The Exit Window: $3 Trillion Heads for the Door

The biggest private companies ever built are filing to go public inside the same 90-day window.

What $1.77 Trillion Actually Means

Read theSpaceXnumber the way an underwriter would, not the way a headline does. At a $1.77 trillion offering valuation, the company has to generate roughly $1.1 trillion of revenue by 2035, a 50% compounded growth rate held for a full decade. No company in history has done that. That figure is not a risk factor buried on page 90 of the prospectus. It is the entire pitch.

Here is what should hold an allocator's attention: the three most valuable private companies ever assembled,SpaceX,OpenAI, andAnthropic, have all begun moving toward public markets in the same quarter. When the most informed, least liquid holders on the cap table simultaneously decide it is time to access public liquidity, that is a top-of-cycle signal for the entire private-market complex. Treat it as one.

The Marginal Seller

A synchronized rush by the largest private companies to list is not a sign of market health. It is a sign that insiders, founders, early VCs, late-stage funds, have concluded that conditions will not get better than this. They are not raising capital. They are pricing their exit. The question for your book is not whether these are good companies. It is whether you want to be the marginal buyer at the moment insiders become the marginal seller.

The Setup: $3 Trillion at the Door

• SpaceX:$75B raise, $1.77T valuation, $135/share, Nasdaq debut targeted for mid-June.

• OpenAI:confidential S-1 filed May 22, targeting an $852B to $1T valuation, September listing.

• Anthropic:estimated ~$900B float expected October.

That is roughly $3 trillion of new equity value asking to be absorbed in 90 days. None of the three is profitable on a net basis. And the market is being asked to underwrite it while the 30-year Treasury sits at 5.2%, the highest since 2007, meaning the risk-free hurdle these stories must clear is the steepest in a generation.

Underwriting a Company That Doesn't Make Money

Unprofitable does not mean uninvestable. Amazon lost money for years. The discipline is separating cash burn that builds a compounding asset from cash burn that funds a structurally unprofitable one. Three questions do most of the work:

1. Unit economics: is the per-unit picture improving or fixed?

The success stories showed contribution margin inflecting upward as they scaled: each incremental customer cost less to serve and the moat widened. The nightmares scaled revenue while the per-unit math stayed broken, growth made the losses bigger, not smaller.

2. Path to profitability: does scale fix it, or is the loss structural?

"We'll grow into it" is a thesis only when the cost curve actually bends. Ask whether profitability is a function of time and volume, or of a problem no amount of volume solves.

3. Capital intensity: how many more dollars of capital per dollar of revenue?

This is where the current cohort separates. OpenAI generated $13.1B of revenue in 2025 and burned ~$22B to do it: a $9B net loss, a projected $14B operating loss in 2026, and an estimated $207B of additional capital needed through 2030 just to honor existing compute commitments. That is not a company growing into profitability; it is a company whose growth is gated by an extraordinary, recurring capital requirement.

Apply the same lens to SpaceX and the picture splits in two. Starlink is a genuine business: $11.4B of revenue and $7.2B of EBITDA in 2025, subscribers doubling to 10.3 million in a year. That throws off real cash. But at $1.77T you are not paying for Starlink's cash flows. You are paying for the story that SpaceX captures 2.4% of U.S. GDP by 2035, a share of national output no single company has ever held. Fortune's math: SpaceX would need to add $360B of revenue in a single year, a figure Amazon reached only cumulatively across six. Morningstar values the company at $780B, roughly half the ask, and flags xAI as a "material threat of value destruction" with an "economic moat indeterminate."

The lesson from every prior vintage holds: the winners were identifiable by improving unit economics and a credible cost curve, not by the size of the headline. So were the losers.

The Pattern: Every Exit Window Rhymes

Wall Street has a reliable tell for late-cycle excess: a sudden acceleration in IPO supply.

• 1999:350+ tech IPOs. Nasdaq-100 forward P/E hit 60x in March 2000. Average first-day pop: 69%. The index then fell 77% by October 2002.

• 2021:1,000+ listings in a single year, ~80% unprofitable. Rivian raised $12B on Nov 10; the Nasdaq printed its all-time high of 16,057 nine days later, then fell for 14 months.

• 2025:Klarna priced at $40 and is down 62%. StubHub priced at $23.50, fell 6.4% on day one and 21% within three sessions.

The current bull run is 17 years old. The Magnificent Seven are 33% of the S&P 500. Multiples are pressing toward dot-com levels. And $3 trillion more is lining up at the door.

Cracks in the System

Standard doom-mongering points at valuation and stops. The more useful work is naming the structural fault lines that make this cycle's plumbing fragile, the things you can actually raise with a client.

1. The valuation is decoupling from any public anchor.

When an independent fair-value estimate (Morningstar's $780B) sits at roughly half the offering price, the gap is not a rounding error. It is the spread between intrinsic value and sentiment. The deal is priced to the maximum the market will bear, not to a margin of safety. Underwriters price the story; they do not price the downside.

2. Liquidity is being manufactured, not discovered.

Nasdaq's fast-entry rules can force passive index funds to buy these names at IPO prices before genuine price discovery occurs. That converts an insider exit into mandatory demand from millions of index holders: liquidity engineered by structure, not earned by fundamentals. If your clients own broad index funds, they may be buying these IPOs whether they chose to or not.

3. You are buying the float, not the franchise.

Elon Musk retains 82% of the vote after the SpaceX offering. Public holders supply the capital and absorb the price risk while control stays entirely inside. A deal the press frames as minting "a new class of billionaires" mints them with outside capital. Read the governance terms before the growth slide.

4. The risk-free hurdle is the highest in 17 years.

Every one of these stories is a long-duration bet on cash flows a decade out. Discounting those flows at a 5.2% long bond is a very different exercise than at the near-zero rates that underwrote the 2021 cohort. The denominator moved against the story.

The Framework: How the Exit Window Works

Stage 1: Insiders decide it's time.

Holders who know the real burn, the real competitive dynamics, and the honest valuation math file an S-1. They are accessing liquidity at a moment they judge to be peak.

Stage 2: Banks price the story.

IPOs are priced to maximum sentiment, not intrinsic value. At the top of the range, the buyer pays for peak confidence with no margin of safety.

Stage 3: The public holds it.

By the time shares reach a brokerage account, the builders have been paid. The window opens when insiders decide conditions won't improve. It closes when the market figures that out, typically 12 to 18 months later. That was true in 2001. It was true in 2022.

What This Means for Allocators

None of these names is permanently uninvestable. The pattern is simply consistent: frenzy peaks, listings disappoint, valuations reset, and the durable entry point appears on the other side of the window, usually 12 to 18 months on. Sophisticated allocators do not chase the Exit Window. They wait for it to close.

The question was never whether to own SpaceX eventually. It is whether $1.77 trillion is the price at which it makes sense.

The Allocator's Checklist: Late-Stage Private Tech, 2026 Vintage

Before adding late-stage private or newly public tech exposure this cycle, require a defensible answer to each:

• Insider behavior:Who is selling, how much, and why now? If the most informed holders are net sellers, ask what they see that you don't.

• Unit economics:Is contribution margin inflecting up, or does growth widen the loss?

• Path to profitability:Is profit a function of scale and time, or of a problem volume can't solve?

• Capital intensity:How many more dollars must be raised before free cash flow turns positive? Who funds the gap, and at what dilution?

• Valuation vs. anchor:How far above an independent fair value and public-peer multiples is the offering price?

• Governance:What do you actually control? Dual-class voting means you own economics, not influence.

• Structural demand:Will index mechanics force you (or your clients) to buy at the IPO price regardless of view?

• Rate context:Does the thesis survive being discounted at a 5.2% long bond?

• Entry discipline:Is this the price, or just the moment? If you can't separate the two, wait for the window to close.

Decision Snapshot

One question to sit with before the bell rings:

Who profits if SpaceX opens up 20% on day one, and who absorbs the loss if it's down 20% six months later? If you can't answer both with specificity, you are standing inside the Exit Window.

SOURCES & METHODOLOGY

• New Constructs, June 2026: SpaceX revenue model; $1.1T-by-2035 requirement; 50% CAGR analysis

• CNBC, June 3, 2026: SpaceX $135/share, $1.77T valuation; Morningstar fair value $780B.

• Fortune, June 6, 2026: SpaceX valuation; $360B single-year revenue requirement; 2.4% of GDP.

• TechTimes, June 7, 2026: OpenAI S-1, $852B to $1T valuation, September timeline, $207B capital need through 2030.

• The National News, June 5, 2026: market-peak warnings; 30-yr Treasury at 5.2%; Mag 7 at 33% of S&P 500; Nasdaq fast-entry mechanics. Rathbones (Sanjiv Tumkur) and Charles Stanley (Garry White) quotes via The National News.

• Yahoo Finance, June 2026: SpaceX IPO and Starlink financials (revenue, EBITDA, subscribers); Musk 82% voting control.

• CNBC (Sept 2025) / IndexBox (2026): Klarna IPO pricing and 62% decline; StubHub day-one and three-session performance.

• Renaissance Capital / Quartr: dot-com IPO counts, Nasdaq-100 P/E, 2021 listing and Rivian data.

The Allocation Memo is published by EM Capital. For RIAs and Family Offices. Built for the allocator's desk. This is market commentary, not investment advice.