The Week in ShortAnthropic calls for foundation model regulations. SpaceX raised $75 billion in an IPO that requires several leaps of faith. Databricks reportedly taps the private markets yet again. Fable’s subterfuge stokes backlash and a course correction. Crusoe stumbles in Wyoming. OpenAI considers token price cuts as tokenmaxxing fever breaks and competitive pressure mounts. Big winners from the SpaceX IPO. The Main ItemRegulate us, please. That’s the message from Anthropic this week.The company has a new warning that foundation models — like the company’s freshly released Fable model — need to be regulated by governments at least as seriously as cars, airplanes, and drugs. Anthropic CEO Dario Amodei wrote that models need to be treated like “powerful technologies essential to the modern economy, but capable of killing large numbers of people if designed or operated poorly.”There’s a cynical, increasingly popular view in Silicon Valley, championed by David Sacks, that Anthropic’s desperate pleas for regulation are primarily self-serving. The argument goes that regulations would establish the current big foundation model providers, like Anthropic, as government-anointed winners, making it nearly impossible for new models to compete. David Sacks@DavidSacksAbout 8 months ago, I warned that “Anthropic is running a sophisticated regulatory capture strategy based on fear-mongering.” This take was controversial at the time; now look how many people are saying it.David Sacks @DavidSacksAnthropic is running a sophisticated regulatory capture strategy based on fear-mongering. It is principally responsible for the state regulatory frenzy that is damaging the startup ecosystem.5:27 AM · Jun 11, 2026 · 634K Views267 Replies · 658 Reposts · 7.91K LikesThe argument continues that even if those regulations never go into effect, the case that Anthropic is making amounts to exceptional self-promotion. Anthropic’s ever-more-dire warnings serve as the best marketing money can buy (certainly better than their Super Bowl ad). Every company in the world wants to get its hands on models that are so powerful that they could crumble global security, create bioweapons, and build rogue superintelligence.On one level, I understand the cynicism here: regulations often end up helping incumbents and there’s a long history of big companies lobbying to make the rules too complicated for new entrants. Facebook once ran the “regulate us, please” playbook in what felt like a calculated effort to push the responsibility for social media’s shortcomings onto legislators who were unequipped to resolve them. (The argument can have purchase even in pro-regulation leftwing and media circles because it fits nicely with the view that all companies are evil and trying to manipulate you.) But there’s such a long record of Anthropic sincerity here. This is a company born out of safety concerns, incubated in the AI-paranoid halls of effective altruism. These are the people who were worried about the release of ChatGPT. They’ve perhaps been over-cautious and sometimes channel a little too much science fiction. But all evidence suggests that they’re on the level when they say they’re very concerned that foundation models could someday soon run amok if we don’t regulate them. It’s endlessly amusing that Anthropic’s do-gooder, pro-regulation stance has actually been good business. While an OpenAI founder-funded super PAC is turning off the company’s employees, Anthropic is recruiting and retaining the best AI researchers by supporting regulation, welcoming criticism, and inviting scrutiny. It’s got to be infuriating for Sacks types that being so lib coded is apparently smart business in a world where winning over AI researchers is the biggest driver of their financial success. Praise be to capitalism.On the substance of Anthropic’s safety position, it really feels like a matter of when not if. My personal view is informed by the self-driving car race where those researchers convinced themselves (and the world) that their technology would be ubiquitous any day now. Of course, it turned out that all the niggling final little details mattered quite a bit and it would take many years to figure out how to make it all work without human supervision. Today, with Waymo’s recent successes, a global society-altering self-driving car rollout seems more and more likely. So what if they got us all excited about a world historical idea a decade too soon? Anthropic’s concerns here — about cyber security, bioweapons, government surveillance, autonomous weapons, and more — seem like things we should get ahead of. Worst case we’re prepared when these problems emerge a few years later than true-believer researchers thought they would. Anthropic is proposing limiting regulations to “models trained using more than 10²⁵ floating-point operations (FLOPs), developed by companies earning more than $500M in AI-related revenue or spending more than $1 billion on AI R&D.” So it’s not exactly mom-and-pop foundation model companies that would be caught in the dragnet of big regulation.While much of the AI industry wants to figure out why Americans are hostile to AI, Amodei doesn’t believe Americans are getting it wrong (except maybe on data centers where their anger is “largely a symbol or outlet for broader economic anxieties about AI”). He writes, “People are worried about AI because they correctly perceive that its risks are real.” He continues, “The key challenge is focusing this concern into constructive solutions and not allowing it to descend into formless anger and violence.”Let’s not let cynicism about big business or smug hostility to regulations blind us to real leadership when it’s staring us in the face. Newcomer PodcastIPO WatchOf all the otherwise accepted business truths that Elon Musk challenges, one of the most fascinating is the Efficient Market Hypothesis.The theory holds that stock prices reflect all available information — that the market, in aggregate, gets valuations roughly right. Skeptics argue markets are also moved by intangibles like momentum and vibes. With SpaceX beginning its trading today, the experiment is about to run again.So far things have been going well. The company raised $75 billion, making it the largest IPO of all time. It’s a wide collection of investors in the round, which Bloomberg reported was four times oversubscribed. BlackRock ordered at least $5 billion worth of shares, per the Wall Street Journal — other buyers included sovereign wealth funds, family offices, and retail investors who reportedly requested more than $100 billion.The longer term story will be trickier. Investors are, largely on faith, accepting a $1.8 trillion valuation for a company with revenue roughly the size of Cheerios-maker General Mills. The most outlandish case has been from Morgan Stanley, which told investors it projects the company will generate $3.4 trillion in revenue and $2.7 trillion in EBITDA by 2040, according to the Wall Street Journal. That’s an 80% margin — software economics, for a business whose plan involves launching its data centers into orbit on rockets.(We’ve written about how farfetched the orbital data center plan is, although the company believes it will have test units up in 2027, according to Reuters.)It’s easy to laugh at these projections for being the pinnacle of Muskian absurdity — the Vibes Market Hypothesis. But there are much larger implications for how the market reacts to SpaceX.If the expected growth in SpaceX’s future is coming from its AI business, then it will also serve as a preview of how OpenAI and Anthropic will fare with public investors. True, both of those have better AI models, less reputational overhang, and more revenue than xAI, but all of them are relying on an unprecedentedly huge market for AI overall to justify their valuations. The SpaceX bet isn’t just that xAI will catch up, but that what they’re all racing toward is as massive a shift in the economy as the tech industry has been predicting.Fund manager and longtime Musk-bear Jim Chanos threw cold water on the whole idea, telling the New York Times that the SpaceX IPO feels like a “don’t look at the man behind the curtain” situation. It’s an ironic statement considering that the man behind the curtain is the reason people believe it’s possible at all.Today’s public listing and the stock fluctuations to follow will make for an entertaining, if terrifying, watch — and a test not just of the efficient market hypothesis and the limits of Musk’s storytelling power, but of whether the public markets actually believe the AI future the entire industry is selling.Eight Notable DealsEnterprise AI automation startup Poetic raised $50 million in Series A funding led by Kleiner Perkins. Other investors include Founders Fund, First Harmonic, and OpenAI.