SpaceX Signs $150M/Month Compute Deal with Reflection AI — What It Signals for the AI Infrastructure Race

The AI infrastructure arms race just got a significant new entry. SpaceX has inked a major compute agreement with Reflection AI, an open-source artificial intelligence lab, that will see Reflection pay $150 million per month for access to cutting-edge hardware — starting July 1, 2026, and running through 2029. The deal grants Reflection AI immediate access to Nvidia's latest GB300 AI chips and supporting hardware hosted inside SpaceX's Colossus 2 data center near Memphis, Tennessee.

The story was first reported by Kirsten Korosec at TechCrunch AI.

This is not a small partnership. At $150 million per month, Reflection AI is committing roughly $1.8 billion annually to raw compute power alone. That figure puts the scale of modern AI development into sharp relief — and sends a clear message about where the industry is heading.

Why This Deal Matters Beyond the Headlines

On the surface, this looks like a straightforward infrastructure lease. But there are several layers worth unpacking for anyone watching the AI space — whether you run a Fortune 500 operation or a 20-person startup.

Open Source AI Is Becoming an Enterprise-Scale Game

Reflection AI is described as an open-source lab, which makes the sheer size of this compute commitment notable. Open-source AI has historically been framed as the scrappy, resource-efficient alternative to closed commercial models. That narrative is shifting. When an open-source organization is committing $150 million per month to hardware, it signals that open-source AI is no longer a low-budget side project — it is competing at the highest tier of the industry.

For businesses evaluating whether to build on open-source AI models versus closed proprietary systems, this is relevant. Well-funded open-source labs with access to frontier hardware can close the performance gap faster than many expected.

SpaceX Is Quietly Becoming a Major AI Infrastructure Player

Most people still think of SpaceX as a rocket company. But Colossus — the data center network that includes the Memphis facility referenced in this deal — represents a serious infrastructure bet. SpaceX is positioning itself as a compute provider to leading AI organizations, sitting alongside the hyperscalers like AWS, Google Cloud, and Microsoft Azure in that conversation.

The fact that a prominent AI lab chose SpaceX over a traditional cloud provider is worth noting. It may reflect competitive pricing, hardware availability, or strategic alignment. Either way, it introduces more competition into the compute market, which is ultimately good news for businesses that rely on AI services.

The GB300 Advantage

The Nvidia GB300 chips at the center of this deal represent the current frontier in AI acceleration hardware. Access to these chips is not just a performance upgrade — it is a prerequisite for training and running the most capable AI models at scale. Securing dedicated access to GB300 hardware through 2029 effectively locks in a competitive advantage for Reflection AI during what may be one of the most consequential periods in AI model development.

What This Means for SMBs and Business Teams

For smaller organizations, deals like this can feel distant. But the downstream effects are real and worth tracking.

When well-resourced labs — open-source or otherwise — secure access to frontier compute, they accelerate model development timelines. That means more capable AI tools reach the market sooner, at lower cost, and with more options for businesses to choose from. The AI tools for business landscape is directly shaped by these kinds of infrastructure investments.

There is also a broader signal here about the pace of change. If open-source AI labs are now operating at this scale, business teams that delay integrating AI into their workflows risk falling further behind — not just behind competitors using closed commercial tools, but behind organizations leveraging increasingly powerful open-source models that will benefit from investments like this one.

Understanding the infrastructure layer also matters when evaluating AI automation for your team. The reliability, latency, and capability of AI tools you use depend heavily on where and how those models are trained and served. Deals like this one shape that foundation.

Platforms like WRRK.ai are designed to help business teams navigate exactly this kind of fast-moving landscape — cutting through the noise to focus on AI tools and workflows that deliver practical value right now, without requiring a team of engineers to keep up.

Original reporting by Kirsten Korosec, TechCrunch AI. Published June 22, 2026. Read the original story at TechCrunch.

Frequently Asked Questions

What is the SpaceX and Reflection AI compute deal?

SpaceX has signed an agreement with Reflection AI, an open-source AI lab, in which Reflection will pay $150 million per month starting July 1, 2026 through 2029. The deal gives Reflection AI dedicated access to Nvidia GB300 AI chips and related hardware at SpaceX's Colossus 2 data center near Memphis, Tennessee.

What is Colossus 2 and why is SpaceX operating a data center?

Colossus 2 is SpaceX's data center facility near Memphis, Tennessee, built to host large-scale AI compute infrastructure. SpaceX has been expanding into AI infrastructure, positioning Colossus as a compute resource for leading AI organizations — putting it in competition with established cloud providers like AWS and Google Cloud.

How does large-scale AI infrastructure investment affect small businesses?

When major AI labs secure access to frontier hardware, they can develop more powerful models faster. Those models eventually power the AI tools that small and mid-sized businesses use every day. More competition in the compute market also tends to drive down costs and expand access, meaning SMBs benefit indirectly from deals like this one over time.

Stay ahead of the AI tools shaping modern business — explore what's new at WRRK.ai.

SpaceX Signs $150M/Month Compute Deal with Reflection AI — What It Signals for the AI Infrastructure Race