'A game of hot potato': An early SpaceX pessimist speaks with Oninvest on key risks

One of Wall Street’s early skeptics on SpaceX, CFRA analyst Keith Snyder, cautions—despite the company’s generally strong first-quarter earnings—that investors are still paying for a compelling narrative rather than proven business growth. In his view, one of the main risks for SpaceX is the hype around AI. Oninvest spoke with Snyder about how he sees the company.

Keith Snyder remains pessimistic on the space company, a rare stance on Wall Street: three out of four analysts, as of August 6, assign the stock a "buy" rating. Snyder, however, was the first to assign SpaceX a “sell,” all the way back in December 2025, setting a price target of $115, which came in 15% below the IPO price. According to Snyder, SpaceX’s second-quarter revenue exceeded his forecast, but “not by much, so the revenue growth was no big surprise.” The company reported revenue of $7.8 billion versus a consensus estimate of $6.9 billion.

The CFRA analyst was surprised though by the growth of the Connectivity (Starlink) segment, which he sees as the engine behind SpaceX currently. Revenue from satellite internet services, SpaceX’s only profitable business, rose 66% year over year to $4.29 billion versus an expected $3.83 billion. The number of subscribers doubled in the second quarter to 12 million.

Snyder reacted coolly to growth in the AI business: it had already been telegraphed by announcements of three contracts with Google, Anthropic, and Reflection for the lease of AI servers, so this had already been priced into expectations somewhat.

In his comments to Oninvest, Snyder said the main risk for the company lies in the AI segment, as SpaceX is investing in it based on the assumption that the AI bubble will continue to grow. If that does not materialize, all these investments will have gone to waste.

“The way I’m thinking about all of this is, why would a company like Google sign a deal with SpaceX for compute? Google being someone who is more than capable of building out its own data centers. It's because they don’t want to be left holding the bag when the bubble pops. In the same vein, why would Apple, the second largest publicly traded company, choose not to invest in AI itself and rather rent it out from other companies? This is a game of hot potato right now, and the smart companies are avoiding being left holding the bag,” Snyder said.

Following the earnings, Snyder trimmed his 2026 revenue growth forecast, as it was clear that it was too aggressive, but raised his 2027 estimate. He has not yet disclosed the figures.

Overall, he sums up his view on the stock like this: “I still feel that most investors and analysts are simply hoping for the growth to materialize. Hope is not a viable investment strategy; I need to see the growth and see a realistic path to said growth. Right now, all that we are seeing is a good story, and it is only a matter of time to see if it's fact or fiction.”

On August 4, SpaceX reported a 92% year-over-year increase in second-quarter revenue to $7.8 billion. Adjusted EBITDA increased 191% to $3.5 billion, while the net loss was nearly halved to $541 million.

The company's stock fell 12% following the results. The market was wary amid the scale of investments in AI, as well as the anticipated pressure on the stock price owing to the end of the lock-up period for insiders on August 6.

Against this backdrop, most Wall Street investment firms and banks maintain “buy” ratings at target prices significantly above current prices. Snyder argues that at such valuations, investors have already paid for several complex and as yet unproven successful scenarios for the company's future, including Starlink, space-based data centers, and the AI business.