Data center IPOs are selling the construction curve

DayOne, SB Energy and Switch filed IPO registrations on September 13, 2026, offering public equity the same underlying: development pipelines still working through county approvals, utility interconnection queues and construction financing. That distinction is the trade.

A stabilized data center is a credit and duration problem, with the lease signed, the power connected, and the valuation moving on tenant quality and lease length; a development pipeline is a construction and energization problem, where the risks are zoning moratoriums, substation upgrades and cost-overrun contingencies before a single tenant payment exists. The same capital market that has been buying privately traded or securitized operating data centers is now being handed the part of the stack it has usually paid private developers to carry on their own balance sheets or fund through private credit.

The filing date matters more than the names. Three registrations landing on the same day suggests a coordinated test of whether public equity will accept pipeline risk at infrastructure-style multiples, and the asset they are selling is the construction curve—precisely the exposure private developers have kept on their own balance sheets or funded through private credit.

Private data center marks have leaned on one reference point: a stabilized portfolio sale, where a leased asset changes hands at a multiple of in-place income. That reference point is increasingly disconnected from the actual pipeline because it does not price the risk that a project never gets energized. The DayOne, SB Energy and Switch filings make that disconnect visible, since a public filing has to disclose the pipeline's route through permitting and power, and the market will soon assign a price to it.

The risk being handed to public investors is not spread evenly across the country. Ten U.S. counties hold 42 percent of the nation's data centers, while 92 percent of counties have none, and Atlanta alone has 2,882 megawatts of data center capacity under construction, with 39 county moratoriums deciding what comes next. A listed data center pipeline is a concentrated bet on a small number of local decisions, not a diversified claim on the sector's growth.

The concentration means the public denominator will be local before it is national. A pipeline in a county with an active moratorium will carry a discount that a project in an open county does not, and the IPO market will make that difference observable. Private transactions have been able to bury a county's permitting risk inside a portfolio-level valuation; a daily traded security cannot.

For private owners, the pricing of DayOne, SB Energy and Switch will not stay inside the public market; it will become the reference point that private data center platforms use to justify their own development pipelines. A private fund that has been marking its pipeline against stabilized sales will now have to explain why its mark should trade through a listed vehicle that carries the same county and power-queue risks but reprices every day, a sharper discipline than a once-a-quarter valuation committee.

If the IPOs price pipelines at multiples close to stabilized portfolios, private marks look validated and the development pipeline becomes a more liquid currency for sponsors; if the public market demands a wider discount for construction and energization risk, the private marks that ignored that discount will have to adjust, and the cost of capital for private data center development will rise just as demand is accelerating. The registrations are the test that decides which way the multiple goes.

The Atlanta figure gives the market something to underwrite against: 2,882 megawatts under construction is a queue of projects that need substation capacity and county consent, not a vague national growth story. The 39 moratoriums sit directly against that pipeline, and a public investor pricing these IPOs will have to decide county by county whether the moratorium is a temporary political friction or a permanent constraint on energization—a fundamentally different analytical job from buying a stabilized data center portfolio.

The three filings are not the first data center IPOs, but they are the first public market test of pipeline risk as the core of the offering. DayOne, SB Energy and Switch may not all price, and the registrations themselves do not yet include a multiple, but the test has begun, and every private data center owner now has to prepare a mark that can survive a public market's version of the same question: what is a development pipeline worth before the power is flowing? The first filing to publish a price range is the next data point—the day private data center marks stop being a quarterly opinion and become a daily comp.