Google and SpaceX: $920M per month for AI computing power
Google has announced a $920 million per month deal with SpaceX to acquire computing power. The news, reported by TechCrunch On June 5, 2026, a question arises from an unexpected inquiry about the recently launched AI products from the Mountain View company. This is an extraordinary figure, surpassing the entire annual budget of many large European technology corporations.
However, the most relevant aspect isn't the number itself. It's the strategic direction this agreement signals: compute infrastructure for artificial intelligence is moving out of traditional data centers and into hybrid architectures, with satellite components and distributed geographical redundancy. Consequently, dependence on a single cloud provider becomes a structural, not just operational, risk. Therefore, Italian SMEs must also start thinking about the scalability and resilience of their digital infrastructure.
We of SHM Studio Let's observe this scenario with close consultation. Indeed, the infrastructure choices of major players are redefining the costs and possibilities of AI services accessible to medium-sized businesses. In summary, understanding what is happening between Google and SpaceX helps in making more informed decisions about one's digital roadmap.
The timeline of an unprecedented agreement
On June 5, 2026, TechCrunch has published the details of the deal between Google and SpaceX. The monthly value is $920 million. On an annual basis, this amounts to over $11 billion dedicated to acquiring computational capacity.
A Google spokesperson stated that the agreement stems from unexpected demand. Specifically, recently launched AI products have generated traffic volumes exceeding any internal forecasts. Therefore, the use of external infrastructure has become an urgent operational necessity, not a planned strategic choice.
This detail is important. It indicates that even a company with Google's resources can be caught unprepared by the speed of AI adoption. Consequently, infrastructure scalability is no longer a topic reserved for large corporations; it's an issue that concerns every organization using advanced digital services.
Why SpaceX and not another hyperscaler
SpaceX's selection as a compute partner is, at first glance, surprising. SpaceX is not a traditional cloud provider. However, its Starlink network offers global coverage with relatively low latency and rapidly expanding bandwidth capacity.
Furthermore, SpaceX has computing infrastructure related to the management of the satellite constellation. These systems require real-time distributed processing. Therefore, the computational capacity already exists and is natively geographically distributed.
On the contrary, further recourse to AWS, Azure, or Oracle Cloud would have further concentrated Google's dependence on direct or indirect competitors. For this reason, a partner like SpaceX offers both technical capability and competitive neutrality. Wired has already analyzed in the past how the convergence between space infrastructure and the terrestrial cloud is redesigning the geography of global compute.
Winners, losers, and those who watch
In this scenario, the immediate winners are clearly identifiable. SpaceX obtains a huge recurring cash flow. This finances further development of Starlink and orbital computational capacity. Similarly, Google maintains the operational continuity of its AI services without having to wait years to build new data centers.
However, there are entities that emerge weakened from this dynamic. Traditional cloud providers—particularly European and Asian ones—are seeing a difficult-to-challenge infrastructural duopoly consolidate. Furthermore, governments aiming for digital sovereignty must contend with critical infrastructures that literally orbit outside their jurisdiction.
SMEs are the ones left watching. In fact, medium-sized enterprises have no say in these agreements. Despite this, they suffer the indirect consequences through the costs of cloud services, the availability of AI APIs, and the latency of generative models. Therefore, understanding these dynamics is the first step to adapting.
SHM Studio Reading: Three Operational Implications
We of SHM Studio Let's read this agreement through a consulting lens oriented towards Italian SMEs. Three concrete implications emerge.
First implication: the AI question is structural, not cyclical. If Google is short on compute, it means AI adoption is growing at a pace that no predictive model had anticipated. Therefore, businesses that are still postponing the integration of AI tools into their processes are accumulating a real delay. Our
News Categories
Related articles
Discover other articles that explore similar topics in depth, selected to give you a more complete and stimulating view. Each piece of content is carefully chosen to enrich your experience.