Hook
Last week, ARK Invest dropped a bombshell: SpaceX told its investors that 90% of its future growth will come from AI. Not rockets. Not Starlink. AI.
A rocket company, built on the back of explosive reusability, is now packaging itself as an infrastructure play for the world’s most compute-hungry market. The logic is seductive: low launch costs → cheap orbital data centers → a new class of AI compute. And the crypto world should be paying close attention—not because it validates the technology, but because it exposes a centralization risk we’ve been warned about for years.
We didn’t see this coming. Or maybe we did. The same forces that consolidate power in traditional tech are now reaching for space.
Context
ARK’s thesis is straightforward. SpaceX controls the entire stack: the rockets, the satellites, the ground stations, and now—through its acquisition of xAI—the AI models themselves. The claim is that a fully orbital data center can be built at 25% lower cost than a terrestrial equivalent, with near-zero energy costs thanks to solar. The key assumption? Launch costs drop to $100 per kilogram, down from today’s $1,500–$2,500 per kg on Falcon 9. The numbers are breathtaking. But they rely on a series of unverified bets: Starship’s success, orbital maintenance feasibility, and a customer base willing to trust their most sensitive AI workloads to space.
At first glance, this sounds like a utopian vision—compute for everyone, cheap and green. But as someone who has spent a decade in crypto, building communities around decentralization, I see the shadow. This is not DePIN. This is not permissionless infrastructure. This is a single company—led by a single individual—building a monopoly on the physical layer of the AI economy.
Core Insight
Let’s examine the numbers through the lens of what they mean for crypto’s decentralized compute networks.
ARK estimates the global data center market at $500 billion today. If SpaceX captures even 5% of that, it’s a $25 billion revenue opportunity—on top of its existing launch and Starlink businesses. But here’s the rub: the path to that capture requires massive capital expenditure, and the only way to raise that capital is through a narrative that justifies extreme valuations. The “90% from AI” line is a signal to public markets: “We are not a cyclical launch provider. We are an AI infrastructure hyper scaler.”
But the decentralization community must ask: who owns the keys to this compute? If SpaceX becomes the dominant orbital data center operator, every AI model trained on its infrastructure will be subject to its terms. No censorship resistance. No trustless verification. Just a single point of failure wrapped in a shiny rocket.
I recall a conversation from 2020, during the DeFi Summer, when I wrote that “Trust is no longer a promise; it’s a protocol.” That principle holds true here. SpaceX may promise low-cost compute, but the protocol—the underlying governance and access—remains opaque. Meanwhile, projects like Akash Network, Filecoin, and Helium are building decentralized alternatives that guarantee censorship resistance and verifiable computation. The gap is real.
Moreover, the narrative that “orbital data centers are cheaper” ignores the hidden costs: hardened electronics, radiation shielding, satellite replacement cycles, and latency. For AI inference, where low latency is critical, a geostationary orbit adds 250ms round-trip delay. That’s unacceptable for real-time applications. So SpaceX’s compute is likely only viable for batch training or non-latency sensitive workloads—a niche, not the whole pie.
Contrarian Angle
Now, let me play the contrarian—not against crypto’s ideals, but against the panic. Perhaps SpaceX’s move is exactly the kickstart the industry needs.
If SpaceX succeeds in driving launch costs to $100/kg, it will commoditize access to space. That could spur a new wave of decentralized satellite networks—small, modular, and open—that build on top of this cheap transport layer. The same way AWS commoditized servers and enabled a thousand startups, SpaceX could commoditize orbital compute and enable a thousand DePIN projects.
The key is to ensure that the infrastructure layer itself is open. Just as Bitcoin and Ethereum built on open internet protocols, the next generation of decentralized compute must run on permissionless orbital nodes. That’s why I’m watching projects like SpaceChain and the emerging DePIN ecosystem. They are not trying to build rockets; they are building the middleware that makes trustless orbital compute possible.
But here’s the catch: trustless systems require trusting relationships. SpaceX may claim openness, but its history shows a tendency toward vertical integration and control. The company can change its API, pricing, or access terms at any time. In a bear market, where survival matters more than gains, we must ask which protocols are bleeding—and which are resilient.
Takeaway
The AI narrative around SpaceX is a mirror held up to crypto. It shows us the future we want: ubiquitous, cheap, decentralized compute. But it also warns us about the future we might get: centralized control dressed in a futuristic suit.
Code is law, but empathy is the interface. The crypto community must build bridges to these emerging physical infrastructure layers, not as customers, but as co-creators. Otherwise, we’ll find ourselves renting compute from a company that can turn off the tap with a single tweet.
The pivot wasn’t about 90% growth. It was about 100% control. Let’s make sure we build the alternative before it’s too late.