Eight zeroes.
$80,000,000,000.
That’s the valuation whispered for Switch’s upcoming IPO. A data center operator, not a software company. Not a chip designer. A landlord for servers.
And yet, the market is treating it like the next Nvidia.
I’ve seen this before. In 2021, when DeFi TVL hit $100B, everyone thought passive yield was free money. I lost 40% of capital to impermanent loss.
Numbers don’t lie. Infrastructure does.
Switch’s story is the same. The AI narrative is real—training models require massive compute, and compute demands power and cooling. But the gap between hype and infrastructure reality is filled with risk.
Let’s cut the noise.
The Hook: Price Action Anomaly
Switch is not publicly traded yet. So why am I writing about it?
Because the private market is already pricing in an AI-driven demand shock. The $80B valuation implies a future earnings multiple that rivals Equinix (EQIX) and Digital Realty (DLR). Those two are publicly traded, with decades of operational data. Switch is a private company going public at a premium.
That’s a signal.
When the market assigns a premium to an illiquid, pre-IPO asset, it’s betting on a structural shift. The shift: AI compute demand outstripping supply of high-density data centers.
But every structural shift has a counter-current.
The Context: Market Structure
Switch operates large-scale data centers across the US and Europe. Its clients? Hyperscalers like Microsoft, Google, and AWS. Also, crypto mining firms—though that footprint has shrunk since 2022.
The core business is selling capacity: power (watts), cooling (PUE), and physical space (rack units). It’s a real estate play with a tech premium.
Here’s what matters: Switch’s “S-Core” architecture claims higher power density per rack than competitors. That matters for AI—NVIDIA’s H100 requires 700W per GPU, and next-gen B200 will need 1000W+. Older data centers can’t handle that without massive retrofits.
So Switch has a moat? Maybe.
But the moat is only as deep as the capital available to maintain it. Switch needs billions to build new facilities. The $80B valuation assumes it can raise that capital cheaply.
Calculate. Execute. Repeat.
The Core: Order Flow Analysis
Let’s break down what the smart money is actually trading.
First, the demand side. In 2024, US data center absorption (new leasing) hit 5.2 gigawatts. That’s double 2023. AI workloads accounted for 60%. The shift from training to inference will only accelerate demand for low-latency edge sites.
Switch is heavily weighted toward core hyperscale locations (Las Vegas, Reno, Silicon Valley). Not edge. That’s a gap.
Second, the supply side. Global capex on data centers is projected to exceed $400B in 2025. That’s a 30% increase. Too much money chasing the same power sources. In Northern Virginia—the world’s largest data center market—new construction permits are bottlenecked by utility capacity.
Switch’s edge? Long-term power purchase agreements (PPAs). It locked in power rates before the AI boom. That gives it a cost advantage.
But here’s the trap: energy prices are rising. Natural gas futures are up 40% year-over-year. Solar and wind have intermittency issues. If Switch’s PPA contracts expire or become uncompetitive, margins compress.
In 2022, I watched a fund lose $1.2M in a week because counterparty risk wasn’t hedged. Same here.
Third, the competitive landscape. Hyperscalers are building their own data centers. Microsoft alone committed $50B to expand its infrastructure. That directly competes with Switch. If the biggest customers become competitors, Switch’s revenue growth slows.
The bull case: smaller AI startups and enterprises can’t build their own data centers. They rent. Switch captures that demand.
The bear case: the hyperscalers eventually overbuild, creating a capacity glut. Rack prices drop. Switch’s margins shrink.
Liquidity vanishes. Lessons remain.
The Contrarian: Retail vs. Smart Money
Retail sees Switch as a “picks-and-shovels” bet on AI. They hear “Nvidia supplier” and think it’s a winner.
Smart money sees something else: a leveraged play on interest rates.
Why? Because data centers are capital-intensive. They require debt financing. Switch’s $80B valuation assumes low interest rates for the next decade. If the Federal Reserve holds rates at 4-5% due to sticky inflation, the present value of future cash flows collapses.
In 2024, I wrote a script to backtest DCF valuations under different rate scenarios. A 1% increase in discount rate reduces fair value by 12-15% for capital-intensive businesses. Switch could be worth $50B, not $80B, if rates stay high.
That’s the contrarian angle retail misses.
Another blind spot: energy regulation. Europe is imposing strict water usage limits on data centers. California is considering a moratorium on new facilities in drought-prone areas. Switch’s largest hub is in Nevada—a desert state with water scarcity. If regulators cap expansion, growth stalls.
Smart money already priced that in. Retail hasn’t.
Data over drama.
The Takeaway: Actionable Levels
Switch hasn’t set an IPO price yet. But based on comparable multiples (EQIX trades at 25x EBITDA, DLR at 20x), here’s my framework:
- If IPO prices below $60B (20x EBITDA): It’s a buy. The market is discounting the infrastructure shortage. Accumulate.
- If IPO prices at $70-80B (23-26x EBITDA): Hold cash. Wait for first earnings dip. Institutions will front-run the hype, then rotate out.
- If IPO prices above $100B (30x+ EBITDA): Sell. The momentum chasers are pricing in a decade of perfect execution. One regulatory surprise or rate hike and the stock gets crushed.
Calculate. Execute. Repeat.
I’ve been on both sides of these trades. In 2017, I lost $7,500 in gas wars on Ethereum because I ignored infrastructure constraints. In 2024, I managed a $5M fund with algorithmic risk models—data centers were a core allocation.
Switch is not a story stock. It’s a machine. Every machine has a breaking point.
The question isn’t whether AI will change the world. It will. The question is whether Switch’s business model can survive the friction between hype and reality.
Numbers don’t lie. But they also don’t predict regulatory whims or energy crises.
You want to trade this? Watch the Fed. Watch power prices. Watch the water wars in Nevada.
Discipline beats conviction every time.
