Labs pledges to publish AI research as Trillion Labs pushes open models
Trillium Labs, a new AI lab, says it'll publish findings on AI self-improvement and model behavior, while South Korea's separate Trillion Labs has open-sourced its Tri series models.
-
ECB survey finds 72% of euro area firms plan to fund AI from their own cash
An ECB survey released October 2 found 72% of euro area firms planning AI investments intend to self-fund them, while bank loans, grants and leasing together cover only about 16%, Crypto Briefing reported.
-
OpenAI, Anthropic, Google and Meta face surge in AI lawsuits, Axios reports
AI companies including OpenAI, Anthropic, Google and Meta face a surge in lawsuits over safety, copyright and liability, according to an Axios report cited by Crypto Briefing.
-
California business owner arrested over alleged $300M AI chip smuggling to China
Federal prosecutors in Los Angeles arrested Greg Lui, owner of Earthmade Computer Inc., on charges he moved more than $300 million in export-controlled AI servers to China via Malaysia and Singapore.
-
Ethereum Foundation's zkAPI goes live on mainnet for private AI API payments
The Ethereum Foundation said zkAPI, built with the Open Anonymity Project, is live on Ethereum mainnet. It lets users pay for AI and other APIs without revealing billing identities, but it doesn't hide prompts or network metadata.
-
Chinese state-backed lender financed Nvidia B300 servers, Bloomberg reports
Semi-Tech Leasing, majority-controlled by the Shenzhen and Beijing city governments, helped fund 700+ servers, including some with export-restricted Nvidia B300 chips, for Glory View, Bloomberg reported via Crypto Briefing.
-
OpenAI launches dots agents on GPT-6 Astra with user sign-off for sensitive tasks
OpenAI introduced dots, autonomous agents on its GPT-6 Astra model that work across Slack and Microsoft Teams, Crypto Briefing reported. The agents need user sign-off before sensitive actions such as deleting data.
-
Google unveils Gemini 4 Argon, an AI model aimed at finding and patching software bugs
Google launched Gemini 4 Argon on September 30, 2026. Google says the model can autonomously discover, validate and patch software vulnerabilities, and for now access is limited to its Fairwind Program.
-
Wells Fargo's Mike Mayo says AI 'scare trade' in bank stocks is overblown
Wells Fargo analyst Mike Mayo called the AI-driven selloff in bank stocks an exaggerated “scare trade,” arguing customer trust is a moat chatbots can't easily overcome, Crypto Briefing reported.
-
EU proposal to triple data center capacity meets growing local pushback
The European Commission has proposed tripling EU data center capacity within five to seven years, a roughly €200 billion buildout, as communities in Ireland, the Netherlands and France push back, Crypto Briefing reported.
-
China's STAR 50 down about 30% since June, on track for worst quarter in recent memory
China's STAR 50 Index has fallen about 30% since the end of June, Crypto Briefing reported, after a July selloff (following a nearly 75% three-month rally) that the outlet tied to doubts about returns on AI spending.
-
AI chiefs sign voluntary safety pact at Trump lunch, Crypto Briefing reports
Leading AI executives signed a voluntary pact backing independent safety audits at a September 29, 2026 lunch with President Donald Trump, according to Crypto Briefing, in a report that hasn't been independently confirmed.
-
Nvidia, AMD ramp up White House lobbying over China AI chip export controls
Nvidia and AMD have stepped up White House lobbying over US export controls on AI chips to China, with Nvidia's spending rising from $640,000 in 2024 to nearly $3.5 million in 2025's first three quarters, Crypto Briefing reported.
-
India lines up roughly $25B for deep tech startups, Crypto Briefing reports
India is marshaling roughly $25 billion in combined public and private money for deep tech startups, according to a Crypto Briefing estimate, with a ₹1 lakh crore research scheme at its centre.
-
Goldman Sachs projects $1.2 trillion in 2027 hyperscaler AI capex, flags energy limits
Goldman Sachs strategists project the five largest US hyperscalers will spend roughly $1.2 trillion on AI infrastructure in 2027, Crypto Briefing reported. They named energy supply as one of three bottlenecks.
-
Modal Labs in talks to raise funding at valuation near $15 billion, report says
Modal Labs, a serverless AI infrastructure company founded in 2021, is in talks to raise new funding at a valuation approaching $15 billion, Crypto Briefing reported. That's more than triple its May 2026 Series C valuation.
-
NEAR AI Cloud joins OpenRouter as provider, serving GLM 5.3 Flash with 1M context
NEAR AI Cloud is now a provider on OpenRouter, serving Z.ai's GLM 5.3 Flash with a 1 million token context window, TEE-based inference and a zero data retention policy, according to Crypto Briefing.