info@bazaartoday.com a property of Inrik
A few years ago, the idea that a rocket company would become one of the world's largest suppliers of artificial intelligence computing power would have sounded absurd.
Today, it is reality.
According to Reuters, Google has agreed to secure access to approximately 110,000 NVIDIA GPUs, along with CPUs, memory, networking equipment, and supporting infrastructure from SpaceX data centers between October 2026 and June 2029. At the same time, Anthropic has reportedly committed to paying $1.25 billion per month through May 2029 for compute capacity hosted by SpaceX and xAI facilities.
These agreements reveal a profound shift occurring in the global technology landscape: AI compute has become one of the most valuable strategic resources in the world.
At the center of this transformation stands NVIDIA.
Just as oil powered the industrial economy of the twentieth century, NVIDIA's GPUs are powering the AI economy of the twenty-first.
Modern artificial intelligence models require enormous computational power to train and operate. Every breakthrough—from large language models to autonomous systems, scientific discovery, robotics, and advanced reasoning—depends on massive clusters of GPUs working together.
The world's most valuable AI companies are not competing merely for talent or algorithms. They are competing for access to NVIDIA-powered compute infrastructure.
The fact that Google, one of the largest technology companies on Earth, is securing long-term GPU capacity years in advance demonstrates how critical these resources have become.
Perhaps the most surprising development in the AI race is not that Google needs GPUs.
It is that Google appears willing to rent massive GPU capacity from SpaceX.
How did a rocket company end up with so much AI computing power?
There are several possible reasons.
First, Elon Musk understood the importance of AI infrastructure earlier than many competitors. While much of the technology industry focused on building AI applications, xAI and SpaceX focused on acquiring the physical assets needed to power AI: GPUs, data centers, networking, electricity, and cooling systems.
Second, SpaceX may have been willing to commit billions of dollars upfront to secure NVIDIA hardware during a period of extreme scarcity. GPU allocation has become one of the most competitive markets in the world. Companies that placed large orders early gained a significant advantage.
Third, SpaceX brings assets that traditional cloud providers cannot easily replicate. Through Starlink, global communications infrastructure, access to capital markets, engineering talent, and large-scale construction capabilities, SpaceX can build AI facilities at extraordinary speed.
The result is remarkable. Instead of merely being a customer of AI infrastructure, SpaceX has become a supplier of AI infrastructure.
In some cases, even companies such as Google and Anthropic appear willing to purchase or reserve capacity from SpaceX-operated facilities.
This reflects a broader shift in the AI economy. The scarce resource is no longer the AI model itself.
The scarce resource is compute.
And whoever controls the compute increasingly controls the future of artificial intelligence.
Some investors have even begun asking whether NVIDIA and SpaceX represent the two most important infrastructure companies of the AI era. NVIDIA builds the engines. SpaceX is emerging as one of the largest operators of those engines.
Together they form a new layer of the digital economy—one built not on software alone, but on physical infrastructure, energy, and computational power.
At first glance, SpaceX appears to be a rocket company. But increasingly, it is becoming an infrastructure company.
The company's Starlink satellite network, global communications capabilities, energy infrastructure, and partnership with xAI position it uniquely in the emerging AI ecosystem.
The answer is simple: AI requires more than software.
It requires:
SpaceX possesses many of these advantages.
As AI models grow larger and more sophisticated, compute capacity itself becomes a strategic asset. Companies that can build and operate AI infrastructure may become as important as the companies creating AI models.
In this new environment, SpaceX is not merely launching rockets. It is helping build the digital infrastructure that powers the next generation of artificial intelligence.
The AI race is increasingly becoming an infrastructure race.
The winners may not necessarily be the companies with the smartest algorithms. Instead, they may be the companies capable of securing sufficient compute resources to train and deploy those algorithms at scale.
Today, the largest AI organizations are locking in capacity years ahead of time because they understand a simple reality:
Without compute, there is no AI.
This explains why billion-dollar contracts for GPUs, data centers, and electricity are becoming commonplace.
The market is recognizing that AI infrastructure is not a support function—it is the foundation upon which the entire industry rests.
The story extends beyond NVIDIA and GPUs.
The real bottleneck may soon become electricity.
A modern AI data center can consume as much power as a small city. As demand grows, utilities must invest billions of dollars in transmission lines, substations, transformers, and generation capacity.
This raises an important public policy question:
Who should pay for the infrastructure required to support the AI boom?
If technology companies generate trillions of dollars in value from AI, many argue they should bear the full costs of the power infrastructure their facilities require rather than shifting those costs onto residential ratepayers.
Communities across America are beginning to confront this challenge as data centers expand into new regions.
The future of AI appears to be developing along two parallel paths.
The first is large-scale centralized compute.
This includes frontier AI models, enterprise platforms, autonomous systems, and global inference networks. These applications require enormous GPU clusters and multi-billion-dollar investments.
The second is local AI compute.
Products such as NVIDIA's RTX Spark and Microsoft's Surface RTX Spark Dev Box suggest a future where increasingly capable AI systems run directly on personal devices. These systems offer lower latency, improved privacy, lower operating costs, and better access to local information.
Both models will likely coexist.
Large cloud-based systems will provide immense intelligence and scale, while local AI systems will provide responsiveness and personalization.
The emerging AI economy is revealing a new hierarchy of strategic assets.
At the top sits NVIDIA, whose GPUs have become the engines of artificial intelligence.
Around NVIDIA is forming an ecosystem of infrastructure providers, including companies such as SpaceX, which are building the power, networking, cooling, and data center capacity needed to support AI's explosive growth.
The most important lesson may be this:
The future of AI will not be determined solely by software breakthroughs.
It will be determined by who controls the compute.
And increasingly, compute means NVIDIA.
By Hamid Porasl
@Bazaartoday
June 14, 2026