The Switch IPO and the Infrastructure Race: Why Decentralized Networks Must Learn from Centralized Giants

Last week, the news hit my feed: Switch, a Las Vegas-based data center operator, is preparing an IPO at a potential $80 billion valuation. The article was thin—just the figure and a mention of internal preparations. But for anyone watching the intersection of physical infrastructure and blockchain, that number is a siren. $80 billion for a company that rents space, power, and connectivity. No software patents. No viral app. Just concrete, cooling towers, and long-term leases with hyperscalers like AWS and Azure. And yet, the market is ready to assign it a valuation that dwarfs most DeFi protocols. Why? And what does it mean for the blockchain projects that rely on the very same physical layer? I first encountered the power of data centers while running community workshops for Aave’s Latin American launch in 2020. Aave’s smart contracts didn’t care where the servers were, but the users did—latency meant slippage, downtime meant lost funds. At the time, I thought decentralized infrastructure could solve that. But after five years of watching DePIN projects struggle to match the uptime of a centralized colo, I’m less sure. Switch’s business model is deceptively simple. It builds massive data center campuses—like its flagship Citadel campus in Tahoe Reno—and leases out space, power, and network connectivity to enterprises, cloud providers, and increasingly, AI companies. The unit economics are brutal upfront: a single building can cost hundreds of millions in construction, with a 2-3 year lag before it generates revenue. But once filled, the cash flows are predictable and sticky. The average lease is 5-10 years; customers like Meta and Google don’t move their servers lightly. Switching costs are astronomical: physical rewiring, network reconfiguration, and the risk of disrupting AI training jobs that cost $10 million a week. So when I read that Switch is eyeing an $80 billion valuation, I didn’t blink. That’s not hype—that’s the present value of a monopolistic bottleneck in the digital economy. Data centers are the new oil refineries. And blockchain, despite its promise of peer-to-peer trust, runs on them. Every Ethereum node, every Solana validator, every Filecoin storage provider—they all live inside buildings like Switch’s. The difference is that blockchain projects often pretend they don’t need them. They talk about permissionless networks, but the hardware sits on rented racks in centralized facilities, subject to the same single points of failure. Let’s dissect what makes Switch worth $80 billion, because it’s a masterclass in competitive dynamics that DePIN projects should study. First, network effects. Switch doesn’t just rent space; it hosts cross-connects. When multiple cloud providers and carriers colocate in the same facility, the value to each tenant increases. A company can peer directly with AWS at sub-millisecond latency inside Switch’s building—more efficient than going through the public internet. That’s a classic two-sided market: more tenants attract more carriers, which attract more tenants. Switch’s 40+ meet-me rooms create a dense interconnection hub that competitors can’t replicate overnight. Second, scale economies. Switch’s massive campuses achieve power usage effectiveness (PUE) near 1.1, compared to the industry average of 1.5-1.6. That means for every watt of power used for computing, only 0.1 watt is wasted on cooling and overhead. Over a 50-megawatt facility, that’s millions in annual electricity savings—and electricity is a data center’s largest operating cost. Switch can pass those savings into lower rents, undercutting smaller players. The barriers to entry are not just capital; they’re operational mastery. A new entrant would need hundreds of millions and years of learning to match Switch’s efficiency. Third, switching costs. Once a customer wires their servers into Switch’s racks, sets up private fiber links, and integrates with Switch’s monitoring and security systems, moving to another provider is a multi-month project involving truck rolls, cable pulls, and network reconfiguration. For an AI company training a model that will generate billions in revenue, a month of downtime is unthinkable. The stickiness is almost absolute. Now, contrast this with blockchain infrastructure. Projects like Filecoin claim to offer a decentralized storage network, but the reality is that most storage miners run their hardware in data centers like Switch’s. The protocol’s tokenomics incentivize cheap power and high uptime, which data centers provide, but the decentralization premium? Not there. Similarly, Helium’s hotspots are often placed on rooftops in urban areas, but the backend network routing still relies on centralized cloud servers. The blockchain layer provides trust in data, but not in the physical substrate. Here’s the contrarian angle that most crypto enthusiasts don’t want to hear: The most valuable infrastructure companies in the world are centralized, and their value is growing faster than the value of decentralized counterparts. Switch’s $80 billion valuation is a bet on centralization—on the idea that a single operator can manage risk, optimize operations, and build trust better than a distributed network of independent actors. And the market is rewarding it. But I’ve seen the other side. In 2021, I worked with Art Blocks to interview 50 female digital artists who used blockchain to assert ownership over their work. They didn’t care about the data center where their NFTs were stored—they cared about the ability to prove provenance without a middleman. That’s the promise of blockchain: not to replace physical infrastructure, but to make the ownership layer permissionless. The data center is just a utility; the trust is in the code. However, the contrarian truth cuts deeper. Switch’s IPO is a warning to blockchain infrastructure projects that ignore the physics of latency, power, and interconnection. Decentralized storage networks like Arweave and Filecoin have struggled to match the throughput and reliability of a centralized cloud object store like Amazon S3. Why? Because physics doesn’t care about consensus algorithms. Data still travels at the speed of light through fiber optic cables that cross the same dark trenches as those feeding Switch’s buildings. The marginal gain from adding a node in a remote village is negligible if the core network has a choke point. I recall a conversation in 2023 with a core developer from a leading L2 rollup. He complained that sequencers were mostly running in a single data center in Northern Virginia. I asked why. “Because it’s cheaper and faster,” he said. “We can’t afford to run six distributed sequencers when one gives us 99.999% uptime.” That’s the trade-off: decentralization versus performance. Switch’s business model proves that investors pay a premium for performance and reliability, not for ideological purity. So what does this mean for the reader who holds tokens in DePIN projects or runs a blockchain node? First, stop pretending that physical infrastructure is fungible. The data center you choose matters. A node in Switch’s Tahoe Reno facility—with direct peering to major cloud providers and low latency to the West Coast backbone—will perform significantly better than one in a small colo in Nebraska. The network effects of interconnection are real. If you’re serious about running a validator, choose your home with the same rigor as choosing your protocol. Second, consider the possibility that the future of blockchain infrastructure is not fully distributed, but a hybrid model where decentralized governance sits on top of centralized physical layers. Already, projects like EigenLayer are exploring “restaking” of trust, but the hardware still lives in data centers. The key insight from Switch’s IPO is that the physical layer is a natural monopoly—data centers get better as they get bigger. Blockchain can govern the allocation of that resource, but it cannot replicate its efficiency at scale without massive capital. Third, understand that every time you read about a DePIN project raising $100 million, that money will likely end up paying rent to companies like Switch. The infrastructure stack is layered: at the bottom, concrete and copper; above, software and protocol. The concrete layer is where the real moats are built. Switch’s valuation tells us that investors understand this. The crypto community should too. Connect first, transact second. Always. I’ve been in this industry long enough to remember the Terra collapse in 2022, when a DAO I advised lost 90% of its treasury. We had to rebuild trust from scratch. One thing we learned was that transparency about physical infrastructure—where your validators run, who controls the keys, what the uptime SLA is—matters more than clever tokenomics. Users want to know their assets are safe. Switch’s success is a reminder that safety and reliability are the ultimate network effects. In a bear market, survival is about trust, not flashy tech. The takeaway is not that centralization is better, but that blockchain projects must acknowledge the physical realities they depend on. The next wave of innovation won’t come from ignoring Switch—it will come from building protocols that can harness the efficiency of centralized infrastructure while preserving decentralized control. Imagine a world where Switch opens up its interconnection fabric as a public good, governed by a DAO, with transparency into power sources and usage. That would be a trillion-dollar opportunity. But that requires a shift in mindset. The Switch IPO is a mirror held up to our own assumptions. We talk about decentralization as an end, but the market values uptime and performance. If we can bridge that gap—if we can build protocols that deliver the reliability of Switch with the permissionlessness of Bitcoin—then we’ll have truly earned the trust of the next billion users. As I write this from Buenos Aires, looking out at a city that still runs on cash, I’m reminded that infrastructure is an enabler, not a destination. The data center is a utility. The blockchain is a promise. The challenge is to make that promise inseparable from the utility. Connect first, transact second. Always. Based on my audit experience with DeFi protocols, I’ve seen teams spend six months optimizing gas efficiency but ignore that their RPC endpoint is single-homed. Switch’s $80 billion valuation is a wake-up call. Build for the physical layer, or become irrelevant.