Rexford's Data Center Strategy: What Comes Next?
Rexford's latest moves signal a transformative shift in data center development. Are you ready for the changes ahead?
Rexford Industrial Realty built its reputation on one simple thesis: buy infill industrial real estate in Southern California and hold it while demand outpaces supply. For years, that playbook printed money. Now, the company is signaling something more ambitious — a move into data center development that could redefine how it deploys capital and, more broadly, how industrial landlords think about their highest-value sites.
The details emerging from Rexford's recent announcements are sparse but telling. A single-tenant site, now vacant, is apparently on the cusp of transformation. The phrase "presumably not for much longer" doesn't appear in a company update by accident. That's a signal, not a footnote.
Rexford's Current Position and What the Vacancy Actually Means
A vacant single-tenant building sounds like a problem. In Rexford's case, it may be exactly the opposite.
Infill industrial sites in Southern California — particularly in the Los Angeles Basin, the Inland Empire, and Orange County — are among the most constrained real estate in the country. Land is scarce, entitlements are brutal to obtain, and power infrastructure near population centers is increasingly difficult to expand. A vacant building on one of these sites isn't a liability; it's a blank canvas sitting on irreplaceable dirt.
The vacancy creates optionality. Rexford can reposition the asset without the friction of an occupied building — no lease termination negotiations, no tenant relocation costs, no waiting for a lease expiration years out. The timing, given surging data center demand along the West Coast, is either very good planning or very good luck. Probably some of both.
What matters strategically is that Rexford isn't talking about this site as a problem to solve. The framing around data center development suggests they see it as an opportunity to capture a market that traditional data center developers have struggled to serve in high-cost, land-constrained metros.
The Shift Toward Multi-Tenant Models — and Why It Changes the Risk Profile
Single-tenant data centers are straightforward but exposed. One customer, one lease, one renewal conversation every 10 to 15 years. When that tenant leaves — as one just did at this Rexford site — you're starting from zero.
Multi-tenant models distribute that risk across multiple operators, each with different lease maturities, different power requirements, and different reasons for being in a given market. The multi-tenant structure doesn't just reduce vacancy risk; it creates a recurring revenue base that looks more like a SaaS business than a traditional real estate lease.
The demand case for multi-tenant data centers in urban and infill markets is real and growing. Hyperscalers — Amazon Web Services, Microsoft Azure, Google Cloud — have largely claimed the large-format, purpose-built campuses in markets like Northern Virginia, Phoenix, and Dallas. What they haven't solved is latency-sensitive workloads that need to be physically close to end users. Edge computing, content delivery, financial services, and increasingly AI inference (as opposed to AI training) all require proximity to population centers.
Southern California, with roughly 24 million people in the greater metro area, is precisely the kind of market where multi-tenant colocation facilities command premium pricing. Operators pay for proximity. They'll pay more for power reliability. And they'll pay more still in a market where there simply aren't many options.
For Rexford, transitioning from a single-tenant industrial model to a multi-tenant data center model means trading simplicity for complexity — more customers, more technical requirements, more operational overhead. But it also means significantly higher rent per square foot and stronger long-term asset value.
The Real Challenges Nobody Talks About Clearly Enough
Data center development sounds like a gold rush. The reality involves a set of structural constraints that trip up even experienced developers.
Power is the first and hardest problem. A modern hyperscale data center can consume 100 MW or more. Even a mid-sized colocation facility might need 20 to 50 MW of dedicated capacity. Southern California's grid is not sitting on surplus power waiting to be allocated. Utility interconnection queues in California can stretch three to five years, and that's before you account for the state's aggressive decarbonization mandates, which complicate the procurement of firm, reliable power at the scale data centers require.
Site selection compounds the difficulty. Not every industrial parcel can support a data center. You need adequate fiber connectivity (ideally multiple diverse paths), sufficient land for cooling infrastructure and backup generation, appropriate zoning, and — critically — that power access. Rexford's infill sites are valuable precisely because of their location, but location doesn't automatically solve the power equation.
Regulatory hurdles in California add another layer. Environmental review under CEQA (the California Environmental Quality Act) can add 18 to 36 months to a development timeline for a project that faces any organized opposition. Data centers, with their diesel backup generators and water consumption for cooling, have drawn scrutiny from local regulators and environmental groups in other California markets.
Cost management is the thread running through all of it. Data center construction costs have climbed sharply — industry estimates suggest $10 to $15 million per megawatt for a fully built-out facility, depending on tier and location, with premium markets pushing higher. For a company like Rexford that has built its model on efficient capital deployment in industrial real estate, the jump to data center development capital intensity is significant. Joint ventures with specialized data center operators or capital partners will likely be part of the equation.
Where Data Center Development Is Actually Heading
The trends shaping the next five years of data center development aren't subtle. AI infrastructure demand is driving investment at a scale the industry hasn't seen before. Microsoft alone announced $80 billion in global data center investment plans for 2025. Google, Amazon, and Meta have made commitments in the same order of magnitude. That capital has to land somewhere, and increasingly, "somewhere" includes secondary and infill markets that previously couldn't compete with purpose-built data center hubs.
The interesting shift isn't in the hyperscale builds — it's in the mid-market colocation facilities that serve the enterprises and cloud operators who can't justify building their own. That segment is where a developer with Rexford's site quality, local market expertise, and balance sheet could carve out a durable position.
Investor interest is following the capital. Data center REITs like Equinix and Digital Realty have traded at premiums to the broader REIT market for years. Industrial REITs repositioning assets toward data center use cases are getting a second look from institutional investors who want infrastructure exposure without the pure-play volatility. Rexford's moves, even at early stages, are being watched for exactly this reason.
On the technology side, liquid cooling is displacing traditional air cooling as AI chips run hotter and denser. Facilities designed even five years ago are struggling to retrofit. New data center development — the kind Rexford would be building from essentially a greenfield position — has the advantage of designing for next-generation thermal requirements from the start rather than retrofitting legacy infrastructure.
What Industry Professionals Should Take From This
Rexford's positioning is a case study in how industrial real estate owners with the right assets can capture a structural shift rather than be disrupted by it. The company isn't abandoning its core competency — it's extending it into an adjacent market where its existing advantages (prime infill locations, entitlement expertise, Southern California relationships) translate directly.
For other industrial landlords, developers, and investors watching this space: the window for repositioning well-located industrial sites toward data center use is open but not unlimited. Power interconnection queues, entitlement timelines, and escalating construction costs all compress the advantage available to early movers.
The developers who figure out the power access problem first — through utility partnerships, on-site generation, or creative grid agreements — will set the terms for everyone who comes after them.
Rexford's next move on that vacant site will be worth watching closely. It may be the clearest signal yet of where smart industrial capital is heading.
Explore the InfraSale Marketplace for more insights and opportunities!
[INTERNAL LINK: data center development]
[INTERNAL LINK: industrial real estate trends]
[INTERNAL LINK: market analysis]