Everything About AI
NextFin News - Anthropic is still moving toward an initial public offering that investors are valuing near $2 trillion, even as its chief executive, Dario Amodei, publicly called on the artificial-intelligence industry to "slow the pace" of model development over safety fears. The company expects to begin marketing the offering in mid-October at the earliest, after filing its prospectus in late September, with a listing that could land days before the U.S. midterm elections in November. The juxtaposition defines the deal: a company asking the world to hit the brakes while its bankers prepare the most expensive equity raise in history.
The Contradiction at the Center of the Deal
On September 12, Amodei published an essay titled "We Must Pace the Frontier" arguing that AI companies must deliberately slow the rate at which they improve model capabilities so that safety work can keep up.
"Over the last few months, I have become convinced that fully addressing the risks requires even more prudence — not just investing in risk prevention, but pacing the rate of capabilities advancement so that risk prevention has time to keep up," Amodei wrote. "We must slow the pace at which we improve the capabilities of AI models. Progress will still seem fast, and we must make wise use of the time we gain."
The essay landed as Anthropic was preparing what could become the largest IPO in history. A $2 trillion listing would surpass the record set in June by Elon Musk's SpaceX, which went public at a $1.77 trillion valuation after raising about $75 billion — roughly three times the previous record held by Saudi Aramco's 2019 debut, which raised more than $29 billion. SpaceX's listing proved that public markets will absorb a near-unimaginable valuation when the growth story is compelling enough. Anthropic is betting the same appetite extends to frontier AI.
The tension is not lost on investors. Anthropic's backers argue that the transparency requirements of being a public company support, rather than undermine, its safety commitments. The company has pledged third-party evaluation of its models and signaled it may slow the release of new models — a stance Amodei framed as a unilateral commitment that he wants governments to require of other frontier developers. The first step of his three-part plan, "embedded evaluators," would give outside monitors employee-like access to verify safety practices and inspect training pipelines, not just finished models.
What the Numbers Say
Anthropic reported more than $47 billion in annualized revenue in May, according to people familiar with the company's finances. Investors are now pricing in annualized revenue of $100 billion to $120 billion by December — growth of more than tenfold from the start of 2026. One investor argued that a company expanding at 800% annually would, at the very low end of reasonable expectations, command a 30-times-revenue multiple, implying a $3 trillion valuation. At the $2 trillion target, the implied multiple on a $110 billion midpoint is roughly 18 times revenue — rich by software standards, but below the 55 times sales at which AI-adjacent names such as Palantir and Nebius have traded this year.
The company last raised capital in late May at a $965 billion post-money valuation, meaning the IPO would roughly double that mark in a matter of months. The funding trail shows how fast the marks have climbed: a Series G-1 round in February priced the company at $380 billion, and the Series H-1 round in May added $41.29 billion at $965 billion. Morgan Stanley, Goldman Sachs, JPMorgan and Citi are among the banks working on the offering, which is also expected to include the finalization of a $15 billion revolving credit facility. The company filed confidentially with the Securities and Exchange Commission in June and has entered a quiet period, declining to comment on the IPO.
The Market's First Read
Wall Street's initial reaction to the safety warnings was swift and one-sided in the hardware complex. On September 14, futures tracking the Nasdaq 100 fell 505.5 points, or 1.72%, while S&P 500 futures dropped 0.70% and Dow futures shed 0.18%. Nvidia fell more than 2% in premarket trading, with Intel, AMD and Marvell Technology each down between 5% and 6%. The chipmakers bore the brunt because a deliberate slowdown in AI development strikes directly at the capital-spending boom that has powered their gains: every month of "pacing" is a month of deferred GPU orders, and the semiconductor complex has been priced for uninterrupted acceleration.
Not every AI-exposed stock fell. Shares of U.S. software companies — which have been pressured by fears that AI could disrupt their businesses — actually climbed, with ServiceNow up 3%, Adobe up 2.5% and Workday up 2.5%. The split captures the two-sided nature of the safety debate: slower AI development hurts the builders of the infrastructure, but it may protect the companies whose business models are under threat from automation. That divergence is the cleanest market signal yet that investors are starting to price AI safety as a redistribution of winners and losers, not just a sector-wide risk.
Why the IPO Cannot Wait
The capital math is the simplest explanation for Anthropic's timing. Frontier AI development is among the most cash-intensive businesses ever created, and private markets, however deep, cannot indefinitely fund the buildout of data centers, chip purchases, and research talent that the company needs to compete with OpenAI, Google, and a well-funded field of Chinese challengers. A public listing converts private marks into liquid currency for employees and early investors, and it opens a permanent tap for equity issuance that private rounds cannot match at the same cost.
There is also a competitive-clock argument. OpenAI is widely expected to pursue its own listing, though sources indicate it is leaning toward 2027 rather than 2026. If Anthropic waits, it cedes the "first major AI lab public" position and the pricing power that comes with it. Going first, at the peak of investor appetite, maximizes the chance of clearing a $2 trillion valuation before safety concerns harden into regulation that could compress multiples across the sector. The window between "extraordinary growth story" and "regulated utility" is narrow, and Anthropic is racing to list inside it.
The Cyclical Fear and the Structural Bet
The central analytical question is whether the safety panic is a cyclical fluctuation that will pass, or a structural shift that changes how the market values AI forever. The evidence points to a cyclical shock layered on top of a structural change — and the distinction determines whether $2 trillion is a fair price or a top.
The cyclical leg is the sentiment wave. History offers three relevant precedents. After the 2000 dot-com crash, internet companies faced a regulatory and reputational reckoning that compressed multiples for years, yet the sector's long-term winners emerged at valuations that looked absurd at their lows. After the 2018 Facebook-Cambridge Analytica scandal, Meta's stock fell sharply on privacy fears before rebounding as advertising revenue proved resilient. And after the 2022 crypto collapses, the industry faced calls for sweeping regulation that ultimately produced a framework the largest players could absorb. In each case, the immediate panic was cyclical — it reverted as earnings proved the business model intact. Anthropic's revenue trajectory, still growing at a triple-digit rate, fits that pattern: if the numbers hold, the safety headlines will read as noise.
The structural leg is different, and it is the real risk to the multiple. Amodei is not merely asking for caution; he is proposing a regime in which frontier AI development becomes a monitored, permissioned activity with embedded third-party evaluators, industry-wide standards, and eventually international coordination. If that regime takes hold, the economics of the industry change permanently. A company that can release a new model every quarter is worth a growth multiple. A company that must clear evaluators, coordinate with rivals, and wait for government sign-off is worth a regulated-utility multiple. The structural shift is not that AI becomes less valuable — it is that the capital intensity and compliance cost become permanent, and the cadence of revenue recognition slows.
The correct call is that the cyclical panic will pass but the structural regime will partially arrive. That combination is the most dangerous for valuation: it is not severe enough to kill the growth story, but it is real enough to cap the multiple. A $2 trillion price assumes the structural regime arrives slowly enough that Anthropic can grow into it. That is a bet, not a base case.
The Counter-Thesis: Slowing Down Is a Moat, Not a Concession
The strongest case against reading Amodei's essay as a threat to the IPO is that it may be the opposite: a strategic move to lock in a regulatory moat. By volunteering for third-party monitoring and calling for government-mandated safety standards, Anthropic raises the compliance cost of competing. Smaller rivals and open-source developers would face higher barriers to entry, while Anthropic — with its deep-pocketed backers including Amazon, Google, and Microsoft — is best positioned to absorb the cost.
Critics have made this point explicitly. Chamath Palihapitiya, an investor and co-host of the "All-In" podcast, wrote that Amodei "makes the case to stop open source and concentrate enormous technological and economic power with Anthropic." Under this reading, the safety slowdown is not a brake on growth but a way of rigging the race so that the incumbent with the most capital and the strongest safety brand wins it. The essay's own language feeds this interpretation: Amodei calls for export controls on advanced chips and a crackdown on model distillation, measures that would slow China's progress "enough to widen America's lead significantly over the next 3–5 years" — a window he describes as the period when AI becomes geopolitically most important.
"I believe that if slowing down bought us even an extra year or two before models reach critical levels of capability, and we used that time to advance alignment, we could greatly reduce the risk that something goes seriously wrong," Amodei said.
That interpretation is plausible, but it does not fully resolve the valuation risk. Even if regulation entrenches Anthropic's position, it also caps the growth rate that the $2 trillion price tag assumes. A company growing at 800% annually is priced for a land grab; a company growing at 50% under a compliance regime is priced like a mature software vendor. The IPO can only clear at $2 trillion if investors believe the safety slowdown buys time without killing the revenue curve. And there are already cracks in that curve: revenue growth slowed in June after the U.S. Commerce Department imposed a temporary export control on Anthropic's best models, and data from payments company Ramp showed businesses reaching the ceiling of what they are willing to spend on AI. The company also remains in active litigation with the Department of Defense, which designated Anthropic a supply-chain risk earlier this year — a disclosure that will sit prominently in any prospectus. Beijing has also rejected the slowdown framework outright: China's Foreign Ministry called the proposal "fear-mongering," and state media described it as a "Cold War playbook" aimed at containing China.
What Would Prove This Wrong
The falsifying signal is concrete: if Anthropic's prospectus shows annualized revenue growth slowing materially below the $100 billion to $120 billion year-end range that investors are underwriting — specifically, if the disclosed run-rate comes in under $90 billion — the $2 trillion valuation thesis breaks. A second signal would be a coordinated international slowdown that includes China. Amodei himself flagged Beijing as the constraint that could force the U.S. to keep racing; China's Foreign Ministry has already dismissed the slowdown proposal as "fear-mongering." If Washington and Beijing fail to coordinate, the "pace the frontier" framework collapses back into a race, and the safety premium embedded in Anthropic's multiple evaporates. A third signal: if embedded evaluators gain real blocking authority over model releases, the structural-regime thesis is confirmed and the multiple should compress toward regulated-utility levels.
What to Watch
The immediate catalyst is the prospectus filing, expected in late September, followed by analyst meetings and a roadshow in mid-October. Key items to watch: the revenue run-rate disclosed for the trailing twelve months, the customer concentration among hyperscalers, the status of the Department of Defense supply-chain litigation, and any language around model-release pacing commitments. The $15 billion revolving credit facility, expected to close before analyst meetings, will also signal how much debt capacity the company needs ahead of the equity raise.
Short term, the equity market will keep pricing AI infrastructure on the margin of safety news — every warning from a lab executive compresses chip multiples, and every sign of regulatory clarity supports them. The September 14 move, with chipmakers down 5% to 6% while software names rose, is the template. Medium term, the question is whether public-market investors will tolerate the capital intensity of frontier AI without quarterly profitability; private markets can wait for inflection, public markets count quarters. Long term, the structural question is whether AI development becomes a regulated, utility-like industry, in which case today's growth multiples are unsustainable regardless of who wins the race.
The base case is that Anthropic lists in late October or early November near the $2 trillion mark, backed by a revenue story that still looks explosive on paper and a first-mover premium over OpenAI's expected 2027 timeline. The downside case is a delayed filing and a valuation closer to the $965 billion private mark if safety fears trigger a broader AI multiple compression before the roadshow. The upside case is a blowout debut that cements Anthropic as the first trillion-dollar-scale AI lab to go public, pulling OpenAI and others into the queue behind it.
The real story is not whether Anthropic lists — it needs the capital too much to stay private. The real story is whether a company can sell investors on unlimited growth while its CEO tells the world the technology needs to slow down. For now, the market is willing to believe both, which is exactly the kind of contradiction that ends badly for late buyers — or pays spectacularly for early ones.
Explore more exclusive insights at nextfin.ai.
