Analyst Doug Anmuth reiterated JPMorgan’s Overweight rating and a $240 price target for SpaceX (SPCX) last week. This wasn’t a new upgrade, but JPMorgan showed that it remains bullish on SpaceX and, by extension, its AI operations, which account for most of SpaceX’s valuation.
The new development was the firm’s growing confidence in Grok after SpaceX completed its acquisition of Cursor. Cursor reached $4 billion in annual recurring revenue in the most recent quarter, and it gives SpaceX solid enterprise revenue alongside the short-to-medium-term revenue coming from Alphabet (GOOG) (GOOGL) and Anthropic.
More News from Barchart
However, Cursor’s most valuable contribution may not be its existing revenue. SpaceX gets that enterprise revenue without years of relationship-building, and it is also getting valuable coding data that is already helping it make progress on its commercial AI model.
That’s more or less the gist of why JPMorgan is bullish here, though it’s worth looking into deeper since Grok is far from perfect.
Surprisingly Good Results From Grok 4.6… But There’s a Catch
While 4.6 is good at coding, most developers use Claude because it dominates the whole ecosystem with Claude Code. Even OpenAI Codex is far more popular than Grok. The issue with xAI is that its coding stack may not keep up, so lots of developers end up avoiding it altogether. Google faces the same issue now with its AI products.
No major company would want to rely entirely on Grok or an xAI ecosystem until SpaceX proves its AI division can keep up, rather than releasing one good AI model and then falling behind for months.
There’s a chance that even with Cursor data, Grok 4.6 might mirror post-2025 Google AI models, where a few benchmark-leading models couldn’t keep up, and subsequent models weren’t capable either. SpaceX needs to rule out that possibility for you to trust this as a solid revenue stream down the line.
I Wouldn’t Focus on the AI Models to Gauge Performance
Consumer-facing LLMs are still financially unviable, whereas the hardware and cloud infrastructure providers that run them remain deeply profitable. The lack of people using Grok ended up being profitable after Google and Anthropic bought that excess capacity, without which SPCX stock likely would’ve been lower, since the contracts exceed $70 billion in aggregate value if both survive to maturity.
