Data center developer Switch is in talks to raise billions of dollars from private equity firms including Brookfield and KKR, at a valuation the company is targeting above $50 billion. That number is worth sitting with for a moment: DigitalBridge and IFM Investors took Switch private in 2022 for roughly $11 billion. Four years later, the same company is seeking a valuation more than four times higher — not because its physical footprint quadrupled overnight, but because the demand curve for the thing Switch actually sells, data center capacity, has bent sharply upward since AI workloads became a mainstream line item in enterprise IT budgets.
What Switch does, and why its asset class specifically is commanding this premium
Switch operates large, campus-scale data center facilities designed for hyperscale and enterprise customers who need significant, reliable compute capacity rather than a handful of racks in a shared colocation space. That distinction matters to understanding why investors are pricing Switch at this level: campus-scale facilities are exactly the asset class that's become scarce relative to demand as AI training and inference workloads have scaled. A company that already owns the land, the power interconnects, and the physical infrastructure to build at that scale has something that can't be conjured quickly even with unlimited capital — data center construction timelines, power grid interconnection queues, and land acquisition in suitable locations all move on multi-year cycles regardless of how much money is chasing the opportunity.
As part of this expansion, Switch has a planned 382-acre data-center campus near Pittsburgh — a specific, concrete illustration of where this capital is actually going. That's not an abstract "AI infrastructure investment" line item; it's a real, large-scale physical build tied to a specific location, reflecting how developers are increasingly looking beyond the traditional data center hubs (Northern Virginia, Silicon Valley) toward regions with available power capacity and land, even if those regions weren't traditionally associated with major data center development.
Reading the fundraising mechanics: PE money, bank involvement, and an eventual IPO
Goldman Sachs and JP Morgan are reportedly working with Switch on this fundraise, and the round could set up the company for an initial public offering as early as next year. That combination — major investment banks structuring the deal, marquee private equity names like Brookfield and KKR as investors, and an IPO explicitly floated as a next step — signals this isn't a speculative venture round on an unproven business model. It's institutional capital positioning itself around an asset class (data center infrastructure serving AI demand) that's increasingly viewed as comparable to other forms of critical infrastructure investment, the kind pension funds and large asset managers allocate toward specifically because of its long-duration, contracted-revenue characteristics.
Separately, Switch also closed an asset-backed securities offering raising approximately $768 million — a distinct financing mechanism from the equity fundraise, and one worth understanding on its own terms. Asset-backed securitization essentially converts a stream of future revenue (in this case, presumably long-term data center lease and services contracts) into an upfront capital raise, backed by those contracts as collateral. Data center developers turning to this specific financing tool, alongside traditional equity fundraising, illustrates just how aggressively the industry is tapping every available capital-raising mechanism to fund the current buildout — a sign of how large and how urgent the capacity expansion race has become across the sector, not just at Switch specifically.
Why this matters beyond Switch's own cap table
For IT and infrastructure leaders who don't have any direct relationship with Switch, this valuation jump is still a useful data point, because it's a proxy for a dynamic playing out across the entire data center leasing and cloud capacity market. When the underlying infrastructure asset class commands valuations quadrupling in four years, that capital cost has to be recovered somewhere in the pricing chain — through data center lease rates for enterprises that colocate their own hardware, through cloud provider pricing for compute-intensive workloads, or through the pricing of AI-specific infrastructure services that depend on this physical capacity existing somewhere.
This is also a useful lens for understanding why capacity, not just price, has become the binding constraint for a lot of enterprise AI initiatives in 2026. If the market for physical data center capacity is bidding up asset valuations this aggressively, it's a strong indication that available capacity is scarce relative to demand right now — which is consistent with the broader pattern of power availability, GPU allocation, and data center capacity all becoming genuine planning constraints for enterprise AI programs in a way that simply writing a bigger check doesn't automatically resolve. Money can now buy priority in a queue faster than it can conjure additional capacity that doesn't yet physically exist.
The regulatory backdrop this fundraising is happening against
This capital raise is also notable for the environment it's occurring in. Several U.S. jurisdictions have started pushing back on unconstrained data center growth over community concerns about power grid strain and land use — a dynamic that's added real regulatory friction to where and how quickly new hyperscale capacity can be built in certain markets. That backdrop makes Switch's specific choice of a Pittsburgh-area location, rather than a more heavily contested market, a data point worth watching: developers are increasingly routing new capacity toward regions with more available power headroom and less regulatory friction, which has real implications for where enterprise customers should expect new capacity to actually come online fastest over the next several years.
Comparing Switch's trajectory to the broader data center financing wave
Switch's fundraising talks aren't happening in isolation, and it's worth situating them against the broader wave of data center capital activity that's defined 2026 so far. Across the industry, developers have been tapping an unusually wide range of financing mechanisms in parallel — traditional private equity rounds, asset-backed securitization, corporate bond issuances explicitly tied to AI infrastructure buildout, and in some cases public listings — often within the same twelve-month window rather than sequentially. That breadth of financing activity, rather than any single mechanism, is the real signal of how capital-constrained the current buildout race has become: developers aren't just choosing the cheapest available financing option, they're pursuing multiple mechanisms simultaneously because the sheer scale of capital required to compete for hyperscale capacity has outgrown what any single financing channel can efficiently provide.
Switch's specific combination — a private equity-led valuation round targeting over $50 billion, paired with a separate $768 million asset-backed securities offering — illustrates that pattern precisely. It's not an either-or choice between equity and debt-like financing; it's both, running concurrently, because the scale of the opportunity these developers believe they're capturing justifies drawing on every available capital source at once. For enterprise IT leaders evaluating data center or colocation partners, the specific financing mix a developer is using is itself a reasonable signal of how seriously capitalized, and how aggressively expanding, that partner actually is relative to competitors still relying on a single, more conventional financing channel.
Practical takeaways
Treat data center capacity, not just budget, as a planning constraint for any AI initiative with meaningful compute requirements over the next 12-24 months — asset valuations rising this sharply are a market signal that physical capacity is scarce relative to demand, and that scarcity doesn't resolve simply because your organization has funding approved. If you're negotiating colocation or dedicated capacity agreements, expect upward pricing pressure tied to this broader capital dynamic, and factor multi-year capacity commitments into contract negotiations now rather than assuming current pricing holds steady. Watch where new hyperscale capacity is actually being sited — markets with available power and less regulatory friction, like the Pittsburgh region in Switch's case, are likely to see new capacity come online faster than more contested traditional hubs, which matters for latency-sensitive planning and regional data residency decisions. If your organization is evaluating build-versus-lease decisions for AI infrastructure, use asset-backed securitization activity like Switch's $768 million offering as a signal of how capital-intensive and how urgently financed this entire category has become — that's relevant context for negotiating leverage with any data center or cloud provider partner. And keep an eye on Switch's potential IPO timeline specifically if your organization has any direct commercial relationship with the company, since a public listing often brings changes to contract terms, reporting transparency, and capital allocation priorities that private ownership didn't require.
Switch's valuation jump from roughly $11 billion to a targeted $50 billion-plus in four years isn't really a story about one company's financial engineering — it's a concrete, dollar-denominated measurement of how much the market believes physical data center capacity is now worth, driven almost entirely by AI demand that didn't exist at this scale when Switch was taken private. That's a leading indicator worth tracking for anyone whose own AI infrastructure plans depend on capacity someone else has to build first.