SpaceX's $13 Billion AI Bet Is Testing Wall Street's Patience

Starlink profits now have to justify SpaceX's AI spending

SpaceX is no longer being judged on vision alone. After more than $500 billion in market value lost since its debut, the stock was trading at roughly $111 per share versus the $135 IPO price. That setup leaves less room for abstract future upside: investors are watching whether SpaceX can fund its ambitions without straining the business that is already producing profits.

What the first earnings report has to show

This quarter matters because it is the first public read-through of the AI-Starlink-launch combination, and SpaceX has not provided formal Q2 guidance. That leaves investors relying more on estimates than commitments. The core tension is straightforward: total Q2 capital expenditure is forecast at $14.05 billion, with the AI segment alone absorbing $10.2 billion, while Starlink is the unit expected to generate the clearest profit bridge. If that profit pool holds up, the market may tolerate a longer payback period. If it weakens, the valuation likely comes under pressure quickly.

The balance sheet is not the problem; the payoff map is

The immediate question is not whether SpaceX can carry the spend. It can, at least for now, with $93.78 billion in cash. The harder question is what that spending will turn into. Estimates for the quarter point to $6.81 billion in revenue and $8.52 billion in total costs and expenses, while capex is projected at $18.58 billion. Because liquidity does not appear immediately constrained, investors are likely to focus less on survival and more on whether the spending can show a credible path to returns.

AI revenue still looks small relative to the buildout

The disconnect is most obvious in AI itself. Estimates cluster around $2.08 billion in AI segment revenue, an $15.5 million adjusted EBITDA loss, a $2.39 billion operating loss, and $13.09 billion in AI capex. That is why the skeptical read is so easy to make: the buildout is already large, while the business being built is still early. Reuters quoted one critic arguing that Starlink cannot single-handedly fund a $30 billion annualized AI capex program.

The bull case is not just that AI could be big. It is that SpaceX may be buying scarce infrastructure tied to a wider vision that includes frontier models, software, and eventually space-based compute. Reuters also noted the upside case that compute contracts could exceed $25 billion annualized once fully ramped. Even so, the market is no longer debating whether AI matters. It is debating whether SpaceX is early enough to justify the spend now.

What investors need from management

Wall Street needs a timing framework, not just a long-term narrative. The key questions are:

• When could AI revenue begin covering AI spending?
• Where is peak spending, and how long must other units help bridge the gap?
• Can Starlink support that bridge without leaving less room for error?

That is why Starlink remains the make-or-break unit. Estimates point to roughly $3.82 billion in connectivity revenue to $3.88 billion in connectivity revenue, with about $1.42 billion in operating profit cited across estimates. If management can show that profit pool is stable enough to support the AI build, the market may accept a longer payoff period. If not, the valuation likely compresses on timing as much as on sentiment.

The disclosure that could change the debate

Investors also need more visibility into how much of SpaceX's compute partnerships are showing up in reported revenue. For now, the debate is still leaning on estimates and narrative. More concrete disclosure would help the market move from psychology to underwriting: if revenue recognition improves and spending milestones become clearer, investors can start modeling returns rather than just financing risk.