SpaceX (NASDAQ: SPCX) is an exceptional business. At $2+ trillion, it is also a significantly overvalued one. The debut of SpaceX that we saw on June 12 made history for the whole market. The price for the share was $135, and it is officially the biggest IPO in history.\Elon raised around $75 billion, and the valuation of the company was ~$1.8 trillion. Right now, the market cap has crossed $2.9 trillion (June 16th).\After a three-day rally that lifted shares 67% above the $135 IPO price to a peak of $225.64, SPCX fell nearly 5% on June 17. The pullback is a preview of what the next six months look like structurally. I’m bearish on the 12-18 month horizon with a specific window in late 2026 into early 2027.For investors thinking beyond the first weeks of a bullish mood, it is very important to separate 3 things:What happened on the first dayWhat's going to happen in 6 monthsWhat this IPO tells us about the broader market coming in 2027\My point is that this is likely to go higher in the short term, as the index inclusion and oversubscription can fuel the whole company. But across the 2026 mega-IPO wave (especially in the AI sector), it looks like the market is absorbing a huge amount of retail liquidity at the worst possible time for us.What You’re Actually BuyingLet's look wider into the narrative with SpaceX and three businesses priced as one:Starlink is the real company. $11.39 billion in 2025 revenue, $4.42 billion in operating income, 63% adjusted EBITDA margins. By March 31, 2026, subscribers reached 10.3 million across 164 countries.Launch services generated $4.1 billion in 2025, growing just 8% year-over-year, primarily from Pentagon and NASA contracts. Most capacity goes to internal Starlink deployment.xAI spent $6.4 billion in 2025 on $3.2 billion in revenue.And now there’s a fourth: SpaceX announced on June 16 the acquisition of Anysphere. It is the company behind the AI coding tool Cursor, and they bought it for $60 billion in an all-stock deal, just four days after the IPO. SpaceX's Four Businesses, Priced as One (Original chart created by Timur Mekhantev, based on data from SpaceX's S-1 filing (via New Space Economy, Tom Tunguz, GraniteShares), TechCrunch, and TradingKey reporting on the Cursor acquisition.)The Valuation ProblemAt roughly 94 times 2025 revenue, the valuation shows that investors accept SpaceX as a flawless SaaS monopoly with a durable AI.\Just for context, Viasat and Eutelsat, mature satellite operators, trade at 2-3x revenue. Applying a generous 20x multiple to Starlink’s $11.4 billion in revenue gets you roughly $230 billion for the core business. The remaining ~$1.6 trillion is optional (Starship, xAI, Cursor, Anthropic compute deals), priced as certainties rather than bets.\The former Nasdaq chief publicly warned that SPCX is not trading on fundamentals, because it’s trading on narrative, float scarcity, and momentum.The ARPU ProblemBlended ARPU is at ~$92 per month implied from S-1 math, but that number is being pulled down fast. SpaceX is adding subscribers in Africa, Southeast Asia, and Latin America at $40-80 per month versus $120+ in the US.\Even as the subscriber base more than doubled, operating income in Q1 barely moved, rising from $1.03 billion to $1.19 billion.\The margin signal points in the same direction. Q1 operating income rose from $1.03B to $1.19B (it is up just 16%) while the subscriber base more than doubled year-over-year.\That's the arithmetic of a business whose marginal subscriber is increasingly cheap to acquire but also increasingly cheap to serve: satellite capacity costs don't scale down with lower-ARPU geographies the way revenue does. If that ratio holds, full-year operating margin could compress from the current ~39% toward the low-30s % by 2027\Justifying a $2 trillion valuation at current ARPU requires somewhere between 150-200 million subscribers. That represents virtually the entire rural broadband access worldwide.The Competition Nobody Is Pricing InStarlink has operated without serious competition for most of its existence.\Amazon Leo (formerly Project Kuiper) confirmed a mid-2026 commercial launch, with enterprise revenue commitments already secured from governments and corporations.\If enterprise/aviation/maritime represents even 15-20% of Starlink's connectivity revenue (a reasonable estimate given the S-1's framing) at 2-3x consumer ARPU, that's roughly $1.7-2.3 billion of 2025's $11.4B in revenue concentrated in the exact segment with a credible, well-funded competitor in 2026.\Losing even 15% of that pipeline to Leo could be up to $250-350 million of annual revenue at risk, concentrated in the highest-margin part of the business.\The service integrates natively with AWS, offering enterprise customers a data and AI analytics advantage that has nothing to do with satellite performance. Big partners, including Delta, JetBlue, and AT&T, have already signed on.\Amazon has committed $10 billion to the program and claims Leo will offer download speeds roughly double Starlink’s typical throughput. It’s one of three companies in the world with the logistics, capital, and cloud ecosystem to compete at scale.\At 94x revenue, there is no margin for a competitor taking even 15% of Starlink’s enterprise pipeline.The Starship DependencyThe next phase of Starlink growth, V3 satellites, requires Starship. SpaceX flew five Starship missions in 2025 against a target of 25, and completed its 12th flight test on May 22, 2026.\Without Starship reaching operational status, V3 deployment slows, capacity growth stalls, and service quality in high-density markets gets worse. The valuation assumes Starship works on schedule, but the execution record says otherwise.The Lockup CalendarThe IPO structure is performance-linked. It means that the better the stock is, the faster insiders can sell it.September 2, 2026: 20% insider unlock kicks in two trading days after the report. A further 10% unlocks early only if the stock holds above $175.50 (30%+ over IPO) for 5 of the prior 10 sessions.December 8, 2026: Standard 180-day lockup expires (~December 8)June 2027: Elon Musk’s shares unlock SpaceX (SPCX) Insider Lockup Unlock Schedule (Benzinga)The pullback on June 17 arrived exactly one day after SPCX options began trading, giving bears their first practical tool to bet against the stock.\Historical precedent: Palantir fell 6-13% (depending on the session window) in a single session when Peter Thiel sold at lockup expiration. Uber hit ATL on its lockup day. Rivian fell 20% when Ford disclosed its stake sale. IPO Lockup Expiration Precedents — Single-Day Price Moves (Sharewise)Cursor Deal Changes the MathFour days after the largest IPO in history, SpaceX spent $60 billion in stock to buy an AI coding tool with a few million users and no disclosed revenue.\Management had access to the most overvalued currency in the market (freshly issued public stock trading at 67% above IPO price) and deployed it immediately into an acquisition that public shareholders had no vote.\When companies use richly valued stock to buy growth, it usually means one of two things:The organic growth trajectory doesn’t justify the valuationManagement sees the current stock price as an opportunity to acquire real assets with inflated currency before the price corrects.\Neither of these interpretations is optimistic for those buying SPCX shares today. The problem is that it reveals how SpaceX management views the current share price relative to fair value.Risks to the Bear CaseThe main risk is timing. Vanda Research believes SpaceX is becoming a core retail asset that could join the Magnificent Seven alongside OpenAI and Anthropic to form what some are calling the “FAB 10.”\If that retail bid is structural, then lockup expirations get absorbed rather than sold.\If Starship execution grows, the V3 timeline moves forward. If Amazon Leo underwhelms at launch, competitive pressure is slower than I model. If the Fed cuts aggressively into year-end, risk appetite stays elevated.\I’d revisit the bear case if SPCX holds above $200 through the September earnings without meaningful insider selling pressure.What is the real worth of SpaceX?Let’s build a simple sum-of-the-parts. Starlink at $15.5 billion projected 2026 revenue, applying a 15x multiple gets you to roughly $230 billion. Apply 20x, and you get $310 billion. Even at 25x, a multiple reserved for the fastest-growing internet infrastructure companies, Starlink alone is worth $390 billion.\Launch services at $4.1 billion revenue, 8x multiple given the structural growth cap: ~$33 billion.xAI and Cursor: pre-revenue at scale, burning cash, with no public comps.\A charitable $150-200 billion is aggressive but defensible if you believe the AI thesis.\That gets you to a range of roughly $400-600 billion on fundamentals, with $200 billion of AI optionality on top. Call it $600-800 billion as a reasonable bull case for the actual business.ConclusionStarlink is an exceptional company, and the launch business is a genuine moat. xAI and Cursor might pay off, but $2+ trillion with declining ARPU, a well-funded competitor entering the market, a $60 billion dilutive acquisition in week one, and a lockup structure that rewards insiders for selling into strength is not the right entry point.\The stock is still 37% above its IPO price after the first pullback. That’s your exit window, not your entry point.\