Industry executives and experts share their predictions for 2026. Read them in this 18th annual VMblog.com series exclusive.
By Joe Minarik, Chief Operating Officer, DataBank
The data center industry enters 2026 at a crossroads. After years of explosive growth driven by AI adoption and hyperscale demand, the market must now confront fundamental questions about capacity, capital, and credibility. The coming year will separate viable operators from speculators and test the industry’s ability to deliver on commitments. It will also reveal whether AI workloads can mature from promise into productivity.
In particular, I predict three different market dynamics will determine which data center operators thrive and which struggle in 2026.
Prediction #1: Supply Shortages Persist Through 2026
Demand for data center capacity remains exceptionally high, and supply won’t catch up anytime soon. Most new capacity coming online in 2026 represents projects first initiated years ago, and even some of those face significant risk.
Power constraints, supply chain bottlenecks, and ongoing labor shortages all create a perfect storm of potential delays. Every operator is scrambling to deliver promised capacity, and the reality is that some will miss their targets. The variables are simply too volatile, with too many dependencies outside data center operators’ direct control.
I believe the supply crunch will begin to ease in late 2026 or early 2027, when the majority of planned capacity reaches the market. Until then, providers with locked-in capacity and proven execution capabilities will meet strong demand, while customers seeking immediate capacity face limited options and extended timelines.
Those companies seeking data center capacity in 2026 need to account for these constraints and partner with providers who have demonstrable track records of execution, not just ambitious announcements.
Prediction #2: The Speculator Shakeout Accelerates
The data center market attracted a flood of speculators over the past two years. In some cases, developers from industrial, warehousing, and commercial office sectors recognized the boom and rushed in, submitting power requests without secured land, committed customers, or realistic capital plans. These speculators jammed utility interconnection queues and created artificial backlogs that masked actual demand.
Utilities are responding aggressively. For example, when Ohio utility AEP Ohio tightened financial requirements for data center projects, their interconnection queue dropped by nearly 60 percent. Additionally, many utilities are now requiring substantial deposits-in some cases, millions of dollars per megawatt-just to enter the interconnection queue. These barriers will force undercapitalized speculators out of the market entirely.
The resulting shakeout should create significant opportunities for well-capitalized operators with proven track records. As speculators exit and release their positions, legitimate providers gain better access to power and will face shorter wait times.
Consolidation will likely pick up toward the end of 2026 as some overleveraged developers sell off projects they can’t continue to fund. This market correction will benefit data center operators who can actually execute.
Prediction #3: AI’s Extended Learning Curve
AI workloads driving data center demand need time to mature before delivering on their potential. Deploying the infrastructure is just the starting point. These systems need extensive training to learn, improve accuracy, and become genuinely useful at scale. The market wants immediate results, yet that’s not how AI works.
We’ll start seeing viable AI models emerge in late 2026 and into 2027, with meaningful market impact hitting in 2027-2028. Agentic AI should make real progress during this window, moving beyond hype to demonstrate practical applications. We’re already seeing this work when AI gets applied to specific, focused problems rather than trying to solve everything at once. The more targeted the training, the better the results.
Here’s the wild card: will AI deliver a breakthrough that justifies all this infrastructure investment? If something spectacular happens in 2026 or early 2027, demand will explode. If breakthroughs don’t arrive until 2028 or later, expect market turbulence. Skepticism will grow. Questions will intensify about whether this AI revolution is actually real. We’ve built the capacity to support transformative applications. Now comes the hard part: waiting to see if those applications materialize on schedule.
Navigating Uncertainty with Confidence
The data center industry has weathered dramatic shifts in the past, but 2026 could demand something different: the ability to distinguish signal from noise. Data center operators who understand these market dynamics, maintain disciplined capital strategies, and resist the temptation to chase hype will have the best chance to emerge stronger.
Equally important will be transparent engagement with communities, countering misinformation about power consumption, water usage, and infrastructure impact before concerns escalate into opposition. The test isn’t surviving disruption. It’s recognizing which disruptions actually matter and positioning accordingly.
##
ABOUT THE AUTHOR
Joe Minarik is DataBank’s Chief Operating Officer and is responsible for all data center operations, engineering, construction, managed services, and IT operations.
