Switch’s confidential IPO filing puts “independent data center” pricing to the test—while Westinghouse’s filing shows the window is still open

What we can verify from primary reporting is that a U.S. data-center operator named “Switch” is moving toward an IPO using a confidential filing route, and that separately a different infrastructure company, Westinghouse, also used confidential draft filing mechanics.

However, the confidential S-1 itself (the underlying file number, full financial schedules, and segment detail) is not fully disclosed in open web snippets, so this article focuses on what the reporting supports with hard numbers (timing/valuation/proceeds) and uses supply-chain logic to map what likely has to be true operationally for the IPO multiple to stick.

Independent operators don’t compete just on IT load—they compete on delivery speed under power constraints. A public-market multiple is basically shorthand for three underwriting beliefs:

1) contracted demand is real enough that expansion capex converts into revenue without long ‘dark’ delay. 2) funding structure (equity vs. debt) won’t force margin compression. 3) build-to-suit and AI-era leasing economics won’t structurally degrade due to oversupply later.

Switch’s confidential IPO reporting implies capital markets expect those beliefs to hold, because a potential ~$80B valuation (including debt) at a time when many private operators have already raised expensive funding would be hard to justify without strong monetization visibility.

This is why the article’s question is not “Will the IPO happen?”—it’s “Will the market pay a premium today for operational reliability that only shows up in future utilization?”

In other words: the confidential filing is a test of whether public markets are still willing to underwrite the AI buildout before the utilization curve is fully revealed.

Because the Switch confidential draft S-1 contents aren’t available to quote here, we can’t compute a precise EV/Revenue or EV/EBITDA-like metric from primary disclosure.

But the reported potential valuation (~$80B incl. debt) and potential proceeds (up to ~$10B) still let investors stress-test the implied story:

• If the operator is worth ~$80B including debt, equity market pricing must assume that contracted or highly visible demand supports large forward capex commitments.
• If it can raise up to ~$10B, underwriting implies the equity story can plausibly fund growth without worsening unit economics enough to trigger a re-rating.

Long-term (1–3 years), the real determinant is whether the operator can show that AI-driven demand doesn’t just raise occupancy initially, but sustains pricing power through renewal and expansion.

The competitive implication: public comps and private sponsor underwriting will adjust based on whether the IPO lands at a valuation consistent with durable gross margin and steady capital turn.

If these prove strong, independent operators can clear public multiples even in a higher-rate macro. If not, investors who bought the IPO narrative too early will likely face re-rating risk once the utilization curve is slower than the market’s story.

Switch’s reported confidential IPO pathway suggests public markets may be ready to pay for independent data-center operators again, but only if underwriting can treat grid-to-rack delivery as controllable.

The nuance for investors: the filing itself is not the edge. The edge is whether the IPO’s eventual pricing reflects a remaining scarcity of credible execution capacity—or whether the market already priced it in through prior private rounds and sector sentiment.