Everything About AI
NextFin News - Anthropic is preparing a stock market listing that investors and bankers have discussed at a valuation near $2 trillion, with a launch that could come as soon as October, and its public debut is expected to lead the traditional year-end rush of initial public offerings. The question is not whether the artificial-intelligence names want to list. It is whether public markets can absorb a cluster of the largest private-company valuations in history without repeating the post-debut fade that has already hit the year's biggest offering.
The Window Opens on the Deepest Pipeline in a Generation
The Labor Day holiday marked the end of summer in the United States and, with it, the traditional start of the push to year-end in the IPO market. Anthropic PBC looms largest in that window. The company filed a confidential draft registration statement with U.S. securities regulators on June 1 and has been working with Morgan Stanley and Goldman Sachs, among others, on a Nasdaq debut that could arrive as early as October.
The scale is what makes this different from a normal IPO season. Anthropic's last private funding round, a $65 billion raise closed May 28 at a $965 billion post-money valuation, more than doubled the $380 billion tag attached to its $30 billion Series G round just three months earlier, in February. The company said its annualized revenue run rate exceeded $47 billion in May, a pace that overtook OpenAI, and by the end of July that run rate had reached $65 billion, according to reporting on the company's investor updates. A $2 trillion public valuation would be roughly double its last private mark and would make the offering the largest public listing ever, ahead of SpaceX's June debut.
Anthropic is not alone. A clutch of other artificial-intelligence and technology companies are headed for the public markets in the same window, and the pipeline behind them is the deepest in a generation. Five private companies — SpaceX, OpenAI, Anthropic, Stripe and Databricks — collectively represent more than $3 trillion in private-market value. SpaceX already went public, pricing on June 3 at $135 a share, raising about $75 billion and debuting on June 12 at a $1.77 trillion valuation, the largest IPO in history. OpenAI filed confidentially a week after Anthropic, on June 8, but is now leaning toward waiting until 2027. Databricks, valued at $190 billion after a $5 billion round on August 13, has called 2026 a "terrible year to go public." Stripe, valued at $159 billion in a February employee tender offer, has not filed and is seen in a 2027-2028 window.
The first half of 2026 set the stage. Traditional IPOs raised approximately $114.1 billion through June 30, more than seven times the $14.8 billion raised over the same period in 2025, according to a second-quarter capital-markets review from PwC. About 97% of the year's offerings opened above their offer price, and nearly half priced at or above the top end of their marketed ranges. Yet the concentration of value in a handful of AI names raises a question the market has not had to answer in years: can it digest several trillion-dollar-scale listings in quick succession?
Stable Rates Open the Window, but Concentration Narrows It
The immediate backdrop favors sellers. The Federal Reserve is expected to hold its benchmark rate in the 3.50%-3.75% range at its September 15-16 meeting, with about 70% of economists in a September survey expecting no change for the rest of the year. A stable rate path matters for IPOs because it anchors the discount rates that growth valuations depend on, and it keeps the risk premium investors demand for illiquid new issues from widening.
But stability is not the same as abundance of appetite. The risk is concentration. In the first quarter of 2026, unicorn deal activity reached $246 billion across 227 transactions, and five companies — OpenAI, Anthropic, xAI, Waymo and Nscale — accounted for 78% of all deal value, while the top 10 captured 81%, according to Russell Investments. The same concentration is now arriving in the public market. When new issuance is this top-heavy, the buyers are not diversified across sectors; they are the same funds that already own mega-cap technology and AI infrastructure.
Investors who buy a new IPO may fund that purchase by deploying cash, but they may also sell existing holdings. If several very large transactions arrive close together, that "making room" process may add to volatility.
That observation, from LPL Research in a commentary on the 2026 IPO class, captures the transmission mechanism. A large IPO does not just take money from cash balances; it can force rotation out of existing holdings. If Anthropic, OpenAI and the others list within months of one another, the "making room" effect compounds.
SpaceX's Debut Is the Stress Test the Market Is Watching
SpaceX provided the first real data point. Its June 3 pricing at $135 a share, across 555.6 million shares, valued the company at $1.77 trillion and raised roughly $75 billion, about three times the previous record held by Saudi Aramco's 2019 offering.
At $74.4 billion raised, SpaceX's I.P.O. would be just about "more than every U.S. I.P.O. combined in the last two years," said Matthew Kennedy, a senior I.P.O. market strategist at Renaissance Capital.
The follow-through is the lesson. SpaceX began trading on June 12 and closed its first day at $161.11, up 19.3%, after reaching an intraday high of $225.64 on June 16 that briefly pushed its market value toward $3 trillion. It then corrected roughly 35%, hitting a low of $145.20 in early July before rebounding to about $153. The slide did not reflect a change in the company's business. It reflected the price at which private-market conviction met public-market scrutiny.
That matters for Anthropic because the $2 trillion discussion mark is already more than double its last private valuation. SpaceX showed that even a record-breaking, category-defining offering can fade once the initial allocation scramble ends. For Anthropic, the bar for a clean debut is higher than the bar for a successful book build.
OpenAI's Hesitation Is a Signal, Not an Outlier
OpenAI's internal deliberations are the clearest evidence that the window has friction. The company filed confidentially on June 8, with Goldman Sachs and Morgan Stanley leading the process, and initially targeted a listing as soon as the third or fourth quarter of this year. Executives then shifted toward 2027. Advisers presented two options: wait until 2027 for a $1 trillion valuation, or accept a lower valuation for a faster listing. CEO Sam Altman reportedly told advisers that any reduction in the trillion-dollar target was a non-starter. The company said it had "not decided on timing yet" and that the filing "gives us the option to go public sooner if that ends up being best."
The company's own numbers help explain the caution. OpenAI reported roughly $13 billion in revenue for 2025 on a loss of about $21 billion, according to financial statements that circulated in June, and it has targeted roughly $600 billion in total compute spending through 2030. Cumulative cash burn is projected at more than $100 billion through 2029. A public market that has to underwrite that loss trajectory quarter by quarter is a different buyer than a private fund writing a growth check.
Databricks reached the same conclusion from a different angle. The data-and-AI platform reported a $7 billion annualized revenue run rate in the second quarter, up more than 80% year over year, and closed a $5 billion round at a $190 billion valuation on August 13. Yet CEO Ali Ghodsi described 2026 as a "terrible year to go public," citing a calendar crowded by SpaceX's blockbuster IPO and anticipated offerings from Anthropic and OpenAI.
"I don't want to also take the company public, and then 2022 happens, and then the market wants to have 30% EBITDA, and we can't invest in these."
Ghodsi said in a June interview, explaining why waiting can preserve growth investment. If a fast-growing data company with a clear enterprise footprint is willing to wait, it is a signal about supply, not about quality.
Cyclical Window, Structural Concentration: Keep Them Separate
This is where the analysis has to separate two forces that are easy to blend. The reopening of the IPO window is cyclical. It is driven by stable rates, a constructive risk backdrop, and the calendar effect that makes the period between Labor Day and year-end the traditional listing season. Cyclical windows close. When rates move, when volatility rises, when a few high-profile debuts disappoint, the window narrows quickly. The 2021 boom, which produced more than $123 billion of proceeds across more than 300 venture-backed IPOs, is the comparison point — and the hangover after it kept most unicorns private for years.
The concentration of value in AI, by contrast, is structural. Private capital has been funneled into a small set of frontier AI and infrastructure companies because the business model itself is capital-intensive: compute, power, talent and distribution at a scale that only a handful of players can reach. That is not a sentiment cycle; it is an industry structure. Those companies will list eventually because their investors need an exit and their balance sheets need public currency. The structural leg does not revert on its own.
The practical implication is that the wave is real, but it will be bifurcated. The top of the pipeline — Anthropic, and eventually OpenAI — has the brand, the revenue and the investor base to clear a crowded window. The second tier — Databricks, Stripe, and the AI-infrastructure names that followed Cerebras — will price only if the first wave trades well. A strong Anthropic debut pulls the rest of the calendar forward; a weak one pushes it into 2027.
The Second-Order Effect: Public Markets Become the Exit for Private AI Valuations
The first-order read of this wave is straightforward: more AI companies will list, and their stocks will move. The second-order effect is what matters more. The private-market valuations built over the past two years — Anthropic at $965 billion, OpenAI at $852 billion, Databricks at $190 billion, Stripe at $159 billion — were set by funds that could underwrite long loss runways and multi-year capex plans without quarterly mark-to-market. The public market marks to market every day.
When those two pricing regimes meet, one of two things happens. Either the public market accepts the private marks and the AI complex is re-rated upward as a public asset class, or the private marks get marked down to meet public reality. Cerebras gave an early read: it priced at $185 on May 14, raised $5.55 billion in the largest U.S. technology IPO since Uber in 2019, and closed its first day at $331.07, up 68%, worth about $95 billion. The next day it fell roughly 10%. That is not a failure; it is price discovery working faster than private rounds ever allowed.
CoreWeave, which listed in March 2025 at $40 a share and raised $1.5 billion at a $23 billion fully diluted market cap, shows the other path: its shares peaked at $183.58 in June 2025 and trade well above the offer price, suggesting that AI-infrastructure names with contracted demand can sustain a public re-rating. The difference between the two outcomes is not the theme. It is the visibility of revenue against the capex bill.
There is also a third-order supply overhang that most investors are not pricing. Standard IPO lockups expire about 180 days after listing. A busy pricing period now means a wave of lockup expirations lands roughly six months out, in the first half of 2027. For a market already defined by high concentration and heavy passive ownership, that is an incremental supply event that arrives just as the second wave of listings would be hitting the tape.
The Counter-Thesis: A Late-Cycle Warning, Not an Opportunity
The strongest case against the wave thesis comes from the record itself. US IPO issuance has surged to a record high in 2026, and Goldman Sachs Research has asked whether this is a late-cycle warning sign and whether the market can comfortably digest so much new supply. History is not kind to IPO booms. The largest waves — 1999, 2021 — peaked just before equity markets rolled over, because the window opens widest when private owners can extract the maximum price from public optimism.
The SpaceX drawdown is the fresh evidence. A $1.77 trillion debut that gives back more than 30% within weeks is exactly the pattern a late-cycle reading would predict: scarcity value at the offering, mean reversion in the aftermarket. If Anthropic prices near $2 trillion on a revenue run rate that public investors will scrutinize every quarter, the same dynamic could repeat at a larger scale. The crowded calendar is not a sign of health; it is a sign that every private owner is trying to exit through the same door at the same time.
This counter-thesis attacks the core of the wave argument: that demand is durable rather than finite. It is backed by the historical pattern that issuance peaks precede market peaks, and by the observable fact that 2026's deal value is concentrated in a handful of AI names whose valuations were set in a private market with no mark-to-market discipline.
The answer is that the counter-thesis is right about the risk but too broad about the timing. Late-cycle warnings can stay right for years while the window stays open in the near term. The 2026 IPO market is not broad-based froth; it is a narrow queue of genuinely scaled companies with real revenue. Anthropic's $47 billion-plus run rate and OpenAI's $13 billion revenue base are not concept-stage numbers. The window will close eventually — the question is which companies are already through it when it does. That is why the bifurcation matters more than the direction: the first wave clears, the second wave waits.
What to Watch: The Offer Price and the First Week
The base case is that Anthropic lists in the fourth quarter of 2026 at a valuation above $1.5 trillion, and its debut determines the pace of everything behind it. A clean first week of trading — price holding at or above the offer — would pull Databricks, Stripe and the remaining AI-infrastructure names into a 2027 calendar that is already crowded. A debut that fades the way SpaceX's did would stretch the window and force later names to either cut valuations or wait.
The upside case is that Anthropic's revenue run rate and its overtaking of OpenAI convince investors that AI revenue is now visible enough to support trillion-dollar public valuations. In that scenario, the 2026-2027 IPO wave becomes the largest transfer of private-company value to public markets in history, and the AI complex is re-rated as a durable public asset class rather than a private-market phenomenon.
The downside case is that the window closes on itself. If the Federal Reserve turns more hawkish than the roughly 70% no-change consensus expects, if Anthropic prices below $1.5 trillion, or if its stock closes its first week below the offer price, the wave stalls. OpenAI's reported shift toward 2027 would then look less like selective patience and more like the first move in a broader delay.
Who benefits and who is exposed is asymmetric. The beneficiaries are the companies at the front of the line with the strongest revenue visibility — Anthropic first, then the AI-infrastructure names that can show contracted demand rather than modelled growth. The exposed are the second-tier listings that price at private-market marks into a window that has already absorbed SpaceX, and the public AI incumbents whose valuations get compared, quarter by quarter, against newly listed rivals with the same narrative and less public history.
The signal to watch is simple and quantifiable: Anthropic's final offer price and its first-week closing price. If the offer comes in below $1.5 trillion, or if the stock closes its first week below the offer, the "wave" thesis is wrong and the calendar slides into 2027. If it holds, the door stays open and the rest of the pipeline follows.
The real story is not how many AI companies list. It is whether the public market will price AI on revenue visibility or on capex promises — and Anthropic's debut is the first time that question gets answered at trillion-dollar scale.
Explore more exclusive insights at nextfin.ai.
