Another infrastructure play going public? Or just another narrative? Switch, a data center operator born in the deserts of Nevada, has filed confidentially for an IPO with a target valuation of $80 billion. That's not a typo. At that valuation, Switch would sit on par with Equinix, a company that pulls in nearly $90 billion in revenue. Switch? Its revenue likely hovers around $10 billion. The market is telling us something: it is not buying data centers; it is buying a story. And that story is about the transformation of 'infrastructure as real estate' into 'infrastructure as platform.'
But as a narrative hunter, I've seen this play before.
Back in 2017, I reverse-engineered the Zeppelin Security Library, not because I was assigned to, but because I sensed the code was whispering a deeper truth about how Ethereum would scale. That same instinct now hums when I look at Switch's confidential filing. The question is not whether the story is true, but whether the market will believe it long enough to price the IPO. And then, whether the underlying technology—the land, the power, the cooling—can deliver on the promise.
Context: The Infrastructure Canvas
Switch is a data center operator, but not a typical one. It has long marketed itself as 'AI-optimized'—high-density racks, liquid cooling, custom thermal management pools. For years, these features were overkill for traditional enterprise workloads. Then the AI boom hit. Suddenly, a single GPU cluster can draw 100 kW per rack, and Switch's excess design became a moat.
The company filed confidentially for an IPO in early 2025, targeting an $80 billion valuation. The timing is strategic: the AI infrastructure market is red-hot, with CoreWeave IPOing at a $350-500 billion cap, and private capital flooding into data center startups. Switch is not a startup; it has been operating for over a decade, with campuses in Nevada, Michigan, and Texas. It owns its own network (Switch CIN), offers colocation, wholesale, and build-to-suit services. But its core asset is not the buildings—it's the power contracts and land permits.
In the AI era, power is the new oil. And Switch has barrels of it.
Core: The Narrative Mechanism Behind $80 Billion
Let's break down the valuation. The article—a detailed analysis from a Chinese research firm—does not provide Switch's financials. But we can infer. Based on industry data, Switch likely generated $10-12 billion in revenue in 2024 (up from $6.8 billion in 2022). At $80 billion, that implies an EV/Revenue multiple of 70-80x. Compare that to Equinix (8-9x), Digital Realty (6-7x), or even CoreWeave (22-30x). The multiple is 2-3x higher than the most comparable AI infrastructure company.
That is not a traditional valuation. It is a narrative valuation.
The narrative is that Switch is not a real estate investment trust (REIT) but a 'AI infrastructure platform.' The market is buying into a paradigm shift: data centers are no longer just about power and cooling; they are the new compute platforms. In this view, Switch is the landlord of the digital economy, but with a twist—it owns the scarce resources (land, power, permits) that are becoming bottlenecks for AI expansion. The narrative is that these resources will appreciate faster than the buildings themselves.
But here's where the narrative gets tricky.
Code speaks, but culture listens.
The culture of AI infrastructure is one of scarcity and urgency. Every hyperscaler—Microsoft, Amazon, Google—is scrambling to secure power capacity. The grid is overloaded; transformer lead times are 2-3 years. Switch's locked-in power contracts are its true moat. But the market is pricing this scarcity as if it will last forever. It won't. New power generation (gas, nuclear, renewables) will come online, and the supply-demand balance will shift.
I've seen this before. In 2020, during DeFi Summer, I identified the 'yield trap' in early Compound forks. The narrative was that yield farming was a new asset class. The reality was that the yields were unsustainable, propped up by token inflation. The Switch valuation has a similar feel: the AI infrastructure narrative is compelling, but the underlying economics may not support the multiple.
The sentiment analysis:
Market sentiment is bullish on AI infrastructure. Investors are desperate for exposure to the AI supply chain, and data centers are a tangible asset. But sentiment is fragile. The Switch IPO will be a litmus test. If it prices at $80 billion, it will validate the narrative and trigger a wave of data center IPOs. If it prices lower or delays, it could trigger a correction across the entire AI infrastructure sector.
Contrarian: The Narrative Mirages
Here is the counter-intuitive truth: Switch's most valuable asset is not its technology but its land and power permits. And those are finite. The $80 billion valuation implies that the market is pricing in a future where power permits become exponentially more valuable. But that is a very specific risk.
Consider: if Switch's revenue is only $7 billion (the conservative scenario), the EV/Revenue multiple jumps to 114x. That is not a growth stock; that is a bubble. And bubbles are driven by narrative, not fundamentals.
Another rug pull? Or just another myth?
The myth is that 'AI infrastructure' is a new asset class. It is not. It is still real estate, with the same risks: tenant concentration, power price volatility, regulatory changes, and overbuilding. The difference is that the tenants are now AI companies, which are themselves burning cash. If the AI funding cycle slows, those tenants will cut costs, and Switch's occupancy rates will drop.
The Cassandra complex is real.
I have been here before. In 2021, I co-founded a newsletter called 'The Digital Totem' that analyzed NFTs not as art but as anthropology. I saw the tribal identity behind Bored Apes and predicted the crash before it happened. Now, I see the same tribal identity forming around 'AI infrastructure' as a status symbol for institutional investors. The same dynamics will play out: a narrative-driven rally, a peak, and then a correction.
Takeaway: The signal in the noise
The Switch IPO is a signal. It tells us that the market is pricing in a paradigm shift, but it also tells us that the shift is still a narrative. The real question is: will the underlying technology—the power, the cooling, the land—deliver on the promise? Or will the narrative collapse under the weight of its own multiples?
As a narrative hunter, I watch the data. The S-1 filing will reveal the truth. Until then, the $80 billion narrative is a beautiful story, but it is still a story. And stories, no matter how compelling, eventually need to be backed by earnings.
Data centers aren't real estate; they're anthropology.
And the anthropology of this market is one of collective belief. The Switch IPO will test that belief. If it succeeds, it will validate the 'infrastructure as platform' narrative. If it fails, it will be a cautionary tale about the dangers of narrative-driven valuation.
Either way, the story is just beginning.
