SpaceX stock (NASDAQ:SPCX) trades at $116 and Cerebras Systems (NASDAQ:CBRS) at $193. On today’s earnings, both look absurd. SpaceX stock is on 597 times forecast 2026 earnings; Cerebras is still loss-making.

But that’s not how I’d value either. Both listed this year, both are pre-scale, and with companies like these the whole question is what the multiple looks like once the growth arrives.

The tedious bit that matters most

I know valuation tables are dull. Read them anyway — this is where the investment case lives, not in the rocket footage (as much as I like to watch the launches).

The consensus forecast has SpaceX earning $4.69 a share in 2028 and $10.80 by 2030 — forward price-to-earnings (P/E) ratios of 24.8 times and 10.8 times. Cerebras gets there faster: $5.85 in 2028 and $16.55 in 2029, taking it from 32.9 times to 11.6 times.

Sales tell the same story. SpaceX sits on 38 times 2026 revenue, falling to 4.1 times by 2030. Cerebras drops from 49 times to 2.4 times. For context, the S&P 500 trades around three times sales.

Why SpaceX could deliver

Starlink is the engine in the near-medium term, not data centres in space. It passed 12m subscribers in June, having roughly doubled in a year, and Quilty Space expects 16.8m by December on $11.3bn of consumer revenue. Payload’s estimate is nearer 18.4m.

The 2030 forecasts are where it gets interesting — and wildly uncertain. MoffettNathanson models 70m subscribers; Morgan Stanley 116m. Around 2.6bn people still lack reliable broadband, Starlink has no comparable rival in orbit, and direct-to-cell opens a second market entirely.

The issue is that those forecast revenue numbers involve near-flawless execution — $365bn of revenue by 2030 against roughly $39bn this year.

What’s more, Starship has to work, and work cheaply. The Space Shuttle put a tonne into orbit for roughly $54m. Falcon 9 does it today for about $2.7m. Starship is targeting a few hundred thousand. This is a big part of the longer-term value.

And Cerebras

Cerebras’s wafer-scale chips lead on AI inference — running models rather than training them — which is where compute spending is shifting. Very simply, many people think these chips offer far more efficiency than peers (Nvidia and AMD) for certain tasks, and that opens up a runway of growth.

However, to meet the forecasts, it needs roughly 20-fold revenue growth in four years while competing with Nvidia and AMD, and very few analysts model beyond 2028. If execution slips, that 49 times sales never falls.

Both are well capitalised

Cerebras holds $2.23bn of cash against $1.36bn of debt — net cash. SpaceX carries $30.6bn of debt against $23.7bn of cash, but net debt under $7bn on a $1.5trn company is a rounding error. This is good news for companies needing to scale.

My take

Neither stock is cheap today. Both are cheap on 2029 numbers — and that gap is the entire bet. And I’d suggest that the growth trajectory would carry on well beyond 2029 — these are really stocks for the 2030s. I own both, and have been buying on pullbacks — they’re both very volatile. I think they’re worth considering.

Should you invest £5,000 in SpaceX right now?

When investing expert Mark Rogers and his team have a stock tip, it can pay to listen. After all, the flagship Twelfth Magpie Share Advisor newsletter he has run for nearly a decade has provided thousands of paying members with top stock recommendations from the UK and US markets.

And right now, Mark thinks there are 6 standout stocks that investors should consider buying. Want to see if SpaceX made the list?