NextFin News - The race to list the world's two most valuable artificial-intelligence companies sharpened this week as Allie K. Miller, founder and chief executive of AI advisory firm Open Machine, weighed in on the prospects for Anthropic and OpenAI going public, with the two rivals carrying a combined pipeline that could test whether investor appetite for frontier AI survives the transition from private hype to public scrutiny.
Miller's appearance on a major business-television program on August 20, 2026, landed at a moment when the IPO queue for AI giants has moved from speculation to active preparation. Anthropic PBC confidentially filed for a US listing on June 1, 2026, and OpenAI followed with its own confidential filing a week later on June 8, 2026, joining Elon Musk's SpaceX in what has been described as a roughly $3.6 trillion pipeline of AI-related listings.
The numbers behind the filings are what make the moment unusual. Anthropic told prospective investors its second-quarter revenue jumped at least 14-fold from a year earlier, and its annualized revenue run rate crossed $47 billion in May. By the end of July, that run rate had climbed past $65 billion, a more than sevenfold increase from the pace at the end of 2025, according to people familiar with the company's investor updates. OpenAI, by comparison, is running at an annualized revenue pace above $40 billion, roughly double its run rate at the end of 2025.
Valuations have raced ahead of those revenue figures. Anthropic raised $65 billion in a May 2026 funding round that valued the Claude maker at about $965 billion. OpenAI raised $122 billion in March 2026, led by Nvidia and Amazon, at a valuation near $850 billion. SpaceX, the first of the trio to list, priced its June 2026 IPO at $135 a share, valuing the company at roughly $1.77 trillion and raising $75 billion — the largest IPO in history.
The market has not greeted that debut with unalloyed enthusiasm. SpaceX shares have fallen about 18 percent below their offer price in the weeks since, putting investors who bought at the offering price underwater for the first time and forcing a conversation about what price public investors will actually pay for AI promises that private markets priced at scarcity.
Against that backdrop, Miller's comments matter because Open Machine sits close to the companies in question. Her advisory firm, founded in 2022, counts OpenAI, Anthropic, Google, Samsung, Salesforce, and Novartis among its clients, and she previously ran machine-learning product and startup efforts at Amazon Web Services and IBM. She is not an underwriter, and she does not set listing timetables — but she is one of the outside voices who speaks regularly with the leadership of both AI labs about how their technology is being deployed inside enterprises, which is ultimately what will decide whether the revenue growth holds up after the lockup expires.
The central tension is straightforward: Anthropic and OpenAI are both racing to show investors that their growth is durable enough to justify trillion-dollar valuations in a public market that has just watched SpaceX give back its first-day gains. The question is whether going public sooner rather than later is an advantage — or a trap.
The Race Is Real, and Both Companies Are Running
The framing of an Anthropic-versus-OpenAI sprint is accurate, and it has only intensified. Anthropic filed first. Its revenue disclosures have been more aggressive. It has pointed to a 2028 break-even target, two years ahead of OpenAI's 2030 profitability goal, a distinction that public-market investors are likely to price as a premium.
But the idea that OpenAI was content to wait has been overtaken by events. Recent reporting indicates OpenAI is laying the groundwork for a public listing in the fourth quarter of this year, accelerating its plans as competition with Anthropic intensifies. The company is holding informal talks with Wall Street banks about a potential offering and has been growing its finance team, including the hire of a new chief accounting officer and a corporate business finance officer to oversee investor relations.
That sequencing matters more than it appears. The first of the two to list becomes the public-market template for how the sector is valued — its price-to-sales multiple, the growth rate investors demand, the tolerance for losses. Anthropic going first at a $1 trillion-plus valuation would set a high anchor for OpenAI. OpenAI going first, or going at a lower multiple, would do the reverse. The company that lists second gets to watch the market's verdict before committing its own valuation to the tape.
There is a reason the race has accelerated on both sides. OpenAI's March $122 billion raise gives it a long runway, but its second-quarter results showed revenue of $6.7 billion, up from $5.7 billion in the first quarter, with its operating margin sinking further into the red — results that disappointed some shareholders who had hoped for more progress catching up to Anthropic. Anthropic's $65 billion raise in May is substantial, but its infrastructure spending commitments — including a reported $35 billion financing package with Apollo Global Management and Blackstone to expand compute capacity — mean the clock is loud on both sides. A company burning cash on data centers has more incentive to access public capital sooner, even if the window is imperfect.
What the Market Is Pricing, and What It Is Not
The market has priced the growth. Anthropic's 14-fold revenue jump and $65 billion run rate are not in dispute. What has not been priced is durability.
The first-order read of the AI IPO wave is simple: these companies are growing fast, they dominate the frontier-model race, and enterprise adoption is still early. That is the pitch in the roadshow. The second-order question is whether the revenue is being bought — literally. A meaningful share of AI revenue flows from companies that are themselves funded by the same pool of venture and growth capital chasing AI returns. When capital tightens, the customers tighten, and the revenue that looked exponential on the way up can mean-revert faster than the headcount and compute commitments built to serve it.
This is where the SpaceX post-IPO action is instructive. SpaceX raised $75 billion at a $1.77 trillion valuation on the strength of Starlink profitability and AI ambitions. Within weeks, the shares were down roughly 18 percent from the offer price. The company was not exposed to an AI revenue slowdown in the way Anthropic or OpenAI would be — it is a hardware and launch business with a profitable satellite constellation. If the market is willing to mark down a cash-generative space franchise on valuation concerns alone, the tolerance for a loss-making AI lab trading at 20 times sales is lower still.
The transmission channel runs through the discount rate and the multiple. A public listing forces a quarterly cadence on companies whose economics are built on multi-year compute buildouts. Capex that private investors underwrote as strategic becomes a margin question every three months. Gross margins in frontier AI are pressured by inference costs and by the pricing competition that comes when three well-funded labs chase the same enterprise contracts. The market does not need the revenue story to break to reprice the stock — it only needs the growth rate to decelerate from "exponential" to "very fast."
Cyclical or Structural: The Call That Decides the Trade
This is where the analysis has to make a call, because the two forces are pulling in opposite directions and blending them produces a muddy verdict.
The cyclical leg is clear and it is mean-reverting. The IPO window itself is cyclical: US IPO proceeds reached $141.2 billion through mid-July 2026, within striking distance of the 2021 full-year record of $142.4 billion, and the second quarter alone raised a record $104.8 billion across 48 deals. Windows like this close. They closed in 2022, and they will close again when liquidity conditions tighten, when a flagship post-IPO name breaks its offer price, or when the Federal Reserve's path diverges from what risk assets have priced. The SpaceX pullback is the first warning sign that the window is not infinitely wide.
The structural leg is equally real and it does not mean-revert on its own. Enterprise AI adoption is a regime shift in how knowledge work is performed. Miller's own observation from her advisory work — that companies are moving from experimenting with AI tools to building systems where AI agents initiate work rather than waiting for prompts — describes a structural change in software deployment, not a sentiment cycle. If Anthropic and OpenAI can convert that adoption into recurring enterprise contracts with high retention, the revenue base is durable regardless of the listing window.
The correct read is that the listing window is cyclical but the underlying demand is structural — and that mismatch is exactly why the sequencing decision is so consequential. A company that lists into a closing window gets punished for a cyclical problem it cannot control. A company that waits too long risks missing the peak valuation even if its business is structurally sound. The optimal move is to list when the company's own metrics are strong enough to survive a window closure six months later.
The Counter-Thesis, and What Would Prove It Wrong
The strongest case against the cautionary read is that the market has already digested this. The SpaceX IPO did not break the bull market; the float was too small to distort liquidity in a meaningful way, and the pullback has been framed by strategists as a stock-specific valuation debate rather than a macro signal. By that logic, the AI IPOs are simply the next tranche of a reopening that has been underway since early 2026, and investors who sat out SpaceX will not sit out Anthropic.
That argument has force, but it rests on a comparison that does not hold. SpaceX came to market with a profitable core business and a clear path to cash flow. Anthropic and OpenAI are listing on revenue growth with losses that will persist for years. Apollo Global Management's chief economist, Torsten Slok, has cautioned investors to approach this year's mega-IPOs carefully, pointing to data showing that IPOs since 2019 have underperformed the broader market in the years after listing. Since 2019, he wrote, the market regime influencing post-IPO performance has been characterized by three things: "peak valuations," "a hostile rate regime," and "low quality, high bar."
"The boom pushed marginal companies public before they were ready while the market-adjusted benchmark was set against an index carried by a handful of mega-cap winners," Slok wrote. "Valuations may re-inflate in the next IPO window, rates look set to stay structurally higher than the 2010s and index returns remain concentrated in a few mega-caps that keep the relative bar high."
The falsifying signal is specific: if Anthropic prices its IPO above a $1 trillion valuation and the shares are trading above the offer price 90 days after listing, the cautionary thesis is wrong — the public market has accepted AI-lab economics at scale, and OpenAI's accelerated timing would look prescient. Conversely, if the first AI lab to list trades below its offer price after 90 days, the window for the second narrows sharply, and the company that waited gains a real advantage.
Conclusion and Outlook
The mechanism cashes out into a clear asymmetry. The beneficiaries of a successful Anthropic or OpenAI listing are the late-stage private shareholders across the AI stack — Nvidia, the cloud providers, and the venture funds holding paper gains — because a clean public comp validates the entire private valuation chain. The exposed parties are the companies that list second into a window that has already been tested, and the public-market investors who buy the first-day pop without a quarter of public financials behind them.
Split by time horizon:
- Short term (sentiment and liquidity): the IPO announcements and any pricing news will support AI-related equities and the brokers and exchanges that service listings. But the SpaceX post-IPO weakness is a live reminder that the window can turn on a single flagship name.
- Medium term (fundamentals): the first two quarters of public results from whichever lab lists first will set the multiple for the entire sector. Gross margin trajectory and enterprise retention will matter more than the revenue growth rate, because growth is already priced.
- Long term (structural): the company that converts enterprise AI adoption into durable recurring revenue wins regardless of listing order. The structural demand for frontier models is not in question; the question is who captures the economics after compute costs and price competition are accounted for.
Scenarios:
- Base case: Anthropic lists in late 2026 or early 2027 at a valuation between $1 trillion and $1.5 trillion, trades flat to modestly higher in the first quarter, and OpenAI follows soon after once the template is set.
- Upside case: the first AI lab to list prices above $1.5 trillion and holds its gains, reopening the window wide enough for the second company and Alphabet to price aggressively.
- Downside case: the first AI lab to list breaks its offer price within 90 days, the window narrows, and the second company delays its IPO until market conditions improve.
What to watch: the IPO pricing of whichever lab files its public S-1 first, the 90-day post-listing price relative to the offer, and the gross-margin trajectory disclosed in the first two public quarters. Those three data points will tell investors more about AI-lab economics than a year of private revenue disclosures.
This is not a race to the public market — it is a race to prove that AI revenue is durable enough to survive it. The company that lists first sets the template; the company that lists second gets to learn from it. In a window that has already shown it can close quickly, patience is not weakness. It is the only leverage a private company has left once the bankers are in the room.
Explore more exclusive insights at nextfin.ai.
