SpaceX Wins Over Two Skeptics as Analysts Look Past the Merger Noise

Two brokerage houses changed their minds about SpaceX this week, and neither did so because of the Tesla question. Bank of America lifted its price target to $275 from $235 while keeping a Buy rating, and PhillipCapital went further still — flipping from Sell to Buy and hoisting its target to $250 from $75. The common thread running through both calls is the quarterly report, not the corporate manoeuvring in the background.

The numbers behind that shift are hard to argue with. SpaceX booked revenue of $14.58 billion in the third quarter of 2026, up 161% year on year, a figure Reuters reported on Tuesday as beating expectations. Net income came in at $4.4 billion, a sharp reversal from a loss of $1.1 billion in the same period a year earlier. Analyst Glenn Thum, who authored the PhillipCapital upgrade, pointed to fresh cloud contracts signed at higher prices, the revenue contribution from Cursor, and expanding business with enterprise and government clients in the connectivity division.

The bill behind the boom

Impressive as the top line looks, the cash going out the door tells its own story. SpaceX spent $19.6 billion on capital investment during the quarter, and management is guiding toward $35–40 billion in the fourth quarter of 2026 alone. Profit and capex are different animals, so the gap between them is not evidence of a funding shortfall — but it does mean the headline earnings figure only paints part of the picture.

The reason for the spending spree sits at the heart of the AI buildout. Leased data-centre capacity underpins the company's artificial-intelligence business, which now sits alongside rocketry and satellite communications. Scaling that up demands power, cooling and a steady stream of new infrastructure. What began as a technological fantasy has turned into a decidedly physical construction project.

PhillipCapital flagged the practical bottleneck plainly: shortages of electricity and cooling capacity could slow growth. A signed cloud contract does not a working site make. UBS reckons SpaceX is targeting 13 GW of AI computing capacity by the end of 2027, with the investment required to get there estimated at roughly $350 billion — an estimate, not an invoice.

Guidance sets the next hurdle

Chief financial officer Bret Johnsen offered a revenue range of $33 billion to $37 billion for the first quarter of 2027. That is a forecast, not a result, and it raises the bar considerably. Strong sales growth and a return to profit provide a starting point; they do not prove the next order of magnitude is within reach. Investors would do well to keep reported performance and forward promises in separate mental compartments.

On the merger front, Elon Musk called a combination with Tesla a "good idea" on Tuesday, while warning against treating it as a done deal. Special committees with independent directors are reviewing a potential tie-up, according to the Wall Street Journal, and Reuters reported that both companies have formed panels to examine the idea. No terms have been agreed, and whether a deal happens at all remains uncertain.

Analysts see a larger structure potentially easing the financing of AI investments and debt raising — relevant given the capital hunger described above. Set against that, Reuters notes possible security reviews tied to Tesla's China business and SpaceX's US defence work. Bulk may make funding easier without making execution any simpler.

What actually matters

The temptation is to let the merger chatter drown out the operating story, or to let growth euphoria drown out the capex bill. Neither extreme holds up. The upgrades from Bank of America and PhillipCapital are re-ratings, not new revenue, and confidence expressed by analysts still has to be validated by the company itself.

The real test for SpaceX is whether announced computing capacity turns into productive infrastructure and whether strong sales convert into durable earnings. Tuesday's results and the guidance that followed give grounds for optimism. Delivering on them is a task the buildout itself must complete.