SpaceX Just Joined the Hyperscaler Club. AI Capex Spending Could Hit $1.15 Trillion

Wall Street has officially decided that SpaceX (SPCX) is more than a rocket company with a satellite internet business attached. Morgan Stanley recently added SpaceX to its hyperscaler basket for the first time, putting its projected $159 billion capex spending alongside Amazon (AMZN), Microsoft (MSFT), Alphabet (GOOGL), Meta Platforms (META), and Oracle (ORCL).

The significance goes far beyond a change in an analyst spreadsheet. It is another sign that the artificial intelligence (AI) infrastructure boom is becoming much larger — and much harder to contain.

Morgan Stanley's latest estimates suggest that hyperscaler capital spending could reach an astonishing $1.15 trillion in 2027, with SpaceX alone accounting for roughly $159 billion. That number should make investors rethink just how long this AI infrastructure cycle can run.

The $1 Trillion AI Spending Threshold Is About to Disappear

Wall Street has repeatedly underestimated how much money the technology industry will pour into AI.

For 2026, analysts previously forecast roughly $527 billion of hyperscaler capital spending. For 2027, the estimate has surged toward $1 trillion — and that's even before SpaceX was added to the group. Now, SpaceX is pushing the projected total to about $1.15 trillion.

The important point isn't simply that the number is enormous. It's that the forecast keeps moving higher.

The traditional assumption has been that AI infrastructure spending will eventually slow as companies digest their enormous investments. Instead, the opposite keeps happening. Companies raise spending plans, new players emerge, and the definition of a hyperscaler keeps expanding.

Oracle was once the notable newcomer. Now SpaceX is joining the party — and that could be the bigger story.

SpaceX Is Becoming an AI Infrastructure Company

SpaceX isn't merely talking about AI. It's already spending at the scale of a hyperscaler. The company spent $15.8 billion on AI compute in the second quarter and has about 1.4 gigawatts of capacity live today. It's targeting more than 2 GW by year-end, while CEO Elon Musk has discussed ambitions approaching 10 GW — potentially 20 GW — by the end of 2027. That's an extraordinary buildout.

SpaceX also has something most AI infrastructure companies don't: vertical integration. Starship could eventually lower launch costs dramatically, Starlink provides global connectivity, and the company is pursuing energy and computing capabilities that could eventually extend beyond traditional terrestrial data centers.

Morgan Stanley has already described the opportunity in terms of converting energy into intelligence at massive scale. Adding SpaceX to the hyperscaler basket is essentially Wall Street putting its stamp on that thesis.

The Real Winners May Be Selling the Shovels

Investors don't necessarily need to own SpaceX to benefit from this spending explosion. If $1.15 trillion of annual hyperscaler capex becomes a reality, somebody has to build it. That means demand for advanced semiconductors and memory, networking equipment, cooling systems, power-generation infrastructure, electrical equipment, and data-center capacity could remain enormous.

This is why the continuing upward revisions matter so much. Every additional gigawatt of AI capacity represents another wave of infrastructure demand. The AI trade may be less about predicting which chatbot wins and more about identifying the companies supplying the computing arms race.

$1.15 Trillion Does Not Guarantee Great Returns

There is an important catch, however: Capex isn't the same thing as returns on capital. Building gigawatts of data-center capacity is extraordinarily expensive. Power availability, supply chains, permitting, talent, financing, and execution all represent potential bottlenecks. Competition could also drive down returns as traditional hyperscalers and new entrants spend aggressively at the same time.

SpaceX carries additional risks because its AI ambitions are still relatively young, while its valuation leaves little room for major execution mistakes. Still, investors shouldn't dismiss the significance of Morgan Stanley's move.

The market keeps expecting the AI infrastructure boom to cool. Instead, the spending estimates keep getting bigger — and another $159 billion in hyperscaler spending just appeared.

The investment thesis is becoming harder to ignore. AI isn't merely increasing demand for computing. It's creating an infrastructure supercycle that could be far larger than Wall Street originally imagined.