Wall Street is fairly bullish on SpaceX (SPCX) stock, and Oppenheimer’s latest price target raise is pushing that narrative further.
Oppenheimer analyst Timothy Horan increased his price target on SpaceX stock to $280 from $250, according to TipRanks. His price target implies about 85% upside, as SpaceX is trading near $151. The raise is a result of Horan’s growing confidence that SpaceX can become a real AI giant.
On the surface, MarketBeat’s data shows 27 of 42 analysts covering the stock rate SpaceX a buy, and this fresh take from Oppenheimer makes the stock look very attractive.
But does the thesis make sense?
Oppenheimer values SpaceX as a “unique vertically integrated AI platform”
The core valuation question is whether you believe that it can capture the market it said it aspires to capture in its S-1. SpaceX’s S-1 states that the company estimates its total addressable market (TAM) at $28.5 trillion, of which $26.5 trillion, or 92.98%, is expected to come from AI.
This means the long-term investment case depends almost exclusively on the success of SpaceX’s AI platform.
Horan believes that SpaceX’s AI play is progressing well and has increased his long-term revenue estimates for the company by about 10%.
He said that the acquisition of Cursor is “transformative” for the company and will help improve its AI platform, including Grok.
The key obstacle for growth is infrastructure capacity, and he believes it will require much higher capital expenditures (capex).
Investors who reviewed SpaceX’s second-quarter (Q2) earnings report likely raised their eyebrows reading that capex claim.
It is really simple: AI revenue in Q2 was $2.56 billion, but AI capex was $15.83 billion, and net loss was $541 million.
We can estimate that, despite the impressive revenue growth from AI, if the capex has to be much higher, as an Oppenheimer analyst said, the company will continue to report net loss instead of income for the foreseeable future.
Related: BofA names 2 SpaceX alternatives with massive upside
With an average price target of $221.2 for SpaceX, Oppenheimer is above consensus. However, this price target is nowhere near the most bullish one, Raymond James’s, at a whopping $800. Raymond James was one of the underwriters, and it certainly has an incentive to see the stock soar.
I’ve explored the problem SpaceX underwriters face, Bank of America more specifically, in my article Bank of America sets alarming SpaceX stock price target.
In short, for the stock to hit any of the price targets, almost everything the company is doing would need to go smoothly. The company needs to achieve many very difficult engineering feats, which are not guaranteed.
This is also what Morningstar equity analyst Nicolas Owens thinks. He is very bearish on SpaceX, and he values the stock at $63 per share.
The AI play is not going off without a hitch, despite Oppenheimer’s growing confidence that it is.
SpaceX’s Cursor and infrastructure build-out face setbacks
On Aug. 28, OpenAI shared that it plans to wind down its contract providing OpenAI models to Cursor. The proposed end date was set to Nov. 12, 2026.
OpenAI’s reaction to Cursor being acquired by SpaceX was fairly easy to anticipate. What is a bit trickier is guessing whether Anthropic will do the same.
Anthropic following suit would be a major blow to SpaceX. The main problem for Anthropic is that it rents capacity from SpaceX, so cutting Cursor off may cause problems on that capacity front.
The company probably has the same limited period to respond to Cursor’s acquisition, so if it also wants to cut off Cursor’s access, it will have to announce it soon.
Nonetheless, Cursor’s loss of access to one of the two generally accepted best coding models is a setback.
SpaceX also hit a bit of a snag in AI data center build-out.
SpaceX has replaced several leaders of its data-center team with executives from its rocket and Starlink operations, according to The Information. This switch follows civil engineering problems and reliability issues at data center sites in Tennessee and Mississippi, as reported on Stocktwits.
The data center sites having issues should not be a surprise to anyone who has followed the work of what used to be xAI.
Hyping SpaceX’s superfast build-out as something good is a very superficial way of looking at it. Building things fast always comes with a cost you pay for later. Building more slowly but in a more reliable and efficient manner seems to be what competitors are doing.
xAI built its Colossus 1 by mixing H100, H200, and GB200 Nvidia GPUs, and this mix is a suboptimal choice for training.
As explained by Tom’s Hardware: “When the faster GB200 chips complete their work first, the entire cluster waits for the slower H100s to catch up — a well-known bottleneck known as the straggler effect. At 220,000 chips, this effect is exponential.”
This led to Colossus 1 being used only for inference, resulting in excess capacity that was rented to Anthropic.
Meanwhile, there hasn’t been any revelation that any hyperscaler has built a supercomputer unsuitable for training.
When analysts give high price targets for SpaceX, they are saying it will win the AI race. These elevated valuations assume SpaceX will capture market share from competitors, including Google, Microsoft, Meta, OpenAI, and Anthropic.
Usually, the same firms are also bullish on Microsoft, Google, and Meta. Investors need to evaluate whether these price targets reflect realistic market growth or double-counted market share across competing sell-side models.
Related: Morgan Stanley finds bigger story in SpaceX’s $100 billion bet
