AI’s voracious appetite is straining energy grids, prompting new strategies and longer planning horizons
In sum — what to know:
A gating factor: Power is now the biggest gating factor for AI infrastructure, with water to follow next.
Bring your own power: To avoid grid struggles, business must consider adopting bring your own power (BYOP) strategies which involve tapping into multiple energy sources, years in advance.
Nuclear power, a distant solution: CEO of DigitalBridge notes that nuclear power as an alternative energy source for AI remains years away.
AI infrastructure is at a crossroads. While on one hand demand for infrastructure is growing at a manic speed, power constraint is choking structural expansion plans.
That’s the message that came from Marc Ganzi, CEO of DigitalBridge — an asset management firm in the digital infrastructure space which was acquired by SoftBank late last year — at Metro Connect USA in Fort Lauderdale.
Speaking at the keynote, Ganzi said, “To enable AI, we all know the cliche is, power availability becomes the currency. That’s really the currency that’s driving critical workloads, that’s driving new data centers, and ultimately drives an ecosystem that feeds off of those data centers.
While there is a lot of free land, he reminded that land without power is of no value, nor a “will-serve” letter from an utility provider stating that it has capacity and intent to supply service.
“A will-cert letter does not mean you have a connection date,” he said, noting that developers today are looking at connection dates between 2030 and 2032.
He said that gaining access to energy is now a power play which really comes down to how politically connected one is in that state.
He argued that it essentially makes the digital infrastructure business more like power enablement business, where, to sidestep state-level politics and long grid queues, developers must orchestrate grid-independent power, micro-grids and multi-source energy strategies as a core part of their business model. Now popularly called bring your own power (BYOP), the strategy promotes onsite power generation over reliance on public grids.
Ganzi reminded that the power problem is here to stay. The industry currently leases 12 gigawatts of power every year, while adding four gigawatts of incremental capacity.
“You guys can do the math,” he said. “We have a deficit of power,” adding that the gap will only widen in the coming years.
And as the energy discourse turns towards nuclear power as an alternative source that can meet AI’s appetite for “baseload” power, Ganzi’s message was blunt but clear: “We don’t believe that magically 40 to 60 gigawatts of nuclear power is going to show up in eight years. It’s not.” He added that it will take another decade and then some years for it to pan out practically.
“If we’re all hoping for some magic bullet to save the power problem, I’m here to tell you it’s not going to happen,” he emphasizes.
With the data center industry already facing public backlash from local communities for building projects in their backyards, Ganzi predicted that the next challenge will be water.
“The next page to turn is the sector is going to be hit hard with water. We’ll have to defend our water usage.” He noted that it is pressing on the industry to design water cooling systems that require no loss of water, adding that work is already underway.
Ganzi urged the industry to think in longer timelines and planning horizons. He recommended a 36 to 48 month build cycle instead of a five-year model, and advised companies to start planning to solve tomorrow’s power supply issues today with several years of time on hand.
“That’s the mentality you have to have for success…The opportunity is huge. The challenges are different. Every year the chessboard changes, and it’s a lot of fun,” he concluded.
Facts Only
Marc Ganzi, CEO of DigitalBridge, spoke at Metro Connect USA in Fort Lauderdale.
DigitalBridge was acquired by SoftBank late last year.
Power availability is the biggest constraint for AI infrastructure expansion.
Developers are facing grid connection delays, with some dates pushed to 2030-2032.
"Bring your own power" (BYOP) strategies involve onsite power generation and micro-grids.
The industry leases 12 gigawatts of power annually but only adds 4 gigawatts of new capacity.
Nuclear power is not expected to provide a significant solution within the next decade.
Water usage for cooling systems is identified as the next major challenge for the industry.
Ganzi recommends a 36-48 month build cycle instead of a five-year model.
The AI infrastructure sector is shifting toward longer planning horizons to address power supply issues.
Local communities are resisting data center projects due to environmental concerns.
Zero-loss water cooling systems are under development to mitigate water usage.
Executive Summary
Full Take
The narrative presents a compelling case for the energy crisis in AI infrastructure, but it’s worth examining the underlying assumptions and potential biases. Ganzi’s emphasis on power as the "currency" of AI development is a strong framing device, but it risks oversimplifying a complex issue. The call for BYOP strategies and micro-grids is pragmatic, yet it assumes that decentralized energy solutions can scale quickly enough to meet demand—a claim that warrants scrutiny. The dismissal of nuclear power as a near-term solution is blunt but aligns with industry skepticism about its feasibility. However, this could also reflect a broader pattern of underestimating technological breakthroughs or policy shifts that might accelerate nuclear adoption.
The focus on water as the next challenge is a valid concern, but the narrative leans heavily on industry perspectives without sufficient counterbalance from environmental or community stakeholders. This could inadvertently downplay the severity of local opposition to data centers, which often stems from legitimate concerns about resource depletion and environmental impact. The recommendation for longer planning horizons is sensible, but it raises questions about whether the industry is prepared to make the necessary investments in sustainability and innovation.
Root cause: The narrative reflects a paradigm where technological growth is assumed to be inevitable, and the primary constraint is resource allocation rather than fundamental limits. This echoes historical patterns of industrial expansion, where infrastructure struggles to keep pace with demand, leading to stopgap solutions that may not address systemic issues.
Implications: The push for BYOP and micro-grids could decentralize energy production, potentially empowering local communities but also shifting costs and responsibilities onto private entities. The dismissal of nuclear power may delay broader energy policy discussions, while the focus on water usage highlights the need for more sustainable cooling technologies. Second-order consequences could include increased competition for energy resources, higher costs for AI development, and potential regulatory backlash if environmental concerns are not adequately addressed.
Bridge questions: How might the industry balance the need for rapid expansion with long-term sustainability? What role should public policy play in addressing the energy and water challenges of AI infrastructure? Could nuclear power or other alternative energy sources see faster adoption if given more political and financial support?
Counterstrike scan: If this narrative were part of a coordinated influence campaign, the playbook might involve amplifying the urgency of the energy crisis to justify deregulation or privatization of energy infrastructure. However, the content does not exhibit clear signs of manipulation, as it presents a balanced view of industry challenges without overtly pushing a specific agenda. The focus on practical solutions and acknowledgment of limitations suggests a genuine attempt to address real issues rather than a coordinated push for a particular outcome.
Patterns detected: none
Sentinel — Human
The article shows strong signs of human authorship, with a distinct voice, specific attributions, and idiosyncratic phrasing. Minimal stylometric or coordination red flags suggest it is likely a human-written piece, possibly with light editorial structuring.