By Wen Wuzhao
First, Lay the Cards on the Table: How Much Money Do These Three Companies Really Want to Drain From the Market?
Any discussion about "whether the market can absorb it" is essentially a liquidity arithmetic problem. Let's look at the hard data for the three companies first.
SpaceX is issuing shares at a fixed price of $135 per share.
Note, it's a "fixed price," not the usual book-building price range. This is a "take it or leave it" posture in itself, a signal of extreme confidence in demand.
The company plans to issue 555.6 million shares, raising approximately $75 billion, with a valuation of $1.77 trillion, surpassing Tesla in one fell swoop to become the seventh-largest company in the U.S. The underwriting syndicate is led by Goldman Sachs, followed by Morgan Stanley, Bank of America, Citigroup, and JPMorgan Chase.
The result? The book-building phase attracted over $250 billion in subscription demand, 3.5 to 4 times the $75 billion fundraising target.
On its first trading day, trading volume exceeded 500 million shares, the closing price settled above $160, and its market cap exceeded $2.1 trillion on day one. As CNBC put it, investors found it "hard to find a word to criticize it" that day.
OpenAI followed closely. The company confidentially filed an S-1 with the SEC on June 8-9, 2026, targeting a valuation of approximately $1 trillion (some reports suggest between $850 billion and $1 trillion). Goldman Sachs and Morgan Stanley are handling the offering, with the listing window between September 2026 and Q4. However, OpenAI's financial profile is much sharper. According to analysis of its finances by European business magazines, the company loses about $1.22 for every dollar it earns.
Full-year 2026 cash burn is projected to be as high as about $27 billion. More interesting is the market's own assessment: on the prediction platform Polymarket, traders priced the probability of "OpenAI truly completing its IPO before the end of June" at only 4%, and about 40% before the end of the year.
This indicates the market sees this confidential filing as a real but distant catalyst, not an imminent event. OpenAI itself used language like "may be a while" to cool market expectations.
Anthropic, among these three, has the most "beautiful" financial story. It confidentially filed its S-1 on June 1, 2026, beating OpenAI, and could go public as early as this fall. Its latest valuation is $96.5 billion, likely crossing the trillion-dollar threshold at IPO, with plans to list on Nasdaq around October 2026.
Its growth curve is almost surreal: annualized revenue went from $87 million in January 2024, to $1 billion in December 2024, to $9 billion at the end of 2025, and then $14 billion in February 2026, $19 billion in March, $30 billion in April. The most critical difference lies in profitability: Anthropic told investors it expects to achieve its first quarterly operating profit of $559 million in Q2 2026, with revenue of $10.9 billion, up 130% from Q1's $4.8 billion.
This means that among the three, Anthropic might be the only one that can prove to the public before listing that it "spends less than it earns."
Summing up these three offerings: the three IPOs together may demand over $200 billion from the public market. The entire U.S. IPO market raised only $45 billion in all of 2025. In other words, these three companies want to drain more than four times the total amount of last year's U.S. IPOs from the market in just a few months.
This is the full weight of the question "can the market absorb it?"
Wall Street's Split Attitude
If you only read the headlines, you'd think Wall Street is cheering. But the real attitude is highly fragmented, and this fragmentation itself is data.
The bulls' logic is the ample liquidity theory.
There is about $8 trillion sitting in U.S. money market funds. SpaceX's $75 billion fundraising is only about 1% of that. That ratio indeed sounds non-threatening. The bulls have another deeper argument: institutional investors have for years only been able to access AI indirectly—buying Nvidia for chip exposure, buying Microsoft for indirect exposure to OpenAI, buying Alphabet for exposure to DeepMind and Anthropic. Now they can finally buy pure-play assets directly. This pent-up demand is real. SpaceX's 3.5-4x oversubscription is empirical proof of this logic.
The bears' logic is the liquidity drain theory.
Bank of America's chief investment strategist, Michael Hartnett, believes this epic IPO cycle is essentially "a massive transfer of risk accumulated by early investors to the public market."
This statement is very harsh but hits the essence of an IPO.
VCs and employees cash out, and retail investors and index funds are left holding the bag.
What the bears worry about is not insufficient money, but the chain reaction of money being siphoned away.
In fact, this concern has already materialized once: on June 5, 2026, the Nasdaq plunged 4.18% in a single day, the Philadelphia Semiconductor Index collapsed 10.3%, the chip sector lost over $1.2 trillion in market value in one day, Nvidia fell about 6%, and Micron dropped 13%. Analysis widely suggests that the impending IPO wave of the three giants is draining Wall Street's liquidity, forcing institutions to sell chip stocks to free up cash.
Then there's the "capitulation-style bullishness."
I believe this best reflects Wall Street's true state of mind. RBC's head of derivatives strategy, Amy Wu Silverman, mentioned a client who called himself a "fully invested bear," whose portfolio looks very bullish, but he stays in the market because "there's career risk in not being long AI stocks." This mentality of "knowing it might be a bubble but not daring to not participate" is precisely a typical feature of the late stages of a bubble.
As for Goldman Sachs internally, even they are arguing. Goldman analyst Covello recalled that he estimated in 2024 it would take 18 months to two years to see if the capital flooding into AI infrastructure could generate returns commensurate with the spending. He told his colleagues, we are well past that timeframe, but the returns have not yet materialized.
His colleague, Goldman Sachs Global Institute co-head George Lee, is more optimistic, but even the optimists acknowledge a sharp premise: the world may ultimately spend $7 to $8 trillion on AI infrastructure, but merely "disrupting" existing profit pools won't fill this hole—the math only works if AI creates a large amount of net new economic activity, and currently enterprise-level ROI data is disappointing. This question, "do the enterprises make or save money implementing AI," is the ultimate question mark hanging over all AI valuations.
Why Are These Three Rushing Out at This Moment?
From the context of events, this is not just about fundraising needs; it also involves some deeper characteristics of the U.S. tech industry.
Capital has a way of pricing distant stories. SpaceX isn't just selling rocket launches and Starlink coverage; it's selling the vision of Mars colonization and global internet. OpenAI and Anthropic aren't selling current revenue or profits; they're selling technological leadership when Artificial General Intelligence arrives. This ability to assign a present-day market value to the "future" is a signature skill of the U.S. capital market. Compared to the 1999 internet bubble era, the difference now is that these companies indeed have real revenue growth. Anthropic went from tens of millions in revenue to hundreds of billions in just over two years—something those internet companies 20+ years ago couldn't do. So where exactly is the boundary—is this time really different, or is history repeating itself—that's where the difficulty of judgment lies.
Founders have exceptionally strong personal control. Musk will hold over 82% of the voting power in SpaceX after the issuance. This means retail and institutional shareholders provide the capital but have little say in decision-making. Musk also amplifies this control through cross-holdings across multiple companies—Tesla holds SpaceX shares, SpaceX merges with xAI, and SpaceX then leases AI computing power to other companies. This network of related-party transactions requires investors to fundamentally trust this person's judgment.
Retail investors are being actively drawn into this round. The traditional retail allocation in IPOs is 5% to 10%; SpaceX raised that number to 30%. OpenAI has also specifically designed channels for retail participation. The surface reason is very American: millions use ChatGPT and Starlink daily, so let users become shareholders too. The deeper effect is that retail investors are taking on the tail risk of these high-valuation assets, whose valuations themselves haven't yet been truly tested by the public market.
The flow of funds has formed a closed loop. Nvidia might invest hundreds of billions in OpenAI, which promises trillions in compute spending, most of which goes to Nvidia chip purchases, and Oracle builds data centers with OpenAI as an anchor tenant. Such a structure benefits the revenue and valuation of all participating parties simultaneously, but also plants a seed of doubt: is this money creating new value, or is it just circulating repeatedly among existing values?
Can the Market Really Absorb It?
Returning to your core question. I believe the answer isn't a simple "yes" or "no," but needs to be viewed in three layers.
First layer, can a single offering be digested? Yes.
SpaceX has already proven this with its first-day $2.1 trillion market cap and oversubscription. The $8 trillion "ammunition depot" of money market funds exists objectively, and the pent-up demand for pure-play AI assets exists objectively. Referencing chip company Cerebras's IPO last month—initially priced at $115-$125, later raised to $150-$160, and finally priced at $185—shows Wall Street's excitement for pure-play AI assets is boiling. Looking at any single company, the market has the capacity to absorb it.
Second layer, can the market digest three consecutive offerings simultaneously without them tripping over each other?
This is the real risk point.
OpenAI and Anthropic filing in the same window could split institutional demand, leading to one or both being undersubscribed, or priced below private market valuations. That's why the three companies' timetables are actually staggered—SpaceX landed in June, Anthropic points to October, OpenAI between September and Q4. Staggering the peaks itself is the underwriters' sober recognition that the "market can absorb it, but not in one gulp."
Third layer, and the deepest one, can the fundamentals support the valuation? This is where the suspense lies.
Here, the fates of the three companies could completely diverge.
SpaceX has real Starlink subscription cash flow and launch business; if Anthropic can truly deliver its first profitable quarter, it proves to the public that an AI lab can also make money, which would be an anchor for the entire sector.
And OpenAI is the most fragile link. Its registration documents will publicly reveal its real revenue and profit margins for the first time. This is a direct confrontation with valuations that have far outpaced disclosed revenue. If the disclosed profit margins are significantly below investor expectations, the $850 billion valuation could face severe compression, triggering a chain reaction of writedowns in VC portfolios holding large AI positions.
It's worth mentioning a differentiating variable for Anthropic: it is a Public Benefit Corporation (PBC) and also has a "Long-Term Benefit Trust" (LTBT), placing safety oversight above short-term investor pressure. For investors pursuing pure returns, this is governance uncertainty; but for those who believe AI needs constraints, this is actually a selling point. This design of "writing the mission into the corporate charter" is itself a very special American-style narrative, using governance structure to hedge against the risk of technological loss of control.
Conclusion
To answer your headline in one sentence: The market can absorb it, but whether it can digest it depends on the moment OpenAI unveils its financial cards.
SpaceX has already proven the appetite exists. The real test comes in the fall, when the audited financial reports of OpenAI and Anthropic are first exposed to the sunlight, and the market will have its first chance to price "humanity's tomorrow" with real numbers rather than private market narratives. The second half of 2026 will be the first real test of whether frontier AI valuations can withstand the scrutiny of the public market.
Historical experience is contradictory. A Bloomberg survey shows strategists generally expect the S&P 500 to be around 7269 points at the end of 2026, about 6% higher than current levels. This gain looks quite meager compared to the 80% cumulative gain since 2023, even below the historical average annual gain of 8%.
Wall Street is putting money on the table while frantically buying insurance to hedge. This mentality of the "fully invested bear" is neither the pure frenzy of 1999 nor the full-blown panic before a crash. It's a sober gamble: knowing the mountain has tigers, but the professional risk of not going up the mountain is greater.
One judgment framework is worth remembering: Covello's question, "do the enterprises make or save money implementing AI." When an AI company's prospectus can answer this question with real customer ROI data, its trillion-dollar valuation will have a foundation; otherwise, what it's selling is still just a very American story about tomorrow.
