At the heart of every infrastructure decision lies a quiet judgment about who we trust. Consider the recent SemiAnalysis report on SpaceX's ambition to add over 10GW of computing power by the end of 2027. This is not merely a technical feat—it is a statement about the concentration of computational authority. When a single entity, even one as visionary as SpaceX, controls a significant fraction of the world's compute, the decentralization ethos that underpins blockchain technology faces a profound challenge. I have spent years translating the Ethereum whitepaper into Portuguese, auditing DeFi protocols, and building open-source toolkits for human verification. These experiences have taught me that trust is not a byproduct of speed; it is a fragile architecture of verification and distribution. The news of SpaceX's computing ambition forces us to ask: are we building the infrastructure for a decentralized future, or are we simply replacing one gatekeeper with another?
The context here is crucial. SpaceX, under Elon Musk's leadership, has long been a pioneer in aerospace and satellite internet. But the pivot to massive-scale computing is a logical extension of the company's resources and vision. According to the SemiAnalysis report, Musk stated that SpaceX's conservative target is to deliver 6-8GW of incremental computing power in 2027, with upside exceeding 10GW. To put this in perspective, a single gigawatt of compute at current capital expenditure rates costs roughly $50 billion. That means SpaceX's 2027 capital expenditures could reach $300-500 billion. This is not a marginal investment; it is a bet on the centralization of AI and cloud computing infrastructure. The report further notes that when OpenAI and Anthropic provide API inference services on GB300 clusters, each GW can generate over $100 billion in revenue per year. At a rental price of $3 per GPU per hour, the annual cost per GW is about $12 billion. The math is staggering: SpaceX could become the world's largest compute provider, dwarfing even the hyperscalers. Microsoft's $250 billion infrastructure agreement with OpenAI in October 2025 corresponds to about 7GW of computing power, and it is possible for Microsoft to sign a computing power contract with SpaceX for about 3GW, with a total value of approximately $150 billion. SemiAnalysis predicts that SpaceX's annual recurring revenue could reach $300 billion by the end of 2027. This is not just a business story; it is a story about the concentration of foundational resources.
The core of my analysis lies in the tension between efficiency and decentralization. From a technical perspective, SpaceX's compute clusters could enable unprecedented AI capabilities. But as an open source evangelist, I see a red flag. Centralized compute is a single point of failure for trust. In the blockchain world, we have long argued that trust must be distributed across many nodes. When a single entity controls the computational substrate upon which AI models are trained and deployed, the risk of censorship, manipulation, and systemic abuse skyrockets. Based on my audit experience with Aave V2, where I identified logic errors in interest rate models, I learned that even the most well-intentioned systems can have hidden flaws. The same applies to compute infrastructure. If SpaceX becomes the dominant provider, the entire ecosystem of decentralized applications, from DeFi to DAOs, becomes dependent on a single corporate actor. This is not a theoretical concern. Consider the recent controversy around AI model training datasets: who ensures that the data is not biased, that the model weights are not tampered with? In a centralized compute model, the answer is the provider's internal policy. In a decentralized model, it is verifiable computation. The SemiAnalysis report highlights the revenue potential: $100 billion per GW per year. But at what cost to the open internet? The infrastructure of the future must be permissionless, not just profitable.
Now, let me offer a contrarian angle. Some might argue that SpaceX's scale is necessary for the next generation of AI, and that centralization of compute is a natural efficiency gain. They might point to the fact that Musk himself has contributed to open source initiatives, such as Tesla's patents. But I disagree. The efficiency argument is a trap. In the 2020 DeFi summer, I saw similar logic: faster, cheaper, more centralized. The result was a series of exploits that cost millions. Transparency is not the oxygen of trust; verification is. A centralized compute provider, no matter how well-intentioned, creates a power imbalance. The SemiAnalysis report suggests that SpaceX's annual recurring revenue could reach $300 billion by 2027. That kind of economic power translates into political power. Who governs the governance of the compute? Today, it is a private company. Tomorrow, it could be a state actor. The blockchain community has always advocated for "code is law," but code running on SpaceX's servers is not law—it is a service. The same applies to AI. If AI models are trained on centralized compute, the outputs are subject to the provider's whims. This is not a call to reject large-scale compute; it is a call to distribute it. We need to build protocols that allow compute to be shared across many providers, with verifiable proofs of execution. This is where zero-knowledge proofs and trusted execution environments come in. My work on the "Verifiable Humanity" initiative with EU Web3 Foundation showed that it is possible to integrate privacy-preserving verification into decentralized systems. The same principle can apply to compute.
Let me ground this in a real-world example. During the NFT cultural critique project "Soulbound Truths," I curated a digital exhibition that rejected speculative flipping in favor of community-building tokens. The project attracted 10,000 unique visitors but zero secondary market trades. This proved that value lies in identity, not liquidity. Similarly, the value of compute lies in its distributed nature, not in its raw speed. SpaceX's 10GW target is a testament to human ingenuity, but it is also a warning. The bear market of 2022 taught me that resilience is built on diversity. The Terra/Luna collapse and FTX bankruptcy were not just failures of code; they were failures of centralized control. In the aftermath, I co-authored "Code as Law, but People as Gods," a 30-page essay on building resilient systems. The essay was downloaded 25,000 times and cited by three major open-source foundations. The lesson is clear: centralization is a vulnerability, not a feature. SpaceX's compute dominance is a vulnerability for the entire ecosystem. We must ensure that the infrastructure of the future is open, verifiable, and distributed. The SemiAnalysis report provides a business case, but we need a moral case.
Finally, what is the takeaway? The news of SpaceX's compute plans is not just a business development; it is a call to action for the blockchain and open source community. We must invest in decentralized compute solutions, or risk ceding the foundation of the digital world to a single entity. The technologies exist: peer-to-peer GPU networks, verifiable computation, and decentralized storage. But they need funding and adoption. The EU Web3 Foundation's grant for zero-knowledge proofs was a start, but we need more. As an evangelist, I believe that the ultimate purpose of blockchain is to preserve human agency in an age of algorithmic automation. That means building infrastructure that is not just fast, but also fair. Code is law, but ethics is soul. The soul of the internet is at stake. Will we allow it to be hosted on a single server?
In the end, the question is not whether SpaceX can build 10GW of compute—they probably can. The question is whether we will build the alternatives. The SemiAnalysis report is a signal, not a verdict. Let us use it as a catalyst for a new wave of decentralized infrastructure. The future of trust depends on it.
