The Space Company Went Public. The Search Company Was Already Printing Money.

There was a time when the advice to buy Alphabet instead of SpaceX made sense because you couldn't buy SpaceX at all. That constraint expired on June 12, 2026. SpaceX went public at $135 a share, raised $75 billion, and briefly hit a $2.6 trillion market cap. Then the stock fell over 40% from its peak and broke below its IPO price in mid-July. The old comparison is dead. A new one has appeared in its place.

Both companies are now public. Both are building future industries - one in orbit, one on the street. Both have a profit-generating core that bankrolls a money-losing moonshot. The question is no longer about access. It's about which business model actually works.

Let's start with the part that's easy to miss. Alphabet reported $9.11 per share in Q2 2026, against a consensus estimate of $2.88. That is a 217% beat. Q1 was similar - $5.11 actual versus $2.63 expected. Alphabet's core advertising business, which most people think of as mature and predictable, just produced its two biggest earnings surprises in company history. Revenue came in at $119.8 billion for Q2, beating the $117.1 billion consensus. The ad engine is still adding torque.

This matters because Alphabet's moonshot - Waymo, the self-driving car unit - is funded by that ad engine. Waymo is real: 500,000 weekly rides, more than $350 million in annual recurring revenue, 125 million autonomous miles logged on U.S. roads. It is also still a net loss. Alphabet's "Other Bets" segment, which includes Waymo, burned roughly $5.5 billion in 2025. Waymo's co-CEOs are targeting 1 million weekly rides this year, but fleet math makes that tight - the current utilization rate of about 20 trips per vehicle per day means hitting a million rides requires roughly 7,200 cars, and the fleet isn't there yet.

Now SpaceX. Revenue was $18.7 billion in 2025, up 33% year over year. Starlink, the satellite internet business, generated $11.4 billion of that - 61% of total revenue - and 10.3 million subscribers across 160 countries. Starlink alone was profitable, with operating profit of $4.4 billion. The problem is the rest of the company. SpaceX lost $4.9 billion in net income for 2025. A large part of that loss came from xAI, the AI company SpaceX acquired in an all-stock deal in February 2026. xAI lost $6.4 billion in 2025. The combined entity is profitable only if you count Starlink and pretend the other segments are temporary.

At a market cap near $1.8 trillion, SpaceX trades at roughly 46 times estimated 2026 sales. Or 95 times 2025 sales. You're paying for a company that lost money last year and is projected to lose money again, priced as if it were already the infrastructure layer of the next economy. There are reasons to believe in the trajectory - SpaceX signed deals worth roughly $26 billion annually in AI compute capacity with Alphabet and Anthropic. If those contracts hold, the revenue base expands dramatically. But the Anthropic deal can be terminated with 90 days' notice. And the company's GAAP first-quarter 2026 loss was $4.3 billion.

Here's the contradiction I kept coming back to. Both companies are option portfolios: a cash cow funding a bet on something that could be bigger than the cow. Alphabet's cow is search and video advertising. SpaceX's cow is Starlink subscriptions and government launch contracts. Alphabet's bet is driverless cars. SpaceX's bet is AI compute infrastructure and orbital economics. The difference is that one cow has been getting bigger and richer while the bet loses money, and the other cow is solid but the bet is a black hole that the company has been honest about.

Most people think the more exciting company is the better investment. That assumption doesn't survive contact with the financial statements. The more interesting question is whether you want to own a business that generates cash while building its future, or one that has to borrow against its future to build it.

SpaceX has a structural complication that makes the second problem worse. About 95% of shares are locked up by insiders. A tiered release schedule starts in August with 20% unlocking after Q2 results, then another 28% after Q3 results, with Elon Musk's roughly 42% of the company locked until June 2027. Twenty-eight percent of the free float is already lent out for short selling. The stock can't establish a real price until it has a real float.

Alphabet doesn't have that problem. It trades freely. Its valuation is a multiple of earnings that are growing at a pace most people didn't expect. AInvest's aggregate signal labels GOOG a Buy, with its fundamental and liquidity ratings both in the upper tier - which is what you'd expect for a company whose Q2 EPS was three times what analysts predicted.

None of this is an argument that SpaceX is a bad company. It's an argument that it's an expensive company whose stock price hasn't finished finding its level. The 40%+ decline from the IPO high is the market working through the gap between narrative and cash flow. That process may not be finished.

I haven't seen current channel inventory data for either stock, and I can't tell you where SpaceX trades in three months. But I can tell you what to watch. If you're holding Alphabet, the test is whether the ad business can sustain its earnings trajectory when the macro environment shifts - and whether Waymo's new vehicle, the Ojai, can cut per-car costs from $150,000 to $32,000 fast enough to turn the loss leader into something else. If you're holding SpaceX, the test is whether Starlink's revenue growth can absorb xAI's losses and whether the AI compute contracts with Alphabet and Anthropic are durable revenue or 90-day cancellable leases.

The deeper point is simpler. When two companies are both building the future, the one that's already paying for it is the one that gets to decide what the future looks like. The other one is hoping the future arrives before the bills do.