JLL: The Perfect Data Centre Land Acquisition Plan

For data centre operators, the days of selecting a site based solely on cheap land and fibre proximity are over.

As we step into 2026, the industry is grappling with a stark new reality: power is the preeminent currency that matters.

A new report from JLL, the 2026 Global Data Center Outlook, outlines a comprehensive roadmap for navigating this constrained landscape. 

The headline figure is staggering – nearly 100GW of new capacity is expected to come online between now and 2030, requiring an investment supercycle of up to US$3tn.

But with grid connection delays stretching beyond four years in major hubs, the playbook for land acquisition has fundamentally changed.

According to JLL, data centre operators’ acquisition plans require a complete inversion of traditional development logic. 

It demands a strategy that prioritises energy independence, proactive community "co-creation" and a flexible approach to the looming shift from AI training to inference.

The power paradigm shift

The most critical takeaway for 2026 is that power availability has emerged as a crucial site selection criterion beyond the location alone.

The report warns that sites without secured power are "increasingly stranded", regardless of their zoning or demand profile.

This crunch is forcing operators to become energy producers.

JLL notes that tenants and developers are increasingly bypassing utility bottlenecks by funding their own "behind-the-meter" generation and exploring co-located battery storage systems (BESS).

This shift is echoed by the industry’s heavyweights.

Satya Nadella, Chairman and CEO of Microsoft, recently highlighted this bottleneck: "The biggest issue we are now having is not a compute glut, but it's power."

He added that “the ability to get the builds done fast enough close to power” is imperative. 

For land acquisition teams, this means the due diligence phase must now include a viable energy production strategy from day one.

Relying on a utility promise for a 2030 connection is no longer a bankable asset.

The ‘acceptance paradox’

Securing power is only half the battle.

JLL’s research uncovers a "universal acceptance paradox" that threatens to derail projects globally. 

While national support for digital infrastructure sits at a high 93%, local acceptance plummets to just 35% – a 58-point perception gap that is fuelling organised resistance.

The report advises that the industry must move from "reactive consultation" to "proactive co-creation", extending pre-development engagement timelines to 24 months.

This is not just about good PR – it is a risk mitigation strategy for valuations.

Brad Smith, Vice Chair and President of Microsoft, emphasised the importance of this integrated approach during a recent update on the company’s Community-First AI Infrastructure initiative.

He promised “a commitment to do this work differently than some others and to do it responsibly”. 

“This commits us to the concrete steps needed to be a good neighbour in the communities where we build, own, and operate our data centres,” he continued. 

“It reflects our sense of civic responsibility as well as a broad and long-term view of what it will take to run a successful AI infrastructure business. In short, we will set a high bar.”

For data centre operators, the sentiment stands that if you cannot bring the community along with you, your powered land is effectively worthless.

Inference vs training: The geography of 2027

For operators planning their land banks, understanding the "what" is as important as the "where".

JLL predicts a pivotal shift in 2027, when AI inference workloads – the application phase of AI – will overtake training as the dominant driver of demand.

This evolution will bifurcate the market. Training clusters will remain in centralised, high-density locations where power is massive and cheap.

However, inference will require "geographical distribution to reduce latency", pushing development towards regional hubs and the edge.

This trend opens new opportunities for land acquisition in secondary markets, provided operators can solve the power puzzle.

The report suggests that by 2030, AI could represent half of all workloads, with inference being the primary engine.

Sustainability as a licence to operate

With the industry consuming an ever-growing share of global electricity, sustainability has moved from a corporate social responsibility metric to a hard regulatory requirement. 

JLL notes that data centres will face "heightened scrutiny" in 2026, with operators needing to prioritise access to compliant energy mixes to avoid reputational risk.

Major operators are already moving to insulate themselves from these pressures.

Andy Power, President and CEO of Digital Realty, confirmed his firm’s progress in this area.

"Digital Realty continued to make strong progress towards its sustainability goals in 2024, reaching 1.5GW of renewable energy capacity under contract and matching 185 data centers with 100% renewable energy," he said.

"This reflects our commitment to build, power, and operate sustainable data centres, which is trusted and relied upon by more than 5,000 customers."

The verdict

The JLL outlook makes it clear that the "perfect" land acquisition plan in 2026 is complex.

It requires looking beyond the dirt to the grid, engaging with communities before a single spade hits the ground and designing for a hybrid future of core training and edge inference.

As the industry moves through this infrastructure supercycle, the winners will not necessarily be those with the deepest pockets, but those with the most foresight.