SpaceX Just Became the Sixth Hyperscaler. Wall Street Sold It. Here's Why That's Bullish for Everything.

SpaceX released its first quarterly earnings as a public company tonight. The revenue beat was clean — $6.93 billion versus $6.88 billion consensus. Every segment outperformed. Starlink connectivity revenue hit $4.29 billion against $3.83 billion expected. AI revenue came in at $2.56 billion against $2.18 billion expected. Space revenue topped at $962 million against $835 million expected.

The stock dropped 6% after hours.

Not because the business is struggling. Because the capital expenditure number was so large it forced a recalculation of what SpaceX actually is.

$18.4 Billion in One Quarter

Wall Street expected $13.2 billion in Q2 capex. SpaceX spent $18.4 billion — 40% above consensus and more than six times the $2.8 billion spent in the same quarter last year. Of that $18.4 billion, approximately $15.8 billion went directly to AI infrastructure.

To put that in context: SpaceX spent more on AI infrastructure in a single quarter than most Fortune 500 companies spend on everything in a year. The Q1 AI capex was $7.7 billion. It doubled in one quarter. Annualized, SpaceX is now running at roughly $60-70 billion in total capex, with the vast majority flowing into AI compute.

That makes SpaceX the sixth hyperscaler — alongside Amazon ($200B), Google ($175-205B), Meta ($125-145B), Microsoft ($120B), and Apple ($100B). A rocket company that went public two months ago is now spending at a rate that places it among the largest AI infrastructure investors on earth.

The market sold it because the loss per share was $0.26 and the cash burn is enormous. But the AI infrastructure trade doesn't care about SpaceX's EPS. It cares about where $15.8 billion per quarter is being deployed — and the answer is the exact same semiconductor and hardware supply chain that StockGecko tracks.

Follow the Money

Every dollar of SpaceX's AI capex flows into the same companies that just posted record earnings:

SK Hynix and Samsung supply the HBM and DRAM. Micron supplies the NAND storage. Nvidia supplies the GPUs. Broadcom and Marvell supply the networking silicon. Bloom Energy and other power providers supply the energy infrastructure. CoreWeave provides supplemental cloud compute.

When SpaceX spends $15.8 billion on AI infrastructure in a single quarter, the revenue shows up in those companies' earnings one to two quarters later. The semiconductor supply chain doesn't differentiate between a dollar from Google and a dollar from SpaceX. It's all demand. And the demand just increased by $60+ billion annualized from a single new entrant.

This is the part the market is missing by selling SpaceX down 6% on the capex number. The capex isn't a cost problem for SpaceX alone — it's a revenue tailwind for every company in the AI supply chain. SpaceX's loss is SK Hynix's revenue. SpaceX's cash burn is Nvidia's backlog. The money doesn't disappear. It flows downstream.

The Hyperscaler Capex Scoreboard

The updated numbers after SpaceX's report:

Amazon: $200 billion (2026 guidance). Google: $175-205 billion (raised twice). Meta: $125-145 billion. Microsoft: $120 billion (cash flow positive through FY2027). Apple: ~$100 billion (announced February 2026). SpaceX: ~$60-70 billion (annualized from Q2 run rate).

Combined: approximately $780-845 billion in 2026 alone. That's up from the $725 billion figure we've been citing — SpaceX just added $60+ billion to the total that wasn't in most estimates because nobody expected them to spend at this rate.

Projected 2027: over $1 trillion. Goldman's upside scenario: $1.4 trillion. Morgan Stanley estimates $1.5 trillion in new tech-sector debt issuance to finance the buildout.

Every time someone says "the AI trade is over," another company adds tens of billions to the capex pile. The spending is accelerating, not decelerating. The companies doing the spending have more data on AI demand than any analyst, any trader, or any account on Twitter. They are collectively betting $800+ billion this year that AI infrastructure generates returns.

Starlink Is Funding the Machine

The financial structure of SpaceX is worth understanding because it mirrors a pattern we're seeing across the AI buildout.

Starlink — the satellite internet business — generated $4.29 billion in Q2 revenue with $1.66 billion in operating income. It runs at roughly 37%+ operating margins with about 63% EBITDA margins. Nearly 12 million subscribers paying recurring monthly fees. This is a profitable, scaling, cash-generating business.

The AI segment generated $2.56 billion in revenue but posted a $1.26 billion operating loss. The Space segment (launches, Starship) generated $962 million but lost $542 million.

The structure is clear: Starlink's profits fund AI and Starship's losses. The profitable connectivity business subsidizes the capital-intensive growth businesses. This is the same model Amazon used for two decades — AWS profits funded everything else. The market punished Amazon's margins for years before recognizing that the spending was building the most valuable cloud platform in the world.

SpaceX is running the same play. Starlink is AWS. AI infrastructure is the next growth engine. The market is punishing the margins. The question is whether it's right to — or whether it's making the same mistake it made with Amazon.

The Unlock Adds Fuel

Two days from now, on August 6, the first lockup window opens. Approximately 911.5 million shares — 20% of the 180-day tranche — become eligible for sale. That's roughly $104 billion in stock at current prices, and it hits a company that's already down 50% from its post-IPO high and 20% below its IPO price.

As we wrote last week, the unlock window is where max fear peaks. The stock is down from $225 to $105. First earnings just printed with a capex number that spooked the market. And the largest share unlock in capital markets history hits in 48 hours.

For traders tracking SpaceX on StockGecko through the Hyperliquid perp, this is the convergence of every catalyst that creates asymmetric setups: earnings reaction, lockup fear, and a company spending $60+ billion annually on AI infrastructure that directly feeds the semiconductor supply chain.

The Stock Is Not the Trade

Here's the distinction most coverage is missing.

SpaceX stock dropping 6% on a $18.4 billion capex print is the market saying "this company is spending too much." That may be a valid concern for SpaceX equity holders dealing with dilution, cash burn, and a $1.4 trillion valuation on negative earnings.

But for the AI infrastructure trade — for the semiconductor companies, memory manufacturers, power providers, and networking suppliers tracked on StockGecko — SpaceX's capex isn't a risk. It's revenue.

The market is selling the company that's spending the money. It should be buying the companies that receive the money.

SK Hynix at 5x forward P/E. Samsung at single-digit multiples after 19x earnings growth. Micron after a 345% revenue increase. Bloom Energy rebounding 14.5% with 79% funding rates. These are the beneficiaries of the same $18.4 billion that just made SpaceX stock drop 6%.

When a new $60 billion annual spender enters the AI infrastructure market, the immediate reaction is to sell the spender. The smart reaction is to buy the supply chain. The money doesn't vanish — it flows downstream into the exact companies that are already posting record earnings.

The Bottom Line

SpaceX just spent $18.4 billion in a single quarter on AI infrastructure. The stock dropped 6%. The AI supercycle just added another $60+ billion in annual demand to a supply chain that is already sold out through 2027.

Total hyperscaler capex is now approaching $800-845 billion in 2026 — up from $725 billion before SpaceX's print. Projected to exceed $1 trillion in 2027. Every quarter, the number gets bigger. Every quarter, someone declares the AI trade is over. Every quarter, the companies doing the spending disagree by writing larger checks.

SpaceX's first public earnings report didn't just reveal a rocket company moonlighting as an AI investor. It confirmed that the AI infrastructure buildout is widening, not narrowing — and the money flows directly into the assets tracked on StockGecko.

The AI trade isn't over. It just added a sixth hyperscaler.

Track every AI infrastructure equity on StockGecko.

Prices and data referenced as of August 4, 2026. This is not financial advice.