Colorado River | Basin 2.0, Measured | Steps Ventures
Steps Ventures Research · Independent analysis · Updated August 2026

Basin 2.0, measured.

The accounting is the crisis, and the water is measurable. Every number here is computed from public data, and the code is published.

What the measurement shows

The river carries far less than it was promised, and the fights over who gives up what are fights over numbers.

-53%Lake Powell water surface, 19 August 2026 against the same day in 2019. Reclamation daily record, now 56,225 acres, 35% of full pool
1.55 MAF/yrEvaporation and system losses, charged to no state’s allocation
37.6 MAFTotal stored water gone since 2002, including the groundwater no meter sees

MAF is a million acre-feet. One acre-foot is about 326,000 gallons, roughly what two or three households use in a year.

Conditions today. Lake Powell is at 3,519.36 ft and Lake Mead at 1,039.55 ft, as of 19 August 2026 (U.S. Bureau of Reclamation). Both set new post-fill record lows this month and both are still falling. Powell sits 29.4 ft above minimum power pool. Both water surfaces are the smallest in Reclamation's daily record. Two independent Sentinel-2 derivations, refreshed 19 August 2026, put Powell's loss at 49% and 52%. They bracket the daily record closely enough to corroborate it and not precisely enough to replace it, which is why the number above is Reclamation's.

Those are the disputed quantities, and they are already measurable. Agreed measurement does not decide who absorbs the cut. It is the cheapest lever nobody is pulling, and it makes every other fight tractable. Every figure below links to its primary source, and the data and code are on GitHub.

Where to go

Four ways in

This is an independent, open analysis of the Colorado River basin. Everything is computed from public data and the code is published. If you want the whole argument in one piece, read the full case, about an hour.

The one move

Big Tech is about to drop $100 billion into data centers and chip fabs across the desert Southwest. Everyone’s furious about it.

Flip it. Use that same buildout to refill the Colorado River.

The dams are losing power as the reservoirs fall, and the desert Southwest is short of grid. The hyperscalers are already building their own solar and storage to run the data centers, so site that build to do triple duty. Power for the data centers. Replacement for the fading dams. And enough surplus daytime solar left over to make water. Reuse, desalination, and recharge are flexible loads. They run when the sun is up and idle when it isn’t, so they soak up intermittent power without paying to store or firm it.

Not charity. Not offsets. A build whose biggest tier pays for itself, whose public-goods tier is priced honestly, and that makes the whole basin measurably better, acre-foot by acre-foot, in public. The companies blamed for draining the desert become the ones who refilled the river.

That’s the move. Who’s in?

Read the $100B proposal → See the evidence