If the harvester's one real advantage is that it never stops, the natural load to hang on it is one that also never stops.
Read this first. This page is a CONCEPT — a thought experiment for the portfolio, not a costed engineering proposal, and nothing here is planned for construction. Every number is illustrative and order-of-magnitude. The point is the logic of the pairing, stated honestly, including where it does not scale.
The Belleville / Second River case argues one thing the storm-only sites cannot: a continuous, behind-the-meter 365-day baseload. A power source that runs every hour of the year invites a load that also runs every hour of the year. Compute is the cleanest such load — a rack of servers draws a flat profile day and night — so the concept is to co-locate a small edge compute node at a continuous HVEH site: power it behind the meter, cool it with the water already flowing past, and let it ride through grid outages on the same islanded microgrid that keeps the pumps running.
Three properties line up. Constant power: mill-mode output is flat, which matches a compute duty cycle far better than solar or storm-peak generation. Water for cooling: a riverside site has ambient water adjacent — attractive for heat rejection, subject to thermal-discharge permitting. Islanded resilience: a node that stays up when the grid is down is worth more than its raw FLOPS — it can host local emergency services, municipal telemetry, or a community edge cache during exactly the storms the HVEH corridor is built for. MODEL
This is where most "river-powered data center" pitches quietly cheat, so state it plainly: a small tributary like the Second River yields on the order of tens of kilowatts continuous, not megawatts. That powers a few racks — an edge micro-node, not a hyperscale facility. A real data hall would need the larger tidal-Passaic sites, a grid supplement, or an aggregation of many modules, and even then the honest framing is "grid-connected with an HVEH baseload contribution," not "off-grid data center." The concept earns its place as a resilient edge node; it does not earn the word "data center" in the industry sense.
| Property | Illustrative figure | Honest caveat |
|---|---|---|
| Continuous power (3 Second River modules) | ~30–60 kW MODEL | Size from USGS-01392500 flow record |
| Compute it supports | ~a few racks / edge micro-node | Not hyperscale; not off-grid at scale |
| Cooling | River-adjacent heat rejection | Requires thermal-discharge permit + ecological study |
| Resilience value | Rides grid outages on the microgrid | The real selling point, not raw capacity |
As a portfolio concept it does one useful job: it turns the abstract "365-day baseload" claim into a concrete picture of what continuous, resilient, behind-the-meter power is actually for. It sits beside the Belleville case as its natural sequel — the demand side of the same continuous-flow argument.