On the Frontier of Modular Edge AI Data Centers with Andrew Lindsey, CEO, Flexnode

The modular data center industry is undergoing a seismic shift in the age of AI, and few are as deeply embedded in this transformation as Andrew Lindsey, Co-Founder and CEO of Flexnode. In a recent episode of the Data Center Frontier Show podcast, Lindsey joined DCF Editor-in-Chief Matt Vincent and Senior Editor David Chernicoff to discuss the evolution of modular and edge data centers, the growing demand for high-density liquid-cooled solutions, and the industry factors driving this momentum.

Lindsey’s career has been defined by the intersection of technology and the built environment. Prior to launching Flexnode, he worked at Alpha Corporation, a top 100 engineering and construction management firm founded by his father in 1979. His early career involved spearheading technology adoption within the firm, with a focus on high-security infrastructure for both government and private clients.

Recognizing a massive opportunity in the data center space, Lindsey saw a need for an innovative approach to infrastructure deployment. "The construction industry is relatively uninnovative," he explained, citing a McKinsey study that ranked construction as the second least-digitized industry—just above fishing and wildlife, which remains deliberately undigitized. Given the billions of square feet of data center infrastructure required in a relatively short timeframe, Lindsey set out to streamline and modernize the process.

Founded four years ago, Flexnode delivers modular data centers with a fully integrated approach, handling everything from site selection to design, engineering, manufacturing, deployment, operations, and even end-of-life decommissioning. Their core mission is to provide an "easy button" for high-density computing solutions, including cloud and dedicated GPU infrastructure, allowing faster and more efficient deployment of modular data centers.

As Vincent noted, Data Center Frontier has closely tracked the increasing traction of modular infrastructure. Lindsey has been at the forefront of this shift, witnessing the market evolve significantly over the last five years.

"Five years ago, we were looking at a graveyard of modular data center companies that leaned heavily on the edge," Lindsey recalled. Many early modular providers focused on latency-sensitive, interconnected solutions—such as base stations at 5G tower sites. However, the market proved premature, hindered by high costs and the scale of deployment within the telecommunications industry.

Now, macroeconomic and technological factors have driven a fundamental shift toward modular data centers. One of the most significant drivers is the rapid evolution of chip design. "A traditional data center design cycle can take a year or 18 months," Lindsey explained. "But if we see radical Nvidia chip advancements every 12 months, your design could be obsolete before you even break ground." The need for embedded flexibility within data center design has made modular solutions an ideal fit.

Another factor accelerating the adoption of modular infrastructure is the labor shortage in construction. "There just aren’t enough people today to build the scale of infrastructure needed for data centers," Lindsey noted. Compounding the issue is an aging workforce, with many skilled professionals nearing retirement. "When they leave, they take decades of institutional knowledge with them."

Modular construction mitigates this problem by shifting labor-intensive processes to manufacturing environments where technical expertise is concentrated. By centralizing production, modular providers can reduce reliance on dispersed construction labor while maintaining high precision and efficiency.

Flexnode is also a leader in the adoption of high-density liquid-cooled infrastructure. Lindsey attended the CoolerChips event last year and has been vocal about the advantages of liquid cooling for modern workloads. "More recently, modular is everywhere," he said. "We at Flexnode are seeing demand hand over fist for high-density liquid-cooled systems that integrate seamlessly with broader building designs."

This demand underscores the shift from the speculative modular edge deployments of five years ago to today’s high-performance, flexible data center solutions. "Modular is no longer just a niche," Lindsey concluded. "It’s a critical strategy for meeting the growing demand for scalable, high-efficiency data center capacity."

The realization that liquid cooling would become a building-wide challenge, rather than just an IT challenge, was a pivotal moment for Flexnode. "Four years ago, we recognized that liquid cooling, which had been around for 10 to 15 years in government and research, was now commercially viable. But very few data centers were truly equipped to deploy it to its full potential," Lindsey explained. Flexnode identified an opportunity to deliver integrated liquid-cooled modules that connect IT systems to building infrastructure through a fully embedded design.

Rather than developing proprietary liquid cooling technology, Flexnode focuses on being "liquid neutral." "The liquid cooling market is advancing well on its own," Lindsey said. "We want to enable OEM-driven solutions like JetCool, Motivair, Isotope, and ZutaCore, ensuring they perform optimally in an environment designed to support them."

Flexnode operates at the building scale, working on innovative heat management strategies that eliminate the need for external heat rejection. "We integrate heat rejection into the panelized construction of our modular data centers," Lindsey explained. This approach pushes forward a broader, integrated building design suited for liquid cooling.

David Chernicoff asked Lindsey whether Flexnode leans toward specific liquid cooling methodologies, such as waterless, multi-phase, or single-phase solutions. Lindsey responded that their focus aligns with OEM and ODM preferences.

"Right now, we're primarily working with direct-to-chip water-based single-phase cooling," Lindsey said. "But as part of our work with the Cooler Chips program, we’re also developing a hybrid immersion approach with Isotope." This hybrid method integrates both direct-to-chip and immersion cooling.

The industry is currently debating whether to move to a single-phase hybrid approach or leapfrog directly to two-phase cooling. "The big challenge with two-phase is the environmental impact of certain chemicals used in the process," Lindsey noted. While companies are actively working to address these concerns, two-phase cooling remains a complex consideration.

Even Nvidia is leaning toward a two-phase future. "From what we've heard at CoolerChips, Nvidia sees the next generation as being two-phase oriented," Lindsey said. "But they can speak better to that."

With liquid cooling now firmly part of the mainstream conversation, the challenge is not just about advancing the technology but also ensuring that the surrounding infrastructure evolves to support it. Flexnode’s approach—integrating liquid cooling at the building level—positions them at the leading edge of this shift.

As the industry evolves, cooling technology decisions are increasingly shaped by customer preferences. "Right now, it's very much customer-driven for us," Lindsey explained. "We're working with sophisticated customers—hyperscalers and GPU-as-a-service providers—who already know what they want to deploy."

While some enterprises may still be evaluating their liquid cooling options, hyperscalers are looking beyond traditional single-phase approaches, including both dielectric and water-based cooling. However, Lindsey emphasized that many of these developments remain in the R&D phase. "We don’t typically recommend one technology over another unless there’s a clear drawback," he said.

One challenge with direct-to-chip cooling, for example, is achieving full heat absorption into the liquid. "That’s where hybrid approaches come in," Lindsey noted. He described hybrid designs that integrate both two-phase direct-to-chip cooling and immersion cooling, as seen in the CoolerChips program. "In some cases, direct-to-chip is single-phase, in others, it’s two-phase. We’re working as a category B provider, helping integrate these technologies at the building level."

Lindsey also touched on sustainability concerns, particularly around immersion cooling. "Immersion is seen as the most sustainable in terms of energy efficiency, but there are still questions about how immersion fluids impact server longevity over time," he said. Factors like glue degradation and cable insulation breakdown raise questions about immersion cooling’s long-term sustainability profile.

Two-phase cooling also presents challenges. "There’s an ongoing discussion about PFAS and finding non-toxic, non-carcinogenic alternatives," Lindsey explained. "Beyond that, two-phase cooling can create cavitational forces that affect motherboard and chip integrity over time. That’s why many in the industry—including Nvidia—are still weighing the trade-offs."

With liquid cooling now firmly in the mainstream, the industry’s next challenge is integrating these technologies seamlessly into modular data centers. "It’s not just about cooling IT gear anymore; it’s about designing buildings that fully support liquid cooling at scale," Lindsey concluded. Flexnode’s modular approach positions them at the forefront of this transformation.

Finally, Flexnode’s modular approach offers extreme configurability. "Our modules can be standalone data centers or integrated into powered shell facilities," Lindsey explained. "We configure everything from 2 MW to 20 MW standalone deployments, and we can scale up to 200 MW campuses."

Beyond footprint flexibility, power integration is a growing focus. "On-prem generation is gaining traction, particularly with fuel-agnostic generators that can switch between natural gas, hydrogen, methane, and propane," Lindsey noted. Collaborating with partners like Hyliion, Flexnode is exploring adaptable power solutions, including fuel cells.

Being behind the meter is another key driver. "Utilities are getting smarter about power allocation," Lindsey said. "In Europe, data centers are facing use-it-or-lose-it policies, and in the U.S., regions like Ohio are imposing tariffs on unused capacity." On-site power generation provides greater flexibility, helping data centers scale more efficiently and participate in curtailment programs that balance grid demand.

As modular data centers become a core part of the industry landscape, Flexnode is pushing the boundaries of what’s possible. "We see modular as a natural extension of utilities—a distributed solution that enhances flexibility," Lindsey concluded. "And we’re just getting started."