Uptime Institute warns that the gap between data center construction timelines and power infrastructure availability will exacerbate industry challenges.
The data center industry is racing toward a power crisis that technology alone will not be able to solve, according to Uptime Institute’s 2026 data center predictions report.
“Critical digital infrastructure continues to expand strongly,” said Andy Lawrence, executive director of research at Uptime Institute, in a statement. “At the same time, our research shows uncertainty about how AI will reshape demand. This is complicating both capacity planning and resiliency strategies. We are also seeing increasing fragmentation in the design and deployment of data centers and expect investment and innovation in carbon capture technologies, in AI, and automation in the data center itself.”
While AI continues to drive investment in digital infrastructure, the growing gap between how quickly data centers can be built and how long it takes to bring new power generation and transmission online is a primary concern.
“This year’s predictions build on those for 2025 and focus on the industry’s continued growth and associated challenges, while also recognizing AI as a powerful, transformative accelerant to growth,” Lawrence said, during a webinar presenting the findings.
Here are Uptime’s top five data center predictions for 2026:
1. The AI ecosystem becomes more concentrated
While enterprises experiment with AI tools, the infrastructure to support AI is becoming more concentrated at the high-end of the market, among a small number of organizations, according to Uptime Institute.
“Despite a broader use of AI tools in enterprises and by consumers, that does not mean that AI compute, AI infrastructure in general, will be more evenly spread out,” said Daniel Bizo, research director at Uptime Institute, during the webinar. “The concentration of AI compute infrastructure is only increasing in the coming years.”
For enterprises, the infrastructure investment remains relatively modest, Uptime Institute found. Enterprises will limit investment to inference and only some training, and inference workloads don’t require dramatic capacity increases.
“Our prediction, our observation, was that the concentration of AI compute infrastructure is only increasing in the coming years by a couple of points. By the end of this year, 2026, we are projecting that around 10 gigawatts of new IT load will have been added to the global data center world, specifically to run generative AI workloads and adjacent workloads, but definitely centered on generative AI,” Bizo said. “This means these 10 gigawatts or so load, we are talking about anywhere between 13 to 15 million GPUs and accelerators deployed globally. We are anticipating that a majority of these are and will be deployed in supercomputing style.”
2. Developers will not outrun the power shortage
The most pressing challenge facing the industry, according to Uptime, is that data centers can be built in less than three years, but power generation takes much longer.
“It takes three to six years to deploy a solar or wind farm, around six years for a combined-cycle gas turbine plant, and even optimistically, it probably takes more than 10 years to deploy a conventional nuclear power plant,” said Max Smolaks, research analyst at Uptime Institute.
This mismatch was manageable when data centers were smaller and growth was predictable, the report notes. But with projects now measured in tens and sometimes hundreds of megawatts, finding that much power quickly has become nearly impossible.
“My prediction is that the scale and severity of the crisis that will begin to emerge in 2026 will catch many operators unprepared, unfortunately. Power generation and distribution equipment will be the deciding factor when choosing what can be built, how it can be built, and where, and the crisis is likely to last many years,” Smolaks said.
Uptime predicts that power costs will rise in tier-one markets, developers will increasingly look to second-tier locations, and data centers may be required to participate in demand response programs—switching to on-site generation during peak grid demand—as a condition of grid connection.
3. Operators look to carbon capture
As operators scramble to secure power through natural gas generation, many face a challenge. They must decide if they will access the generation needed to support growth or meet facility-specific net-zero carbon commitments.
Uptime Institute predicts that carbon capture and storage (CCS) systems will move from theoretical to practical for some operators in 2026 and beyond. While CCS technology has long been viewed as uneconomic, several factors are shifting the calculation, according to the report. Improving technology, the high cost of alternatives, and rising prices for carbon offsets are making CCS look more competitive. Still, Uptime reported that CCS only makes sense in geographies where captured carbon can be stored locally or transported easily at low cost.
“One of the ways that this will be done is through natural gas generation, which will very often be in both grid-based gas turbines and on-site gas turbines. We are seeing a lot more interest amongst many data center operators to work with companies that will integrate CCS technology,” Lawrence said. “The essence of this prediction is that we think it is becoming realistic at scale at some of these sites, and that’s partly because of improving technology, the high cost of alternatives. So, there are things going on in CCS. It is becoming a realistic technology and something that can be deployed at scale at certain sites.”
4. Scale adds new challenges
As data centers grow larger and more concentrated in specific regions, Uptime warns that resiliency risks are also growing.
“Scale, huge data centers, not only the size of individual data centers and campuses, but also the concentration of large data centers in specific geographies, is introducing quite a lot of new risks,” Lawrence said.
Several factors are creating new challenges, according to the report: Aging transmission equipment and increasing shares of intermittent renewable generation are creating conditions for grid instability. Large concentrations mean data centers can affect grid stability. Regulators have taken notice, Uptime notes. Authorities in several countries are working on new grid connection rules that will require data centers to tolerate higher voltage fluctuations and avoid or slow down load disconnection during grid disturbances.
“The key takeaway, or what we see in the next year or two, is that the industry isn’t moving towards full autonomy. It’s actually going to be moving toward supervised, practical automation that will be deployed carefully, scaled gradually, and designed to support operators rather than actually replace them,” said Rand Talib, research analyst at Uptime Institute.
5. AI automation in the data center moves to production
AI-driven automation within data centers will begin transitioning from experimental pilots to supporting daily operations, Uptime predicts.
“In 2026, data center operations will begin shifting from experimentation with AI to early, targeted deployments,” Talib said. The automation will focus on reinforcement learning and hybrid digital twins for cooling and power optimization, industrial copilots that help operators with workflows, and smarter rules-based orchestration for automated responses to live sensor data, according to the report.
“The goal in the next year or two is not going to be full autonomy,” Talib said. “It’s going to be more to reduce the manual effort and improve consistency.”
Hyperscalers and large colocation providers will advance more quickly, while most enterprises will progress cautiously through incremental upgrades. According to Uptime Institute, the easier solutions to power constraints have been exhausted, and the rush to support AI is straining grids, supply chains, and traditional approaches to resiliency.
“AI has only recently become the dominant narrative, and most operators are continuing to plan in the face of uncertainty rather than executing on clearly AI-driven demand,” Lawrence said.
