How 22 Data Centers Are Shaping the Future of Tech

Three gigawatts. That's not a number you stumble across often outside of power grid engineering discussions — but it's exactly what one company now manages across 22 operational data centers, with a fundraising push underway that signals even bigger ambitions ahead.

To put that in perspective: 3 GW of capacity could power roughly 2.25 million average American homes. Instead, it's running the compute workloads, storage systems, and network infrastructure that big tech increasingly can't live without. That reallocation of power — from households to hyperscale infrastructure — tells you something important about where capital and electrons are flowing in the modern economy.

The Scale Behind the Numbers

Twenty-two data centers sounds like a lot until you realize what each one represents: years of permitting, hundreds of millions in construction costs, complex power procurement agreements, and cooling infrastructure engineered to handle heat loads that would overwhelm most industrial facilities.

Managing 3 GW across two dozen facilities isn't just an operational achievement — it's a logistical proof of concept that very few organizations in the world have demonstrated.

For context, the largest hyperscale operators — Microsoft, Amazon, Google — each manage dozens of campuses globally, but they've spent decades and hundreds of billions building those portfolios. A company reaching 3 GW through 22 facilities has compressed what used to take much longer into a tighter execution window. That compression is only possible because the demand signal has been deafening.

The demand isn't coming from one corner of the market. Enterprise cloud migration, AI model training, real-time inference workloads, streaming, financial services compute — these are distinct use cases with distinct infrastructure requirements, and they're all growing simultaneously. A 3 GW portfolio positioned to serve multiple verticals is worth considerably more than one optimized for a single customer type.

What Big Tech Actually Needs from Infrastructure Partners

Here's something the press releases don't usually say clearly: the major cloud and AI players don't just need megawatts. They need *reliable, scalable, well-located* megawatts — and that distinction is doing a lot of work.

Reliability means redundant power feeds, N+1 or 2N cooling configurations, and interconnection to fiber routes that don't have single points of failure. Scalability means a campus that can absorb a 200 MW expansion without rebuilding the substation from scratch. Well-located means proximity to renewable energy sources, transmission capacity, permitting-friendly jurisdictions, and — increasingly — water availability for cooling.

Most real estate developers can build a shell. Very few can deliver a 100+ MW powered shell with a 99.9999% uptime SLA and a direct fiber path to three major exchange points.

The companies that have cracked this combination are the ones attracting long-term lease commitments from hyperscalers. Those leases — often 10 to 15 years with extension options — are what make data center portfolios attractive to institutional capital. The infrastructure itself becomes almost bond-like in its cash flow predictability, which explains why the fundraising environment for proven operators remains resilient even when broader commercial real estate is under pressure.

Investment Trends: Why Capital Keeps Flowing

The fundraising activity surrounding this 22-facility portfolio reflects a broader pattern: institutional investors have largely concluded that data center capacity is not cyclical infrastructure. It's secular infrastructure — meaning demand doesn't contract meaningfully during economic downturns the way office or retail occupancy does.

During the 2020 pandemic, while other commercial real estate categories cratered, data center absorption accelerated. Companies weren't cutting their cloud spend — they were increasing it, because remote operations and digital services became mandatory overnight. Investors noticed, and the capital allocation toward data center fundraising has been aggressive since.

That said, not all data center investment opportunities are equal. The spread between a well-capitalized operator with long-term hyperscale leases and a smaller colocation facility with month-to-month tenants is enormous — in both risk profile and valuation. Investors who understand data center growth trajectories look specifically for operators with demonstrated land banking strategies, existing utility relationships, and the engineering depth to execute at scale.

A portfolio already at 3 GW with 22 operational sites has cleared the hardest hurdles. The power contracts are signed. The facilities are proven. The remaining question is whether the capital being raised will be deployed efficiently into the next tranche of capacity — and whether the demand to fill it materializes on schedule.

It almost certainly will. AI compute requirements alone are projected to drive data center power demand to unprecedented levels through the remainder of this decade, with some analysts estimating that AI-related workloads could account for 20% or more of total data center electricity consumption by 2030, up from low single digits just a few years ago.

Where Growth Is Headed

The next phase of data center growth won't look exactly like the last one. The hyperscale campuses of the 2010s were largely concentrated in established markets: Northern Virginia, Phoenix, Dallas, Chicago, Seattle. Those markets are now facing real constraints — power availability, water stress, permitting backlogs, and community opposition in some cases.

The smart operators are already working in secondary markets: the Midwest, the Southeast, and portions of the Mountain West where renewable energy is abundant and land costs haven't yet reflected the demand pressure. A company with an existing 22-facility footprint has built institutional knowledge about navigating these development cycles — knowledge that translates directly into competitive advantage when expanding into new markets.

Technological advancement is also reshaping the economics. Liquid cooling — both direct-to-chip and immersion cooling — is moving from niche to mainstream as GPU-dense AI racks push power densities above 30, 50, sometimes 100+ kW per rack. Traditional air cooling becomes economically and physically impractical at those densities. Operators who have invested in liquid cooling infrastructure aren't just future-proofed — they're positioned to win the AI infrastructure contracts that air-cooled competitors simply can't bid on.

The Real Challenges Nobody Wants to Talk About

For all the capital flowing into the sector, data center development has genuine friction points that don't disappear with fundraising success.

Power procurement is the most acute. Utilities in high-demand markets are quoting interconnection timelines of 4 to 7 years in some cases — meaning a data center breaking ground today might not have full utility power until the end of the decade. Companies managing this successfully are doing so through a combination of on-site generation, power purchase agreements with renewable developers, and creative interconnection strategies that don't rely solely on the utility queue.

Water is a quieter crisis. A large air-cooled data center can consume millions of gallons annually for evaporative cooling. In drought-stressed regions, this is becoming a genuine political liability, with some municipalities refusing permits specifically over water concerns. The shift to liquid cooling partly addresses this — closed-loop liquid systems use dramatically less water — but the transition takes time and capital.

The operators who thrive through the next growth cycle will be the ones who treat power and water not as utilities to purchase, but as strategic resources to develop.

Workforce is the third constraint. Skilled data center technicians, power engineers, and critical infrastructure specialists are in short supply across every major market. Companies building at scale need to think about talent pipelines the same way they think about power pipelines — with long lead times and deliberate investment.

What This Means for the Market

A company reaching 22 operational data centers and 3 GW of managed capacity has effectively demonstrated the full stack of what institutional-quality data center development requires. That proof of execution is what commands premium valuations, attracts anchor tenants, and sustains the fundraising momentum necessary to keep building.

The more interesting question isn't whether demand for data center capacity will continue — it will, at rates that will surprise even optimistic forecasters. The question is which operators have the power relationships, the engineering talent, the balance sheet discipline, and the market positioning to capture a meaningful share of what's coming.

Portfolios built on 22 proven facilities and 3 GW of operating history are starting from a position most competitors can't replicate quickly. In infrastructure, that kind of head start tends to compound.

Ready to explore the future of data centers? Visit [InfraSale Marketplace](https://infrasale.com/marketplace) to discover more.

[INTERNAL LINK: data center investment trends]

[INTERNAL LINK: infrastructure partnerships]

[INTERNAL LINK: AI compute requirements]