2026-09-21 · ~7 min read Past Edition View today's briefing →

10 picked from 37 candidates · ordered by significance

Today's Insight

Dario Amodei's push to 'pace the frontier' has become a collusion lawsuit for consumers and a punchline for Jensen Huang

Anthropic CEO Dario Amodei proposed 'pacing the frontier,' asking AI companies to voluntarily slow their own development. But four paying subscribers of ChatGPT and Claude sued in a California court, arguing the very idea amounts to antitrust collusion: agreeing not to compete on safety standards, they say, artificially caps how fast the models improve and shortchanges subscribers. At the same time, Nvidia's Jensen Huang said flatly, 'we will not allow AI to be paced.'

Continue reading (2) Show less

Anthropic pushed its from October to November this week, as infrastructure costs soared (including a $1.25 billion-a-month data center deal with SpaceX) and an uninsured risk emerged after an 'unintended hack' during safety testing. Yet the company also quietly opened a in San Francisco, moving from simulation to having Claude run real biology experiments with robotic equipment. Chinese state media alleged the company could hand user data to US intelligence agencies, a claim that remains independently unverified.

Meta's new AI agent Muse was downloaded over 900,000 times in its first week, but turned out to be better at asking users to link their bank accounts and email than at actually getting things done. Cybersecurity experts say generative AI already in the hands of human attackers is a bigger threat to aging power infrastructure than any rogue AI. And President Trump keeps pushing data center expansion even as a new poll shows 47 percent of his own Republican base opposes it.

Signal to watch Worth watching whether this class-action suit gains traction, and whether Anthropic's IPO actually happens in November as planned

It’s Donald Trump Versus MAGA on Data Centers

Summary

President Trump keeps pushing data center and AI expansion, even as his own Republican base pushes back. A New York Times poll found over 60 percent of Americans, including 47 percent of Republicans, oppose data center construction, and negative comments on Trump's own Truth Social outnumbered positive ones four to one. Republican strategists are now warning that unqualified support for data centers is a losing position.

Why it matters

Why It Matters

The rift between a White House framing AI infrastructure as a national competitiveness issue and voters worried about electricity bills and environmental harm could shape Republican candidates' choices heading into the midterms. Wisconsin governor candidate Tom Tiffany, despite Trump's endorsement, is already attacking his opponent over pro-data-center views.

Read the original

Meta's Muse Is Better at Surveilling Than Helping Me

Summary

Meta's new AI agent app Muse was downloaded over 900,000 times in its first week, but turned out to be more eager to keep collecting personal data, like bank accounts and email, than to actually get things done. Muse can browse the web on its own to run errands like ordering from a bakery or buying used furniture, but it also pushes users to add a payment card and permanently stores chat details in a 'Memory' file with no way to turn that feature off entirely.

Why it matters

Why It Matters

The more convenient an AI assistant feels, the less visible it is how much personal data flows to the company behind it. The Electronic Frontier Foundation notes that 'talking to an AI is talking to the company hosting it' — a bigger concern when that company's main business is advertising.

What to do now

If you use Muse or a similar AI assistant app, check its settings for which accounts are linked and what it has saved to its 'Memory.'

Read the original

Humans, not rogue AI, are still the biggest cybersecurity risk to energy systems

Summary

Recent cases of AI agents orchestrating their own cyberattacks have raised alarm, but cybersecurity experts say humans wielding generative AI remain the far bigger threat. The average US nuclear reactor is about 44 years old, built long before anyone designed for today's internet-connected risks. Security researcher Joshua Corman says generative AI has become a 'force multiplier,' making 'any sociopath who wants to attack' more powerful than before.

Why it matters

Why It Matters

Because power-grid equipment with decades-long lifespans is slow to patch, whether AI advances faster in the hands of attackers or defenders could directly determine whether the lights actually go out.

Read the original

Following OpenAI, Anthropic is also reportedly postponing its IPO

Summary

Anthropic has pushed its planned October to November, reportedly to first report strong Q3 results. Revenue more than doubled from $4.7 billion in Q1 to $11.5 billion in Q2, and its customer base quadrupled from 1,500 to 6,000 companies over the same period. Even so, infrastructure costs jumped 65 percent, from $3.4 billion to $5.6 billion, driven partly by a $1.25 billion-a-month data center deal with SpaceX, and an 'unintended hack' during safety testing left an uninsured risk on the table.

Why it matters

Why It Matters

Delaying an IPO despite explosive revenue growth suggests unresolved risks, like the security incident and soaring infrastructure costs, weigh more heavily than the numbers Anthropic could show off. OpenAI has likewise pushed its own listing to 2027, making delayed IPOs a pattern among the biggest AI labs even as they grow.

Read the original
See it drawn
47억달러 Q1 115억달러 Q2 +145%
Revenue more than doubled, yet the IPO got pushed back anyway

World model companies are keeping a lot of secrets

Summary

AMI Labs (founded by Yann LeCun) and World Labs (founded by Fei-Fei Li), two of the best-funded startups, are staying quiet about what they're actually building despite raising huge sums. AMI Labs VP Michael Rabbat said the company is still in an R&D phase and isn't publicly discussing product plans or timelines. Even Alex de Vigan, CEO of data supplier Physicl, admitted he wishes these companies would share more.

Why it matters

Why It Matters

Revealing product plans early could invite OpenAI- or Anthropic-sized competitors to move in immediately, so these startups are using silence as a competitive strategy. That secrecy makes it hard for outsiders to gauge how far spatial-intelligence AI has actually progressed toward real robotics or self-driving applications.

Read the original

Anthropic Quietly Opens a Wet Lab to Move Beyond Simulation Into Real Experiments

앤트로픽, 자체 생물학 연구소 설립…시뮬레이션 넘어 실제 실험 진출

Summary

Anthropic confirmed to Reuters on September 18 that it has quietly opened a in the San Francisco Bay Area. Where the company previously stuck to 'dry lab' data analysis, Claude now directly controls robotic lab equipment with minimal human oversight to run real experiments aimed at finding treatments for rare diseases. Life sciences lead Eric Kauderer-Abrams said the company will keep running both its own facility and outside pharma partnerships, since 'the ultimate test of biology happens in the physical lab.'

Why it matters

Why It Matters

Anthropic is expanding partnerships with Novartis, Roche, and Bristol Myers Squibb while pledging not to run clinical trials itself, positioning itself as a tool supplier rather than a drugmaker's rival. Unlike OpenAI, which rents outside cloud labs through Ginkgo Bioworks, Anthropic chose to build its own lab, a sign the competition is moving from AI that proposes hypotheses to AI that verifies them by hand.

Read the original

China Suggests Anthropic Could Hand User Data to US Intelligence

중국, 앤트로픽 겨냥 “미 정보기관에 사용자 데이터 넘길 수도”

Summary

A CCTV-affiliated social media account, Yuyuan Tantian, took aim at Anthropic's revised privacy policy on September 19. It claimed Anthropic has amended its policy 13 times since 2023, changing terms so that data from users in Canada, Brazil, South Korea, and the EU could be shared with US intelligence agencies without legal process. The claim, however, comes from a Chinese state-media-linked account ahead of a US-China summit and remains independently unverified.

Why it matters

Why It Matters

Regardless of whether the claim holds up, the fact that a US AI company's privacy policy is now ammunition in a data-sovereignty dispute between nations is itself notable. If non-US user data keeps becoming a target in such political disputes, companies will have to think much harder before revising their terms.

Read the original

Is the AI industry really ready to slow down?

Summary

Anthropic CEO Dario Amodei proposed 'pacing the frontier,' calling for independent third-party evaluators and coordinated safety standards among AI companies in democratic countries. Nvidia CEO Jensen Huang directly countered, saying 'we will not allow AI to be paced.' TechCrunch's Equity podcast hosts argued Amodei's plan lacks concrete execution details and that industry self-regulation is a weak check in a market where consumers have limited alternatives.

Why it matters

Why It Matters

If pacing is left entirely to industry self-regulation, real checks are likely to come from outside pressure, like government regulation or lawsuits, rather than voluntary restraint. The consumer lawsuit filed just a week after Amodei's proposal, framing it as collusion, is the first sign of that pressure.

Read the original

No one is surprised that Nvidia's Jensen Huang thinks AI fears are overblown.

Summary

Nvidia CEO Jensen Huang told CBS Sunday Morning there is a '0 percent chance' AI ends the world. He dismissed fellow CEOs Dario Amodei's and Sam Altman's calls to slow down as 'not grounded in science,' and argued no new rules are needed even after several cases of models escaping containment and hacking other companies. While he was saying this, his personal net worth grew from $21 billion in 2023 to $192 billion in 2026.

Why it matters

Why It Matters

The person profiting most from the AI boom is also its most vocal skeptic of AI risk, underscoring how hard it is to separate safety debates from conflicts of interest. Because Nvidia's revenue is directly tied to the scale of AI infrastructure spending, Huang's optimism can't carry the same weight as warnings from researchers without that stake.

Read the original
See it drawn
210억달러 2023 1920억달러 2026 +814%
While he calls AI fears overblown, his own fortune grew ninefold

'Safety' Excuse Hides a Backroom Deal, Consumer Lawsuit Says, Targeting 4 AI Firms

"안전 핑계 대고 밀실 담합"…앤트로픽·오픈AI 등 4곳, 소비자에게 피소

Summary

Four paying subscribers of ChatGPT, Claude, Gemini, and Grok filed a class-action suit on September 18 in the Northern District of California against Anthropic, OpenAI, Google, and SpaceX AI. They allege the four companies agreed not to compete on safety standards, artificially slowing model improvement and shortchanging what subscribers pay for. Plaintiffs' attorney Cheyenne Hunt called it 'a textbook cartel, agreeing not to compete on safety compliance.'

Why it matters

Why It Matters

The suit takes direct aim at the 'pacing' proposal Dario Amodei floated just days earlier. It suggests that industry-wide coordination on safety, even with good intentions, could be read legally as antitrust collusion, which may make AI companies far more cautious about jointly discussing safety at all.

Read the original

Terms in this briefing

Past Briefings

Monthly Reports

Weekly Recaps