Kevin Ooley has spent 15 years watching other people not understand what he does for a living. Most of the time, even his family didn’t have a clean answer to the party’s question. Now the sector he used to build a job can’t avoid it—included in every heat wave
DataBank, founded before anyone talked about big language models, has ridden three different waves since Ooley joined: the outsourcing of corporate data centers since 2011, the cloud explosion in the mid-2010s, and, since late 2022, the AI wave that has reshaped the scale of everything the company does. “The data bank has been riding those waves,” Ooley tells the Observer. The arithmetic of that trip is wrong. When DigitalBridge acquired DataBank in 2016, the company was operating at approximately $25 million in annual EBITDA; DataBank expects to close this year near the $1 billion run rate. It has grown from six data centers in three markets to what Ooley describes as the largest of any data center provider in the US—about 65 to 66 centers in 26 markets, with about 10 more under development. The company hit its 1,000th employee, or “data banker,” in March.
What has changed since 2022 is not the growth trajectory but its position. The DataBank was used to build selectively—a 10-megawatt facility, combine a few megawatts, sell it, build the next hall. As the project grew, the company began purchasing land adjacent to its existing campuses. Now it’s popping up on 200- to 300-acre campuses that are clearly built for expansion, and they’re selling to that capacity faster and earlier than ever. “In 2019, you needed to have a white-floor data center, ready for a client to visit,” Ooley said. “Now you have negotiations when you’re approved, and you get your operating agreement. You can pre-lease twelve, eighteen months in advance.”
I Water Something
Ooley knows that “data center” has become a shorthand for public concerns, and he has some complaints about what facts are being lost in that conversation. Older DataBank facilities, built since 2008, use energy-efficient evaporative cooling technology.
In power, he is alert. Costs are rising—inflation, prices, transformer and grid maintenance costs—but he posits the need for a data center as a rare opportunity for utilities that have seen little load growth over the years to finally invest in modernizing and robust infrastructure that benefits everyone, not just their tenants. He also shows a desire to participate in the responses. Because the data center’s load is flat rather than spiky, DataBank can pull large chunks of usage offline during peak periods, with the idea of saving customers who live in a particular strain zone. That those figures come out well for taxpayers, in particular, isn’t a claim he’s trying to scotch right away. “I don’t want my housing prices, and my family’s prices, to increase,” he said, “but we think that over time with this investment, in partnership with utilities, we can move to renewable investments… and ultimately reduce electricity costs.”
AI Bubble Talk
Asked directly if the industry is inflating a bubble it can’t sustain, Ooley doesn’t shy away, but he doesn’t overpromise either. He and much of DataBank’s leadership lived through the dot-com crash. In his words, he was “a paper millionaire in 1999 and 2000,” and that memory creates a deliberate wariness about overextending. Construction costs have increased in terms of inflation and interest, and DataBank has passed some of that on to prices. Today, he says, the industry is in a clear state of supply-demand imbalance, as demand outstrips supply and social backlash continues to squeeze new capacity—situations that, in the near future, point to price pressures for AI customers rather than relief.
Where it comes in the long term, Ooley won’t confidently predict. He expects consolidation in line with what happened after the dot-com crash and in the cloud sector, where hundreds of providers have reached a handful of prominent players and experts. He draws a contrast between the AI use cases already woven into everyday infrastructure — the kind no one wants to release, the way no one wants to go back to calling an airline to reserve a seat — and the open question of how far AI reaches into consumers’ lives. At the moment, he points to concrete, little-used applications that work in DataBank institutions: pancreatic cancer detection research in health clinics, AI-driven fraud detection in financial institutions. “Fraud detection has always been around,” he says, “but this is, you know, on steroids.”
What He Would Do Differently
Ooley says he’s never been an AI skeptic, but he underestimated how quickly the return on investment is for end users. DataBank used Claude Code internally to speed up its development work, he noted as one example among many. He compares the adoption curve to self-driving, which spent a decade feeling stuck before suddenly gaining real traction. The speed and rate of AI adoption, he says, is surpassing anything the industry has seen. “If I could go back, we would have built bigger and faster five years ago.”
Succession, Culture and Part Four
Ooley explains his reaction to the CEO’s announcement with a small, telling anecdote. Coming out of the ceremony where the change was announced, Martynek told him that he felt a weight had been lifted from his shoulders. “Hey, thanks a lot, Raul,” Ooley said, “that weight just came off my shoulders.” He puts his excitement further, making the existing plan with the existing leadership team, continuing to work with Martynek in his new position as Executive Chairman, and expanding the DataBank employee ownership program, in which any data keeper employed for a year or more participates in the company.
His concerns are very strategic: getting the right power in the right places, being a reliable partner to host communities, and managing capital-intensive, large-scale buildings without overbuilding or falling behind. He cites DataBank’s work in Culpeper, Virginia, where the company helped establish a designated “technology zone” for data center development away from residential areas, partnering with Dominion Energy on energy availability.
When hiring, Ooley credits a culture change from four or five years ago, caused by friction between people and workflows across DataBank’s various acquisitions. The third party helped the company identify the “archetypes” of behavior – the “hero” who tries to solve the problem and the “caretaker” who tries to protect the customer, both pursuing the same goal in different ways – and from that process, DataBank built what it calls its cultural foundations: put people first, take a positive aim, focus on data. Leadership speaks openly Good for Grandfatherit prioritizes high competence paired with low ego, and has direct conversations about problems rather than people. The company publishes its Net Promoter Scores—customer satisfaction ratings, now at a high level—internally, and an Employee Net Promoter Score used to gauge employee sentiment and adjust culture accordingly.
He added a fourth to the traditional trio of investors, customers and employees: community neighbors, a group he expects to be in the middle of his first year as CEO. DataBank has three solar projects underway, including what Ooley described as the largest rooftop solar installation in the Houston metro area. The company says it is 80 percent renewable energy today, working toward a 100 percent goal by 2030, up from 40 percent three years ago.
Who Gets Space
How does DataBank actually decide who will lease space when demand exceeds supply? Is it simply a high-end buyer, or is there a deliberate mix of tenants, as a landlord might choose to sell near restaurants? About 70 percent of DataBank’s business comes from enterprise customers—financial services, health care, content distribution, software—and 30 percent from hyperscalers. The company prioritizes investment-grade, long-term clients because that client profile helps secure cheap capital when DataBank raises collateralized bonds or equity to finance new buildings. Existing customers also tend to grow closer to DataBank; about 80 percent of the company’s growth comes from accounts already on the books, served by a team of about 100 salespeople, with only 10 percent focused on new business.
That discipline has a cost. By 2023, Ooley says, DataBank could have filled all available megawatts with so-called neocloud providers—the GPU-as-a-service startups that emerged alongside the AI boom—but deliberately closed that share to save room for business customers to grow. Ooley says: “If we don’t have a place for them to grow with us, they will grow somewhere else.”
