深潮 TechFlow

wrote a column ·

An AI read SpaceX's S-1 filing and drafted this investment memo in 12 minutes

Author: Nick Prince

Compiled by: DeepTide TechFlow

Deep Tide Summary:An AI agent autonomously completed a task that would typically take an investment analyst team several days: it read through SpaceX’s 226MB S-1 filing, purchased real-time market data using USDC on the Base chain, and generated an investment committee memo featuring multi-perspective analysis, valuation models, and a risk matrix—all for a total cost of just $1.87. This isn’t a demo; it’s a record of actual paid API calls. Wall Street’s working model is being reshaped as AI agents gain the ability to pay for data and autonomously decide their analytical pathways.

An AI agent read SpaceX’s 226MB S-1 filing submitted on Monday, purchased real-time market data using USDC on the Base chain, and generated this investment committee memo within 12 minutes. Total cost: six paid API calls, $1.87 in USDC, with no API keys required.

SpaceX owns three businesses that competitors cannot replicate. First, near-monopoly access to commercial spaceflight—accounting for 80% of global orbital mass since 2023, with Falcon missions boasting a 99% success rate and reusable technology a decade ahead of rivals. Second, the world’s only deployed low-Earth-orbit broadband network—Starlink serves 10.3 million subscribers across 164 countries, up 49.8% year-over-year, with segment-adjusted EBITDA of $7.2 billion. Third, following its acquisition of xAI in February 2026, it became the only AI lab vertically integrated down to the launch vehicle level, with plans to deploy orbital computing capacity. By any reasonable valuation methodology, this constitutes a generational asset.

The connectivity business is real and profitable. But everything else is either burning cash at an alarming rate—the AI division reported a $6.4 billion loss against $3.2 billion in revenue in 2025—or hinges on Starship, which has completed 11 test flights but has yet to deliver payloads to orbit. This IPO is partly a refinancing event. SpaceX took out a $20 billion bridge loan to acquire xAI, maturing in September 2027, and the bridge lenders are precisely the underwriters of this IPO. If the valuation exceeds $500 billion, you’re paying for unproven execution capability, corporate governance you have no say in, and a refinancing deal the underwriters must make succeed.

Starlink is an excellent standalone business. In 2025, it generated $11.4 billion in revenue (+49.8%), $4.4 billion in operating income (+120%), and segment-adjusted EBITDA of $7.2 billion (+86%), supported by 10.3 million paying subscribers on premium-priced plans.

Launch services are unparalleled. Since 2023, they have accounted for over 80% of global mass-to-orbit, with Falcon rockets achieving a success rate exceeding 99%, and the Falcon 9 first stage having flown up to 34 times.

Vertical integration is real and compounding. Rocket → satellite → spectrum (EchoStar’s AWS-4/H-band deal has received FCC approval) → AI computing power (two COLOSSUS clusters totaling approximately 1 GW).

Government reliance is a moat, not a risk. Primary launch provider for U.S. national security: executing 11 of 12 National Security Space Launch missions in 2025 and all five of NASA’s crewed and cargo flights.

Orbital AI compute carries significant option value, with deployment planned for 2028. If Starship achieves even 50% of its targeted economics—reducing launch costs by 99%—the addressable market could expand by an order of magnitude.

The AI segment is a bottomless pit burning over $6 billion annually. In 2025: $3.2 billion in revenue versus a $6.4 billion operating loss, segment adjusted EBITDA of negative $1.2 billion, and capital expenditures of $12.7 billion. In Q1 2026 alone: $818 million in revenue against a $2.5 billion operating loss and $7.7 billion in capital expenditures. Annualized AI capex now exceeds $30 billion, while AI revenue stands at just $3.2 billion.

The actual debt burden is approximately $42 billion, not the headline figure of $29 billion. It comprises roughly $20 billion in SpaceX bridge loans (maturing September 2027), about $6.7 billion in X Corp.'s Term Loan B-1 and approximately $6 billion in X Corp.'s Term Loan B-3 (both maturing October 2029, with effective interest rates of 10–12%), and around $9.1 billion in 'other financing,' including obligations arising from the failed sale-leaseback of AI infrastructure. Interest expense from X-related loans alone amounts to roughly $1.2–1.3 billion annually, allocated to the AI segment.

The $19.6 billion EchoStar spectrum commitment will be fulfilled by November 2027. The deal involves equity and cash consideration in exchange for 65 MHz of U.S. spectrum and a global mobile satellite service license. This represents a binding capital commitment beyond the bridge loan and fiscal year 2026 capex.

The option agreement with Cursor could trigger termination fees as high as $10 billion. In April 2026—one month before this S-1 filing—SpaceX signed a compute and option agreement with Anysphere (Cursor), implying a $60 billion valuation for Cursor. If either party terminates the agreement, SpaceX must pay Cursor a $1.5 billion termination fee plus $8.5 billion in deferred service fees, payable in cash or Class A shares.

The $45 billion Anthropic contract is the AI segment’s largest single external revenue source. A cloud services agreement signed in May 2026 requires Anthropic to pay $1.25 billion per month through May 2029. SpaceX is selling its COLOSSUS compute capacity to Anthropic, a direct competitor in frontier models, creating extreme counterparty concentration risk.

The balance sheet recognized a $530 million litigation reserve for the Grok image generation class-action lawsuits—Jane Doe v. X.AI Corp. (January 2026), Jane Doe 1 (March), and the Baltimore case (March). Plaintiffs are seeking compensatory, statutory, and punitive damages. The S-1 explicitly states that the range of potential additional losses cannot be estimated.

Q1 2026 revenue growth slowed to 15.4% ($4.69 billion vs. $4.07 billion year-over-year), down from the full-year 2025 growth rate of 33.2%.

SpaceX will be a controlled company with four classes of equity. Musk holds majority voting power post-IPO. The company will rely on Nasdaq’s controlled company exemption, waiving requirements for an independent compensation committee and an independent nominating committee.

Adjusted EBITDA has been inflated by approximately $9 billion. Management’s headline figure for 2025 is $6.6 billion in 'adjusted EBITDA,' compared to a GAAP operating loss of negative $2.6 billion. Adjustments exclude depreciation, stock-based compensation, and segment-specific exclusions.

SpaceX (Space Exploration Technologies Corp.; SEC CIK 0001181412) designs and operates reusable rockets, the world’s largest LEO satellite constellation (approximately 9,600 broadband satellites plus about 650 direct-to-cellphone satellites), and—as of its acquisition of xAI in February 2026—gigawatt-scale AI training infrastructure. It reports three segments: Space, Connectivity (10.3 million Starlink subscribers), and AI (Grok models, the X social platform with 550 million monthly active users, and the COLOSSUS/COLOSSUS II compute clusters). Full-year 2025 revenue was $18.7 billion; GAAP operating loss was negative $2.6 billion; cash on hand stood at $15.85 billion against $29.1 billion in long-term debt as listed on the capitalization table cover page.

The corporate chain warrants retracing. SpaceX acquired xAI in February 2026. xAI had acquired X Holdings in March 2025. X Holdings acquired Twitter in October 2022. Result: Twitter/X is now embedded within SpaceX’s AI segment, with its own balance sheet items, its own litigation exposure, and its own debt structure.

Scale: Supported 1.3 billion accounts over the past 12 months, with 550 million monthly active users (up from 520 million in December 2025) and 350 million posts per day. Of these monthly active users, 117 million use Grok features—X serves as the model’s primary distribution channel. The Money suite (payments, banking, financial services) launched in beta in November 2025 and is progressing toward full availability. X Ads Manager began a phased rollout in April 2026.

Financial contribution: The AI segment’s revenue in 2023–2024 came almost entirely from X—advertising, X Premium subscriptions, and data licensing. Advertising revenue alone declined by $595 million year-over-year in 2024 due to 'X losing advertising partners,' partially offset by a $157 million increase in X Premium subscription revenue and a $90 million rise in data licensing revenue.

Adding the $20 billion SpaceX bridge loan (due September 2027) and the $9.1 billion in 'other financing' brings total long-term debt to approximately $42 billion—not the headline $29 billion figure shown on the capitalization cover page.

X-specific risks absent from SpaceX's other businesses. Enforcement of the EU’s Digital Services Act against very large online platforms. Advertiser brand safety reversibility on short-term ad contracts that can be canceled at any time—the mass exodus of 2024 could replay within a single news cycle. Money products trigger payment/money transmission/banking regulations in all 50 U.S. states and every foreign jurisdiction. Reversals in content moderation policies could simultaneously trigger advertiser suspensions and user migration.

This comparison table was assembled in real time during analysis by paying $0.10 to Jintel’s GraphQL endpoint to retrieve bulk fundamental data for all five comparable companies. No Bloomberg terminal or FactSet contract required.

ASTS operating margin reflects pre-revenue heavy investment. Source: Retrieved via Base chain x402 from Jintel entitiesByTickers, retrieval date May 22, 2026.

Interpretation of the peer group: Rocket Lab’s 104x price-to-sales ratio is the closest narrative benchmark—investors are willing to pay extremely high multiples for scalable, reusable launch capability plus LEO optionality, even with negative margins. SpaceX should command a higher multiple than RKLB, but blindly applying 104x to SpaceX’s connectivity-only revenue of $11.4 billion implies an equity valuation of $1.2 trillion, which cannot be anchored to anything tangible. AST SpaceMobile’s 345x is purely a pre-revenue narrative valuation, serving only as an upper-bound reference for direct-to-cellphone optionality. Iridium’s 7.4x sales and 14.8x EBITDA represent what mature, profitable LEO communications looks like—applying 7.4x to Starlink’s $11.4 billion in revenue yields an $84 billion standalone valuation for Starlink (the bear case anchor). NVIDIA’s 31.7x EV/EBITDA corresponds to 85% revenue growth, which is the level the AI segment would need to reach to justify a fundamentals-based valuation. It hasn’t gotten there yet.

Notable signal: Rocket Lab filed a 424B5 prospectus supplement on May 20, 2026—the same day SpaceX released its S-1. RKLB issued secondary equity during SpaceX’s news cycle, indicating management believes the IPO window is open and competitive supply pressure is imminent.

Each of these four items is individually material and mutually reinforcing. Two were signed within 60 days prior to this S-1 filing.

Why this matters for valuation: A clearer view of 'adjusted net obligations' is as follows: $42 billion in total debt plus $19.6 billion in EchoStar commitments plus up to $10 billion in contingent liabilities related to Cursor, minus $15.85 billion in cash on hand, equals approximately $55 billion in net obligations—before accounting for any IPO proceeds. This is three to four times the figure implied by a cursory reading of the capitalization cover page, materially altering the bear case.

Method 1 – Based on transaction multiples for the Connectivity segment alone, as it is the only segment with positive standalone economics.

This point is buried deep within the underwriting section and rarely covered in news reports, yet it is significant. The affiliated parties of the five lead underwriters (Goldman Sachs, Morgan Stanley, Bank of America, Citi, and JPMorgan) plus five additional bookrunners (Barclays, Deutsche Bank, Royal Bank of Canada, UBS Group, and Wells Fargo & Co) were all lenders in SpaceX's USD 20 billion bridge loan, and they are now pricing the IPO intended to refinance that loan. Morgan Stanley also separately acted as advisor to SpaceX on its acquisition of xAI (which was funded by the bridge loan). The underwriting syndicate has a direct financial incentive to maximize the IPO proceeds. This should prompt the investment committee to remain vigilant on pricing discipline.

No single relationship appears alarming on its own. What’s concerning is the density—entities controlled by Musk have at least nine distinct financial touchpoints with SpaceX. Public company governance committees typically review one or two such relationships. Here, it’s an order of magnitude higher.

Upgrade to overweight if the deal prices at an implied equity valuation of USD 350 billion or below, Starship achieves commercial payload delivery in the second half of 2026 as guided, and Connected Services revenue growth exceeds 40% year-over-year in Q2 2026.

Downgrade to avoid if the deal prices above USD 510 billion, or if a Starship vehicle loss event delays V3 satellite deployment beyond 2027, or if the AI division’s cash burn accelerates in Q2–Q3 2026 to an annualized operating loss exceeding USD 8 billion, or if the FAA imposes a prolonged grounding of Starship.

D+1: First-day price performance benchmarked against comparable IPOs

D+30: First quarterly earnings report (Q2 2026) — triggers early release of lock-up tranches (20% released immediately; an additional 10% released if share price is +30% above IPO price)

D+70, +90, +105, +120, +135: Phased early releases of lock-up tranches, each releasing 7%

D+90: End of quiet period; sell-side analysts initiate coverage

D+180: Expiration of all standard lock-up tranches

Second half of 2026: Starship guidance targets achievement of commercial payload delivery

Q2–Q3 2026: Procedural milestones in the Grok image-generation class-action lawsuit (monitor whether the $530 million reserve is increased)

April 2027: First anniversary of the Cursor option agreement — watch for exercise or termination signals

September 2027: Maturity of SpaceX’s $20 billion bridge loan (must be refinanced or repaid)

November 2027: Completion of the $19.6 billion EchoStar spectrum transaction — global rollout of V2 Mobile is contingent on this

May 2029: $45 billion Anthropic compute contract expires; renewal terms will define the economics of the AI segment for years to come

October 2029: Combined $12.7 billion in Company X’s B-1 and B-3 term loans mature

SpaceX S-1, SEC filing number 0001628280-26-036936, filed on May 20, 2026

Real-time comparable fundamentals via Jintel GraphQL entitiesByTickers, on Base chain x402, retrieved on May 22, 2026

Real-time consolidated SEC profiles via x402helper /companies/profile for RKLB, IRDM, VSAT, retrieved on May 22, 2026

Industry IPO context via Parallel Search, on Base chain x402, retrieved on May 22, 2026

Four scenarios for a SpaceX IPO – Acadian Asset Management

Generated by the IPO analysis package on agentic.market. Six paid x402 calls. $1.87 USDC on Base chain. No API key required. No registration needed. Pay-per-request.

A Bloomberg Terminal seat costs $24,000 per year. This memo demonstrates what agents can now produce when they can pay for data themselves.

142K Views