Google has closed a huge agreement with SpaceX to reinforce its computing capacity in the midst of the AI ​​race, and the move fits perfectly with the pressure that the entire industry is experiencing today. The pact contemplates payments of $920 million dollars per month between October 2026 and June 2029, a figure that brings the contract to nearly $30 billion dollars.

What did Google and SpaceX agree on?

SpaceX announced the deal in a regulatory filing and detailed that Google will gain access to approximately 110,000 GPUs, CPUs, memory and other related components from NVIDIA. According to TechCrunch, the deal is similar in length and scale to the one SpaceX signed with Anthropic in late May, although in this case Google would pay for about half of the compute Anthropic has on Colossus 1.

The detail that stands out the most is that Google is not buying satellites, nor space bandwidth, nor a stake in SpaceX. What you get is pure and simple computing power, something that today is worth gold for training, scaling and sustaining artificial intelligence products.

The objective of the pact is quite clear. Google needs more infrastructure to respond to the growth of its AI products, especially Gemini Enterprise and its agent platform, which according to the company itself have been in higher demand than expected. The company described the contract as a temporary solution to ensure “bridge capacity” while it adjusts its own infrastructure.

In simple terms, Google buys time, technical muscle, and room to maneuver. When a company of this size is forced to go to market seeking billions in capacity, the message is powerful: demand for AI continues to push infrastructure to the limits.

What does Google gain from this agreement?

Google gets access to a massive block of high-performance hardware without having to wait for all of its internal expansion to be ready. This allows it to maintain the pace of its AI services, absorb usage peaks and avoid bottlenecks at a stage in which each product improvement depends, to a large extent, on how many chips and how much memory you have available.

There is also an important strategic nuance. Alphabet has already committed more than $180 billion in capital expenditures this year and anticipates that figure will grow even more in 2027. In that context, such a contract is not only a technical purchase, it is also a way to ensure operational continuity in an industry where falling short of computing can slow launches, responses and commercial deployments.

Infrastructure for AI is becoming more expensive

This agreement shows something that was already being felt in the sector, but now it is clearer than ever. Infrastructure for AI is becoming the new battlefield, and big technology companies are willing to pay huge amounts to avoid losing speed. Google also maintains a long-term relationship with SpaceX and is also an investor in Elon Musk's company, whose stake could exceed $100 billion after the planned IPO.

There is another piece of information that helps read the panorama with more context. SpaceX signed this deal shortly before its expected IPO on Nasdaq, as it seeks to raise its valuation to around $1.75 trillion. That makes the contract a relevant financial piece for both parties, not just a technological play.

At its core, the deal says a lot about how power is shifting in the age of AI. The advantage no longer depends only on brilliant software, but on who manages to secure enough chips, data centers and power to make it all work at scale. Google, in this case, pays a fortune not to be left behind in that race, and SpaceX monetizes an infrastructure that has become strategic even outside of its space business.

If this pact ends up materializing as announced, we will be facing one of the largest technology contracts in recent years, and also a fairly clear sign of where the digital economy is moving. Computing is the new oil, and companies like Google are acting accordingly.