SpaceX listed on Nasdaq on June 12, 2026, at $135 per share. It opened at $150 and closed its first day at $160.95. By June 16 — four days after listing — it had reached $225.64. That peak lasted for approximately twenty-four hours.
SPCX is now trading at approximately $152. That is 32 percent below its all-time high, twelve days after that high was set. It is $17 above its IPO price. It is $8 below where it closed on its first day of trading. Ninety-six percent of shares remain locked until December.
On June 26, two separate reports confirmed that OpenAI and Anthropic — the two most anticipated AI company listings in the world, with private valuations of $852 billion and $900 billion respectively — have each cited SpaceX’s stock performance as the primary reason for pushing their planned 2026 listings into 2027. OpenAI’s CEO Sam Altman was described as “spooked by the SpaceX tumble.” Anthropic explicitly named “SpaceX’s stock performance and prevailing market conditions” in its delay rationale.
The phrase now circulating in IPO coverage is “the SpaceX scare.”
That phrase warrants examination. The SPCX correction is not a scare in the sense of an unexplained market event. It is a predictable, mechanistically grounded outcome that we documented in three separate analyses before it reached its current price. The question this week is not why it happened. The question is what it means that the two companies best positioned to understand listing mechanics have concluded that watching it happen was enough reason to step back from the queue.
The SPCX Numbers, In Order
June 12: IPO price $135 per share. Day one open: $150. Day one close: $160.95.
June 16: Peak price $225.64 — 67 percent above the IPO price in four trading sessions, reached before any short-selling mechanism existed. The float was 4 percent of total shares. There were no put options. There was no borrowable supply for short sellers. The only directional trade available in the market was long.
June 17: Options trading began. For the first time, the market had a mechanism to express a bearish view on SPCX. The stock has not recovered its June 16 high since.
June 28: Approximately $152. Thirty-two percent below the June 16 peak. Six percent below the June 12 day-one close. Seventeen dollars above the IPO price. The 52-week range: $135.00 to $225.64 — a range whose upper extreme was reached and abandoned in four days.
In our June 23 analysis of the float mechanics and lockup structure, we documented why this trajectory was the predictable outcome of the listing architecture: 4 percent float suppresses price discovery, generates a run on supply that cannot be borrowed, and creates a peak that holds until the first structural change (options, partial lockup expiry, earnings) introduces new information or new supply. The June 17 options start was that structural change. The correction followed within the same week.
The trajectory from $225 to $152 in twelve days is faster than we initially modelled. It suggests the market absorbed the mechanics quickly once a short mechanism existed — that institutional holders who understood the overvaluation had been waiting for the options market to open before moving.
What “The SpaceX Scare” Actually Is
The phrase “SpaceX scare” frames the SPCX correction as a warning that was sent to the market and received by the next companies in line. That framing is accurate but incomplete. It does not capture what specifically OpenAI and Anthropic observed.
They did not observe a bad company. SpaceX is one of the most technically accomplished companies in history — a launch provider with near-monopoly position in heavy commercial lift, a satellite constellation with real consumer and enterprise revenue, and an AI and aerospace conglomerate following its merger with xAI. The SPCX correction is not a signal that Space is a bad business. It is a signal about what happens when a good business is listed at the wrong price in a structure that cannot support price discovery until supply normalises.
What OpenAI and Anthropic observed was the following: a company with genuine fundamentals, listed at $1.77 trillion against analyst fair value estimates of $780 billion (Morningstar) and $1.3 trillion (Damodaran), on a float structure that suppressed short selling, reached an extreme of $225 on day four, and has spent twelve days returning to a price only $17 above where it launched. The correction happened before any earnings were reported, before any fundamental change in the business, and before any meaningful new information about SpaceX was released. It happened because the mechanics resolved: options allowed the market to price what the full float would eventually price, and the result was a rapid convergence toward the range that fundamental analysis had predicted.
The institutional community watching this process does not call it “the SpaceX scare” because SpaceX scared them about AI or about technology. They call it that because the mechanics of what happened are now visible to every company preparing to go public at a narrative-peak valuation on a thin float. SpaceX ran the playbook for them. The result is being marked to market in real time.
OpenAI’s Delay and What It Reveals
OpenAI’s path to an IPO has been unusual at every stage. The company filed a confidential S-1 with the SEC around May 22, with the filing confirmed publicly on June 8, 2026. At the time, September 2026 was the internal target. Goldman Sachs, Morgan Stanley, and JPMorgan were engaged as lead underwriters.
By June 26, reporting from multiple outlets indicated that OpenAI was “tilting toward” a 2027 delay. CEO Sam Altman was described as having been “spooked” by the SpaceX correction. CFO Sarah Friar had been internally advocating for 2027 on separate grounds — the company’s $3.7 billion annual burn rate, its compute infrastructure commitments, and the disclosure burden of quarterly public reporting were all cited as operational considerations that argued for more time as a private company.
The external and internal reasons for the delay are interesting separately. Altman’s concern about SPCX’s performance is a market-signal response: he is watching the correction and concluding that the September 2026 window is the wrong moment to bring a $1 trillion listing to a market that has already processed and corrected one $1.77 trillion AI listing. Friar’s concern is operational: the burn rate, the compute commitments, and the reporting burden are real constraints that do not disappear if the market improves.
The combination matters because it suggests two independent decision-making paths are converging on the same answer. Altman reads the market and sees the wrong moment. Friar reads the business and sees insufficient readiness. Together they produce a 2027 outcome from different inputs.
But the Altman response is the one that is new. The Friar concerns were known. The $3.7 billion burn rate has been public information. The compute commitments are disclosed in partner agreements. What changed in the last two weeks was the SPCX correction — visible, fast, and now being discussed in every IPO preparation meeting in Silicon Valley.
There is a secondary dimension to OpenAI’s delay that connects directly to what we documented in our June 24 analysis of OpenAI’s listing structure: Altman is “holding firm on $1 trillion.” He will not accept a lower valuation than $1T at listing. The market, after processing SpaceX’s correction, may not be willing to offer $1T on thin float mechanics in the immediate aftermath of SPCX’s return toward fair value. The delay is not only strategic patience — it may be the only path to the valuation Altman has committed to publicly.
If the September 2026 market would offer OpenAI $800 billion and Altman wants $1 trillion, 2027 is the answer by arithmetic. The SPCX correction compressed what the market was willing to pay for narrative-peak AI exposure. Waiting for that compression to fade — or for the SPCX lockup expiry in December to resolve whether SpaceX represents a floor or a precedent — is rational calendar management for any company that needs a specific valuation target.
Anthropic’s Delay and What It Reveals
Anthropic’s situation is structurally different from OpenAI’s. Anthropic filed its confidential S-1 with the SEC on June 1, 2026 — eleven days before SPCX began trading. At that point, Anthropic was targeting an October 2026 listing at a valuation of approximately $900 billion to $965 billion. Goldman Sachs, JPMorgan, and Morgan Stanley were engaged as lead underwriters — the same three banks handling OpenAI.
Anthropic’s S-1 filing was completed and submitted. The regulatory process was underway. The company had passed the point at which a delay is costless — filing fees, banker time, management bandwidth, and internal preparation resources had already been committed. Yet by June 26, less than four weeks after filing, the company had concluded that the October 2026 window was no longer optimal, citing SpaceX’s performance explicitly.
The speed of the pivot is notable. From S-1 filing (June 1) to SPCX listing (June 12) to SPCX peak (June 16) to SPCX -32% from peak (June 28) to Anthropic delay announcement (June 26): twenty-five days. A company that had spent months preparing for an October listing recalibrated its entire go-public timeline in less than a month of observing a single stock.
This is not a sign of indecision. It is a sign of sophisticated market reading. Anthropic’s founders and board understand the listing mechanics. They watched SpaceX run the playbook — thin float, narrative peak, no short mechanism, options as structural change, rapid correction. They concluded that following SpaceX into the market in October 2026, four months after the SPCX correction, does not offer the window they want. The retail FOMO that drove SPCX to $225 has been educated. The next company in line does not benefit from a buyer population that has already watched one AI listing correct 32 percent.
What Anthropic is waiting for is not necessarily better fundamentals or a different business. It is a market in which the SPCX correction has been absorbed, processed, and forgotten — or at minimum, in which the December lockup expiry has provided a new floor price for AI listings. Once SPCX trades with its full float and the market prices SpaceX against genuine supply, the reference point for AI listing valuations resets. That reset may be up or down from current levels, but it will be cleaner than the current state, where SPCX is an unresolved question mark hovering above the entire AI IPO category.
Anthropic’s governance structure adds another dimension to the delay’s logic. The governance of Anthropic — which will involve trust arrangements and mission-driven controls that “will be the single most-debated feature of any S-1,” according to institutional analysts — is easier to defend when it is not entering a market that has already punished a novel governance structure (OpenAI’s nonprofit Foundation control) for the first time. Waiting for the market to process one complex governance structure before introducing a second is sensible sequencing.
The Queue Reordering
The AI IPO queue that was described as a “three-whale” event — SpaceX done, OpenAI fourth quarter 2026, Anthropic October 2026 — has now reordered in ways that the original plan did not anticipate.
SpaceX completed its listing on June 12. That part of the plan executed. The $75 billion capital raise was the largest IPO in history. The mechanics produced the outcomes the plan predicted: four-day peak, options-start structural break, ongoing correction.
The second and third whales are now both targeting 2027. The $3.7 trillion combined market capitalisation that was going to drain global IPO liquidity “from June through the end of the year” has been reduced to approximately $1.77 trillion for 2026 — SpaceX alone — with the remainder deferred. Polymarket traders are pricing roughly 30 to 40 percent odds of no OpenAI IPO by end-2026, reflecting the uncertainty that reporting has introduced into a timeline that was previously described as nearly certain.
Databricks remains on its original schedule, with an S-1 filing expected in Q3 2026 and a listing likely in Q4 2026 or Q1 2027. The company has not announced a delay. At $134 billion to $175 billion in private valuation — significantly below the trillion-dollar range of OpenAI and Anthropic — Databricks may calculate that the SPCX correction is less relevant to its own listing. A $134 billion listing faces less narrative-peak pricing pressure than a $1 trillion listing. The multiple compression risk (32x revenue versus Snowflake’s 7x) remains, but the FOMO quantum is smaller and the float mechanics may be more conservative.
The queue reordering has an unintended consequence for the N4 thesis: it extends the timeline of observable data points. If OpenAI and Anthropic had listed in 2026 as planned, we would have three live datasets by December — SPCX in month six, OpenAI in month two or three, and Anthropic approaching its lockup. Each would be at a different point in the correction cycle, providing a rich set of observations across the listing mechanics.
With both pushed to 2027, the 2026 data set is SPCX alone. The December lockup expiry becomes the definitive 2026 test — the only observable moment at which a company that ran the playbook faces the full supply. Whatever SPCX prices at in December, when 96 percent of shares become tradeable, will be the reference point that OpenAI and Anthropic are watching when they make their final go-public decisions in early 2027.
Why the Correction Happened This Fast
The speed of SPCX’s correction — from $225.64 to approximately $152 in twelve days — was faster than we modelled when we established the N4 framework in June. In the original analysis, we expected the correction to unfold across a longer arc, with meaningful pressure events at three points: the June 17 options start, the August partial lockup expiry, and the December full lockup expiry.
The June 17 options event produced a correction far sharper than a typical options-start effect. The explanation is likely that the June 16 peak concentrated an unusual amount of conviction in both directions — buyers who had accumulated above $200 were deeply in profit and eager to lock it in, while the institutional short interest that had been structurally impossible before options was immediately available and material. The options market did not gradually introduce bearish positioning. It opened and institutional capital moved into it the same day.
This is the acceleration effect of modern market structure applied to a thin-float listing. When the only barrier to a trade is the absence of a mechanism (no options, no borrowable shares), the removal of that barrier produces instant deployment of capital that was waiting. Institutional investors had weeks to construct their thesis, model their target price, and prepare their trade. They were executing plans they had built during the float-suppressed run-up, waiting for the structural change. The June 17 options open was the starting gun, not a slow ramp.
The implication for future listings is that the correction window may be shorter than the listing-to-lockup timeline suggests. The mechanism does not require waiting for December. It requires waiting for the first structural change. For SPCX, that was five days. For future listings, the market has now been educated: whenever a thin-float AI listing runs on narrative, the first derivative product creates an immediate exit for institutional holders who bought the theory but not the valuation.
What the December Test Will Now Mean for the Queue
The SPCX December lockup expiry — when 96 percent of shares become tradeable for the first time — will now carry two audiences: conventional investors tracking SpaceX’s fundamental trajectory, and the OpenAI and Anthropic boards who are watching for a reference price.
If SPCX prices above its IPO price of $135 in December, it suggests the market can sustain some premium above fundamental value even with full supply — that the quality of the underlying business provides a floor above the analyst range. This would give OpenAI and Anthropic reason to believe their own listings at significant premiums to any DCF estimate are supportable.
If SPCX prices below $135 in December — below the IPO price at which the original buyers acquired shares — the thesis is confirmed in its fullest form: the listing event was priced above what the full market will support when given access to supply. Retail buyers who purchased above $135 in June will have lost money on a company with excellent fundamentals, solely because they bought at a listing event that structurally concentrated them at the narrative peak. OpenAI and Anthropic boards watching this outcome in December will face a question about whether any listing at their valuations — $1 trillion, $965 billion — can clear the full-supply test at all.
The August partial lockup expiry is the preview. Gary Black of The Future Fund specifically flagged August as the moment to revisit SPCX following the partial insider release. The August event is not the full test, but it introduces additional supply before December and will indicate whether the holding cohort above $200 maintains or distributes.
What the Thesis Has Predicted and What Has Come True
The framework we established across this series rests on a single observation: the specific market psychology and structural mechanics that crypto markets normalised — listing event as narrative peak, thin float as price discovery suppressor, lockup expiry as the real exit event — have migrated into traditional equity markets via a wave of high-anticipation private company listings.
In the first article in this series, published June 21, we established the mechanism: SpaceX listed at $1.77 trillion against analyst estimates of $780 billion to $1.3 trillion, with the same FOMO concentration pattern that academic research (Xu and Livshits, 2019) documented in crypto pump-and-dump events — retail buyers absorbing supply at maximum narrative concentration.
In the second article, published June 23, we quantified the float mechanics: 4 percent circulating supply, 96 percent locked, options as first short mechanism, and the Cursor acquisition as an example of how the inflated listing price concretely benefited SpaceX as an acquirer.
In the third article, published June 24, we documented SPCX’s early confirmation (-27% from peak at that point) and the queue behind it: OpenAI at $1 trillion with CEO equity listed as “TBD,” Databricks at 32 times revenue against Snowflake’s 7 times.
What has come true in the four days since that third article: SPCX has fallen from approximately $164 to approximately $152. OpenAI and Anthropic have both cited SPCX to explain delays. The correction is not slowing. The queue has reorganised around the correction. And the “SpaceX scare” is now the phrase being used in financial media to describe what we called the FOMO contagion thesis.
The thesis has not predicted events. It has described a mechanism. The events are following from the mechanism.
The Governance Question for the Queue
One element of the delay that has not received adequate attention in the initial reporting is what the delay period allows OpenAI and Anthropic to resolve.
OpenAI’s CEO equity situation — where Altman’s ownership stake was listed as “TBD” in the confidential S-1 — is one of the most structurally unusual disclosures in the history of large technology IPOs. The delay to 2027 provides a window to resolve that disclosure before the public filing. Whether Altman receives a negotiated equity grant, a deferred compensation arrangement, or a disclosed zero is information that institutional investors will require before the S-1 becomes public. The delay is partly time to finish a negotiation that was not finished when the confidential filing was submitted.
Anthropic’s governance trust arrangement — which its advisors describe as “the single most-debated feature of any S-1” — is similarly complex. The delay provides time to test governance structures in private, to negotiate with institutional anchors about what governance protections they require, and to see how the market reacts to OpenAI’s governance disclosure once it becomes public. Anthropic listing after OpenAI files its public S-1 means Anthropic can calibrate its governance disclosures against the market’s actual response to OpenAI’s, rather than entering blind.
Both companies are, in different ways, using the delay to resolve the specific governance ambiguities that we identified as the new dimension of the FOMO contagion problem — the feature that distinguishes the OpenAI and Anthropic listings from SpaceX, where Elon Musk’s ownership stake was known, large, and unambiguously aligned with the stock price.
What the Thesis Now Requires for Completion
The N4 framework has three remaining observable events before the picture is complete for 2026 and early 2027.
The first is SPCX in August, when Gary Black’s flagged partial lockup expiry introduces new supply. The August event is not definitive, but it shows whether early holders above $200 are waiting for December or distributing earlier when partial mechanics allow.
The second is SPCX on September 2, when the company reports its first post-IPO earnings. That earnings call will be the first moment SpaceX must answer, in public reporting, the questions Damodaran’s DCF raised: how does the revenue trajectory compare to the implied assumptions at $135 per share? What does the Cursor integration add to the growth model? What are the unit economics of Starlink at scale? Earnings do not determine the stock price on their own, but they set the fundamental reference that the December lockup expiry will price against.
The third — and most important — is December, when the full 96 percent unlocks. That is the test. Whatever the market is willing to pay for SpaceX when every holder can sell, and every buyer has access to the full float, is what SpaceX is worth in the post-FOMO equilibrium. That number will be cited by every OpenAI and Anthropic board member when they finalise their 2027 listing plans.
The FOMO contagion thesis has moved from an analytical framework to an active market signal. The correction we documented is now the reference point that two $1 trillion companies are using to calibrate when they go public. The mechanism we described has become the shared vocabulary of the AI IPO category — “the SpaceX scare” is our thesis in three words.
December will determine whether the scare was warranted or whether it becomes the moment the market decided it was safer to wait and missed the window. Either way, the data will be observable, and the thesis will be tested against it.
