Everything About AI
NextFin News - Anthropic PBC has lodged an Australian prospectus with the nation's corporate regulator, opening its initial public offering to retail investors in a market starved for direct exposure to a frontier artificial-intelligence company, as the Claude chatbot maker targets a valuation of up to $1.4 trillion. The offer, which opened Wednesday through participating brokers, is expected to close on Nov. 17, with shares set to begin trading on Nasdaq on Nov. 20 under the ticker ANT. The filing marks the final regulatory step before one of the largest-ever technology floats reaches Australian households, giving local traders a seat at a deal that has drawn scrutiny over whether AI capital spending can justify record valuations.
The Australian channel is one of a small set of international retail on-ramps - alongside Japan, Canada, parts of Europe and the United Kingdom - for a company that had grown almost entirely behind closed doors. It also arrives in a bruised local market: Australia's $4.8 trillion superannuation industry and its retail investors have had almost no way to own a hyperscaler or frontier model developer, and the recent collapse of data-centre builder Firmus's planned A$7.1 billion float has left sentiment fragile. Anthropic is betting that a product customers are already paying for is a different proposition from a capital-intensive expansion story leaning on debt.
The Deal: Numbers Behind the Float
Anthropic's IPO terms, set out in its amended U.S. registration filing, are concrete enough to test. The company and certain selling shareholders plan to offer 825 million Class A shares at a proposed price of $70 to $78 each. At the top of that range, the base public transaction would raise about $64.35 billion before fees. On a fully diluted basis - counting employee options and other equity-linked claims - the offer implies a valuation of roughly $1.3 trillion to $1.45 trillion, a figure some reports have rounded to $1.4 trillion. That would sit above the roughly $965 billion valuation from a May funding round, but well below the $2 trillion some investors had discussed earlier in the year.
The structure matters. Of the 825 million public shares, about 801 million are newly issued by Anthropic itself, while roughly 24 million come from existing shareholders. Underwriters also hold an option for up to about 124 million additional new shares. The dual-class architecture means the headline valuation depends heavily on which share count is used: one widely cited calculation using expected outstanding shares put the top-of-range value near $1.02 trillion, while fully diluted measures run higher. Investors are not buying a clean claim on future profits; they are buying one slice of a capital structure that also includes non-voting Class N shares carrying economic rights.
Anthropic is not seeking a listing on the Australian Securities Exchange. CommSec, the Commonwealth Bank of Australia's broking arm, said it will act as lead Australian retail broker for the local offer, while Macquarie's capital-raising unit is expected to lead the Australian institutional book among the 22 underwriters on the deal. The Australian retail offer is being distributed through eight intermediaries: Bell Potter, CommSec, LGT Wealth, Macquarie, Morgans, Morgan Stanley Wealth Management, Sharesies and Stake.
Why Australia Matters: A Market Short of AI
The Australian angle is a story about scarcity. Compared with the technology-heavy markets of North Asia and the United States, Australia has almost no semiconductor or hyperscaler exposure. The shortage of local proxies for the AI buildout has been acute enough to draw retail demand to offshore offerings: CommSec handled SpaceX's Australian retail marketing earlier this year and said it received more than 28,000 applications, a record for the brokerage.
The timing is delicate. Firmus, an AI data-centre developer, scrapped its planned A$7.1 billion ASX float in October after demand from international investors failed to materialize, a blow to a market already grappling with a shrinking listings base. The ASX had 1,891 companies listed in September 2026, down from 2,066 a decade earlier, after a wave of privatizations and a weak pipeline of new issues. Against that backdrop, Anthropic's arrival gives Australian investors something the local market has rarely offered: a direct stake in a company that builds frontier models rather than merely uses them.
"Anthropic is a very different proposition to Firmus. Anthropic is selling a product that customers are already paying for and only increasing their spending on," said Cameron Gleeson, head of investment strategy at Betashares.
Josh Gilbert, lead analyst for Asia Pacific at eToro in Sydney, said the offering "should draw strong local interest after years of the biggest names in AI growing behind closed doors." The distinction between a revenue-generating model lab and a balance-sheet-heavy infrastructure story is the core of the local sales pitch - and the core of the risk, because Anthropic's own spending commitments are large enough that the two narratives can blur.
The Counterweight: $895 Billion in Commitments
The same filing that powers the growth story also carries its heaviest warning. Anthropic disclosed aggregate committed spending of $895.4 billion, primarily for compute and infrastructure. Of that total, $69.8 billion falls due within the 12 months after September 30, 2026, and about $447.5 billion is scheduled before the end of 2030, with the remainder extending to 2045. Nearly half of the commitments are expected to be incurred before December 31, 2030.
Against those obligations, Anthropic reported revenue of $33.2 billion for the first nine months of 2026, up from $2.5 billion a year earlier - growth of more than 1,200 percent. Its operating loss narrowed to $1.3 billion in the period from $5.2 billion in the same nine months of 2025. The company said it held $146.6 billion in liquidity as of late September. The arithmetic is unforgiving: even the full $64.35 billion IPO proceeds, and even combined with a reported $26 billion private placement from seven chip and cloud companies, would not cover the committed spend on their own. The deal funds the next leg of the buildout; it does not retire the obligation.
That gap is where the valuation debate lives. A $1.4 trillion equity value asks investors to underwrite years of spending that must eventually convert into revenue growth that has, so far, been spectacular but unproven at scale. The commitments are not all fixed cash outflows - a portion can be cancelled or depends on conditions being met, and some are leases not yet commenced - but they are a real claim on future free cash flow.
Is This Structural or Cyclical? A Regime Shift With a Cyclical Entry Point
The central question for investors is whether frontier AI is a structural shift or a cyclical capital-spending boom. The answer is both, and the distinction determines who wins and who gets hurt.
The structural leg is real. Frontier model development has shifted the cost base of the software industry from engineering headcount to compute at a scale that did not exist five years ago. Anthropic's own disclosure - nearly $900 billion of commitments - is evidence that the industry has moved into a regime where access to capital and chips is the binding constraint, not ideas. Customers are already paying, and spending is increasing. That is not a speculative demand curve; it is a contracted one. This part of the thesis does not mean-revert on its own.
The cyclical leg is equally real and more dangerous. Capital expenditure in technology has historically overshot demand, then corrected. The risk is not that AI disappears; it is that capacity comes online faster than monetization, margins compress, and the market reprices the multiple before the revenue catches up. Anthropic's operating loss, though narrowing, shows the model is not yet self-funding at current spend rates. If revenue growth slows from its current triple-digit pace while commitments remain fixed, the equity duration stretches and the valuation becomes sensitive to discount rates and sentiment - classic cyclical dynamics layered on top of a structural trend.
The second-order effect is what most investors are not pricing. The obvious read is that a successful IPO validates AI valuations and lifts the sector. The less obvious read is that a $1.4 trillion capitalization converts private AI risk into public-market risk, and in doing so changes the incentives. Once public, Anthropic must report quarterly, and its spending discipline - or lack of it - becomes visible. If the company signals any slowdown in committed spend, the signal would ripple through the entire AI supply chain: chip makers, cloud providers, power and data-centre operators. The IPO is not just an exit for early investors; it is the moment the AI capex cycle becomes auditable.
The Adversarial Case
The strongest argument against the bull case is simple: at $1.4 trillion, almost all of the good news is already in the price. The company is being valued on revenue of $33.2 billion for nine months, or roughly $44 billion annualized - a price-to-sales multiple in the low thirties, before accounting for the dilution embedded in the dual-class structure and the $895 billion of future obligations. If revenue growth decelerates to even a still-strong 50 percent while spending commitments hold, the multiple has nowhere to go but down. Bears would argue that Anthropic is not being priced as a software company but as a utility in formation, and that utilities do not command software multiples while they are still borrowing to build.
The bull's answer rests on the private placement and the customer base. Seven of the world's largest chip and cloud companies - Alphabet, Nvidia, Microsoft, AMD, Broadcom, Micron and Taiwan Semiconductor Manufacturing - are reported to be investing $26 billion alongside the IPO. Those are not passive index funds; they are counterparties with direct visibility into AI demand. Their participation is a vote, backed by cash, that the spending will be absorbed. Combined with more than 300,000 business customers and large accounts growing nearly sevenfold year over year, the bull case is that demand is contracted, not hoped for.
The signal that would break the structural thesis is specific and observable: if Anthropic's revenue growth falls below 40 percent year over year for two consecutive quarters while near-term committed spend remains above $60 billion a year, the claim that this is a durable regime shift priced reasonably would be wrong. At that point the story would have flipped from structural adoption to cyclical overbuild, and the $1.4 trillion valuation would need to be re-underwritten on utility-like returns.
What Comes Next
In the short term, the focus is mechanical: whether the Australian offer closes on schedule on Nov. 17, whether the final price lands at or near the top of the $70-$78 range, and how ANT trades in its first sessions on Nasdaq from Nov. 20. Australian retail investors should note that the final price will be determined by the company and underwriters, and applicants remain bound to their allocation once the offer closes.
Over the medium term, the first quarterly reports as a public company will matter more than the listing-day pop. Investors should watch three metrics: revenue growth versus the current run rate, the pace at which committed spend converts into deployed capacity, and whether operating losses continue to narrow. The base case is a successful listing followed by volatile early trading as the market digests the spending schedule. The upside case requires revenue to keep compounding at triple-digit rates while margins improve - a path that would justify a premium multiple. The downside case is a growth deceleration that collides with fixed commitments, forcing a multiple contraction that would hit not just Anthropic but the listed AI ecosystem.
For Australia specifically, the stakes extend beyond one stock. A clean, oversubscribed Anthropic retail offer would restore some confidence after Firmus and signal that global tech giants still see value in accessing Australian capital - both institutional and household. A weak or undersubscribed offer would deepen the ASX's listings problem and push more investors toward offshore platforms.
The central judgment: Anthropic's Australian prospectus is less a gift to retail investors than a test of whether the AI boom can survive the transition from private hype to public accounting. The structural trend is intact, but the entry price assumes the cycle stays benign. This is a market being asked to price a decade of infrastructure today on the strength of a year of revenue growth.
Explore more exclusive insights at nextfin.ai.
