AI data centers are forcing a grid reality check. Who is solving the downtime problem for hospitals and everyone else?

Pew Research put a hard number on what a lot of people are feeling anecdotally: US data centers used about 183 TWh in 2024, a bit over 4 percent of total US electricity use, and the same analysis points to that demand more than doubling by 2030. Grid Strategies went a step further and looked at utility load forecasts and found data centers are roughly 55 percent of forecast demand growth over the next five years. NERC has also been flagging surging load growth and naming data centers as a major driver in multiple reliability assessments.

Even if you think "the grid will adapt," the lag is the problem. New generation, transmission, and substation upgrades do not appear overnight. In the meantime you get tighter operating margins, more congestion, and more local reliability and power quality headaches. For normal households that is annoying. For mission critical facilities it is existential.

Hospitals and long term care are the obvious stress test. They cannot tolerate downtime, and the real pain is not only multi hour blackouts. It is sags, brief interruptions, and messy transfers that can trigger alarms, force resets, or put staff into manual mode during the worst moments. Backup generators solve duration, but the facility still needs fast continuity and clean power. That is why resilience is turning into a separate spend category, not just an insurance policy.

This is where I think microgrids stop being a "green project" and start being a reliability product. A properly designed microgrid can island, stabilize local power, and keep critical loads running even when the utility side is unstable. The reason I am posting this in an investing sub is that we may be entering a multi year capex cycle where customers pay for uptime directly rather than trusting the public grid to guarantee it.

One example I have been tracking is NextNRG (NXXT). They are not a pure play utility, but they sit on two angles that map to this problem:

1. Facilities resilience - they announced 28 year microgrid PPAs in healthcare settings, which is basically "no upfront capex, long duration service contract" for solar plus storage plus controls.

2. Keep the lights on logistics - their EzFill mobile fueling business is about getting fuel to fleets and equipment. In outage scenarios, the unsexy bottleneck is often fuel logistics for generators. If critical facilities start thinking in terms of "resilience supply chain" rather than "one generator," that kind of service matters.

They also recently disclosed preliminary operational numbers for December 2025: about $8.01M revenue, up 253 percent year over year, and about 2.53M gallons delivered, up 308 percent year over year. Separately they signed an MOU with A123 that explicitly references 20 foot, 5 MWh battery storage units, which is the kind of format that fits the "facility microgrid" use case.

None of this is a free lunch. NXXT also disclosed lender litigation around an alleged default, which is a real risk if you are thinking about financing, dilution, or project execution. But stepping back, the bigger question is whether the market is underpricing the demand shift: AI load growth -> tighter grid conditions -> customers paying for on site resilience.

From an investing lens, do you think resilience spend becomes "non discretionary capex" the same way cybersecurity did, or does the grid catch up fast enough that this stays niche?