The 2026 Mega-IPO Pipeline: SpaceX, OpenAI, Anthropic, Kraken, and More

Introduction

The 2026 mega-IPO wave is the deepest in modern ECM history, and its centerpiece has already traded. SpaceX priced its Nasdaq IPO on June 11 at $135 per share, raising $75 billion in the base deal and $86.2 billion once the underwriters exercised their over-allotment option, nearly three times Saudi Aramco's $29.4 billion 2019 record. In July, SK hynix sold $26.5 billion of ADRs on Nasdaq, and AI chipmaker Cerebras had already raised $5.55 billion in May. The rest of the pipeline proved less certain: OpenAI says it will not list in 2026, Databricks and Kraken pushed their plans back, and Anthropic is reported to be targeting November. The general-reader story is in how SpaceX's record listing came together and why the 2026 IPO boom happened, and aggregate volumes are in the 2026 ECM market picture; this article covers the deals and their mechanics.

The 2026 wave did not appear from nowhere. It is the release of pressure that built through a multi-year freeze. After the 2021 IPO boom, the market essentially shut: the 2022-2023 rate shock drove the IPO window into its deepest drought in over a decade, and only 17 venture-backed unicorns went public in 2025. The result was an unprecedented logjam of more than 800 private unicorns sitting on a combined $4.3 trillion of locked value, with employees, early investors, and limited partners all waiting for liquidity that private secondary markets could only partly provide.

Several forces converged to crack the window open. Rate stability removed the single biggest source of valuation uncertainty: after three cuts in late 2025, the Federal Reserve held its target range at 3.50-3.75% at every meeting from January through July 2026, so issuers and public investors could price growth assets with far more confidence than during the 2022-2023 tightening cycle. The AI capital-expenditure supercycle created companies with genuinely insatiable funding needs: frontier models and data-center compute require tens of billions that even the deepest private rounds struggle to supply, pushing the largest AI names toward public markets as a matter of necessity.

The top of the backlog had also used the freeze to build public-company reporting and governance, and several names reached profitability. The first quarter of 2026 marked the strongest opening for US IPOs in five years, and when SpaceX, Anthropic, and OpenAI all filed between April and June, the reopening looked like a flood.

Why the Fall Window Stalled

The variable that opened the window then turned against it. With inflation still elevated, the Fed raised its target range to 3.75-4.00% on September 16, its first increase since 2023. Rising yields hurt IPOs twice: a higher discount rate shrinks the value of long-duration growth equity, and a higher risk-free return makes allocators less willing to pay for distant earnings. With fast-rising debt yields and the US midterm elections on November 3 weighing on sentiment, few issuers moved to price, and Dealogic's nine-month ECM review expects the window to reopen in late November or early December. Rates were not the only brake: OpenAI, Databricks, and Kraken each stepped back for reasons of their own, covered below.

SpaceX's IPO is the centerpiece of the 2026 pipeline, both for its size and for what its first months of trading taught.

Mega-IPO

An IPO transaction at exceptional scale (typically $5 billion+ of proceeds and/or $50 billion+ market cap at pricing). Mega-IPOs require dedicated institutional capital absorption, multi-bank syndicates with substantial distribution depth, cornerstone investor anchoring, and careful calendar staggering. Recent benchmarks, each including the over-allotment: SpaceX 2026 ($86.2B, the new record), Saudi Aramco ($29.4B in 2019), Alibaba ($25B in 2014).

Filing, Valuation, and Pricing

SpaceX filed a confidential S-1 with the SEC on April 1, 2026, codenamed "Project Apex," targeting a $1.75 trillion valuation (up from $1.25 trillion when SpaceX merged with xAI in February 2026). On June 11, 2026, it priced the deal at $135 per share (555,555,555 shares), raising $75 billion in the base offering. The underwriters also held a greenshoe, an option to buy up to 15% more shares from the company at the offer price, which covers the extra stock they sell short at allocation. With SPCX trading well above issue, they exercised it in full, and the offering closed on June 15 at 638,888,888 shares. Dealogic records the full deal at $86.2 billion, nearly 3x the $29.4 billion record set by Saudi Aramco's 2019 IPO (SpaceX's own $85.7 billion is net of the underwriting discount). The pricing valued SpaceX at roughly $1.77 trillion, making it the seventh-largest US public company at debut, ahead of Tesla (around $1.6 trillion).

Roadshow, Bookrunners, and Retail Allocation

SpaceX ran an abbreviated roadshow the week of June 8, 2026, including an event for roughly 1,500 retail investors on June 11. Shares began trading on Nasdaq under ticker SPCX on June 12, opening at $150 and closing the first day at $160.95, up 19%. Goldman Sachs and Morgan Stanley took the largest underwriting commitments, with BofA, Citi, and JPMorgan as fellow representatives and 17 other banks in smaller roles, 22 underwriters in all. SpaceX CFO Bret Johnsen indicated retail would be a bigger part than in any IPO in history.

The Aftermarket: A Record Pop, a Break Below Issue, a Recovery

The aftermarket became the deal's real lesson. With a thin free float (only the newly sold shares could trade) and much retail demand unfilled, SPCX ran to an intraday high of $225.64 on June 16, 67% above the offer price. Then it unwound. The stock fell through $135 in mid-July, and SpaceX's first earnings report on August 4 paired a revenue beat with soaring AI spending. The shares dropped 13.6% the next day to an all-time closing low of $108.27, 20% below issue and more than half off the peak. Within a week they were back above the IPO price, and through September and early October they closed between about $140 and $172.

Valuing a Company That Has Never Traded

Pricing SpaceX at roughly $1.77 trillion required valuing very different businesses bundled into one listing, with no clean public comparable for the whole. The natural approach is sum-of-the-parts: launch and Starshield on contract-backed cash flow, Starlink on subscriber and ARPU-driven growth multiples (the segment most investors treat as the real engine of the valuation), and the merged xAI assets on the revenue-multiple and strategic-scarcity framing applied to OpenAI and Anthropic.

SpaceX listed as one entity rather than spinning off Starlink, which forces investors to underwrite the whole conglomerate and accept management's capital allocation across all of it. Bulls argue the combination captures real synergies (Starlink funds launch, launch enables Starlink, xAI consumes the compute) and that a single mega-cap is easier to index and own. Skeptics counter that a conglomerate structure can mask weaker segments and invite a discount; the August sell-off over AI spending showed that risk in practice. For an ECM analyst, SpaceX is a live case study in how listing combined or separated shapes both the valuation narrative and the investor base.

A $75 billion base offering is not simply a larger version of a normal IPO; the mechanics change at scale. The defining challenge is absorption: no single investor, and no small syndicate, can place tens of billions of new equity into the market in a single morning. This is why the SpaceX deal carried 22 underwriters rather than the three or four a typical large-cap IPO uses. The syndicate is a distribution machine, with each bank responsible for placing stock with its own institutional relationships while the lead active bookrunners coordinate the book and the remaining banks broaden reach.

Several execution levers become essential at this scale:

The unifying theme is that mega-IPO execution is a problem of distribution capacity and risk staging, not valuation alone: placing this much stock without breaking the aftermarket is genuinely hard, as the contrast with a conventional offering shows.

For candidates, the lesson is to resist treating a mega-IPO as just a bigger headline number. The size is precisely what forces a different execution model, and an interviewer who asks why SpaceX needed 22 underwriters is testing whether you understand distribution, not arithmetic.

The most unusual structural feature of the SpaceX IPO was its embrace of retail investors. Where a typical large-cap IPO reserves roughly 10% of shares for retail, SpaceX first signaled an allocation as high as 30%, then trimmed it as institutional demand built. Individual investors ended up with about 20% of the base deal (roughly $15 billion), hedge funds 10%, and buy-and-hold institutions 70%, CNBC reported, citing Reuters: double the Wall Street norm and the largest retail push in IPO history. The company made shares available to individual investors through major brokerages including Charles Schwab, Fidelity, Robinhood, SoFi, and E*TRADE, with eligibility ranging from no account minimum at several platforms to a $100,000 liquid net worth requirement at others.

The rationale is part brand, part strategy. SpaceX is a rare consumer-recognizable name, and channeling that enthusiasm into the shareholder base broadens distribution and builds a base of long-term holders less likely to sell early. A large retail tranche also absorbs supply the institutional book might strain to take at the target price, even on a deal reported to be roughly 3.3x oversubscribed.

The two leading frontier labs filed confidential S-1s within a single week in June, which looked like the start of a race. Instead, the AI-lab group split.

OpenAI filed on June 8, 2026, with its last private round valuing it near $852 billion. Early reporting pointed to a debut as soon as September, but in late June The New York Times reported that advisers had laid out a choice between listing sooner at a lower valuation and waiting until 2027 to pursue $1 trillion, and that Altman had ruled out cutting the target. On September 11, chief executive Sam Altman told Fortune the IPO would come "not 2026," tying the delay to AI safety concerns rather than market conditions.

Anthropic filed on June 1, 2026, days after a $65 billion Series H at a $965 billion post-money valuation. Its timing has moved more than once: The Wall Street Journal reported on September 18 that the target had shifted from October to November, and Bloomberg later reported that formal marketing could begin as soon as the week of November 9. The company has not confirmed a date or made its prospectus public, though it shared third-quarter revenue of $16.8 billion with prospective investors, CNBC reported. Reported targets of about $2 trillion in valuation and around $100 billion in proceeds stay unconfirmed until the deal prices. With OpenAI out of the 2026 calendar, Anthropic carries the sentiment read for the whole AI cohort, and Dealogic frames fourth-quarter ECM as hinging on its listing.

Beyond the AI labs, the 2026 pipeline split into deals that priced, deals that were postponed, and names still waiting.

SK hynix: The Year's Other Mega-Deal

The year's second-largest listing came from outside the AI-lab group. SK hynix, the Korean memory maker whose high-bandwidth memory feeds Nvidia's AI accelerators, priced 177.9 million American depositary receipts at $149 on Nasdaq in July, raising about $26.5 billion, as The Korea Times reported, to fund new fab and advanced packaging capacity in Korea. The ADRs closed their first day about 13% higher. Because SK hynix already trades in Seoul, Dealogic files the deal as a US secondary listing rather than an IPO; the Level III ADR mechanics are covered in cross-border listings.

Cerebras, Databricks, Kraken, Stripe

Cerebras Systems, positioned as an Nvidia alternative, was the year's first large AI listing: it priced 30 million shares at $185 in May, above a range it had already raised, for $5.55 billion of proceeds, then closed its first day up 68%. Like SPCX, it later gave much of that back, hitting a post-IPO low in early October as its lock-up expired. Databricks went the other way: despite $5.4B of annualized revenue, positive FCF, and a reported $175 billion valuation target, chief executive Ali Ghodsi said in June that 2026 was a terrible year to go public, with the calendar crowded by SpaceX and the AI labs. Crypto exchange Kraken filed confidentially in November 2025 after raising at a $20 billion valuation, then paused in March 2026 amid weak crypto markets; a Deutsche Börse stake purchase in April implied about $13.3 billion, and CoinDesk reported in September that its parent, Payward, now sees the second quarter of 2027 as its earliest window. Stripe, valued at $159 billion in a February 2026 employee tender, is in no rush given profitability and secondary-market liquidity. Revolut ($75B), Canva (around $42B in 2025), and Shein ($50B) round out the international names, none with a 2026 listing date.

The risk most often named for the 2026 wave was capacity: allocators forced to choose among several giants might spread commitments thin or sit out later deals if the first few disappointed. The self-reinforcing calendar worked as feared, only through the aftermarket rather than the books: SPCX's slide below issue, Cerebras's fall to a post-IPO low, and the Fed's September hike narrowed the window behind SpaceX almost as quickly as SpaceX had opened it.

Down-Round IPO

An initial public offering priced at or below the company's most recent private funding-round valuation. Once rare for marquee names, down-round (or flat) IPOs grew more common after 2022 as public investors applied stricter discipline than the late-stage private and crossover funds that had bid valuations up. A down-round IPO is not necessarily a failed deal; it can still raise enormous capital, but it signals that public markets are repricing the asset below its last private mark.

The subtler risk is valuation, and OpenAI showed the usual response to it. Rather than accept less than the reported $1 trillion target (against a last private mark near $852 billion), it is waiting, even if Altman's public reason was safety. When the public market looks likely to clear an asset below its owners' price, issuers tend to choose delay over a down-round IPO. Public investors apply scrutiny that crossover and venture funds, racing to own a piece of the AI frontier, did not. For companies that raised at peak private valuations, the IPO can become a moment of repricing rather than celebration, and the gap between the private mark and the clearing price is exactly where money is left on the table in either direction.

The 2026 mega-IPO wave above is the most structurally important component of the 2026 ECM outlook. The next article walks through the PE-backed sponsor IPO backlog, where the multi-year build-up of aging sponsor-backed assets is driving sustained 2026 IPO and follow-on activity beyond the mega-IPO headlines.