The signal

In October 2025, the Nobel Prize in Chemistry went to Omar Yaghi, Susumu Kitagawa, and Richard Robson for metal-organic frameworks: crystalline sponges so porous that a gram of the stuff can have the internal surface area of a football field. The press coverage filed the story under chemistry. It was about elegant lattices, gas storage, carbon capture, molecular Lego.

The US patent office files it under something else. When you count granted patents that describe pulling drinking water out of thin air, the curve looks like a startup’s pitch deck. One or two a year through the 2000s. Five in 2017. Nineteen in 2024. Nineteen more in 2025. And when you read what is actually inside the newest ones, the invention that keeps showing up is not a better refrigerator. It is a gas sponge, repurposed as a water utility.

Two ways to squeeze water from air

There have always been two physics for getting water out of the atmosphere, and they are not close cousins.

The old way is condensation. Cool a surface below the air’s dew point and moisture beads up on it, the way it does on a cold glass. This is what a dehumidifier does, and it works fine in Miami. It fails in Phoenix. In genuinely dry air, the dew point sits below freezing, so you are spending enormous energy chilling an entire desert’s worth of atmosphere to wring out a few drops. The machines that do it are the size of shipping containers and drink electricity.

The new way is sorption. Instead of cooling the air, you let a material grab water molecules directly out of it, one at a time, and hold them. Then you warm the material slightly, often with nothing but sunlight, and it releases the water as a small, concentrated, humid puff that condenses easily. You never chill the desert. You chill a thimble. The whole trick lives in finding a material that will bind water tightly at 15 or 20 percent humidity and let go of it again when you want it back.

That material is a metal-organic framework. And the reason 2025 was the year the patents piled up is that the chemistry finally got good enough, and cheap enough, to build a business on.

What the newest patents actually claim

Read past the titles and the engineering is unusually consistent, which is the tell that this is a real technology cluster and not a keyword coincidence.

A grant issued in September 2024 to Water Harvesting, Inc. describes a system with a bank of separate modules, each packed with a different metal-organic framework tuned to a different “adsorption threshold humidity.” A controller decides, in real time, which modules to run based on the humidity outside, the energy cost of squeezing each one dry, and how much water is already in the tank. It is a software-orchestrated array of sponges, each optimized for a different slice of the day. Morning fog and afternoon bake get handled by different chemistry inside the same box.

Others are chasing the same idea through different materials science. One 2024 grant describes an aerogel, itself one of the lightest solids ever made, cross-linked with a metal-organic framework so the composite can cycle water continuously without the powder clumping or washing out. Another describes weaving the framework into a nanofiber mat, a discontinuous phase of active crystal dispersed through a polymer that holds it in place. A third stages a “subcooler” heat exchanger specifically to recover the heat penalty of driving water back off the sorbent, because the energy accounting is where these systems live or die.

Strip the phrase “atmospheric water” from all of them and they still describe the same machine: a cyclic sorbent that adsorbs at low humidity, desorbs with modest heat, and condenses a small concentrated stream. That shared spine is the difference between a technology and a buzzword.

The lab bench to the desert

The trajectory was visible in the literature years before the Nobel committee noticed. In 2017, a team from Yaghi’s Berkeley lab and Evelyn Wang’s group at MIT ran the first field demonstration, powered only by sunlight, and reported in Science that it worked. Berkeley News wrote it up. By 2018 the group was running trials in the Arizona desert; later prototypes were tested in Death Valley. A 2019 paper in ACS Central Science, “Rapid Cycling and Exceptional Yield in a Metal-Organic Framework Water Harvester,” has since been cited more than 560 times, the kind of number that marks a paper the whole field builds on.

The company Yaghi co-founded to commercialize it, Water Harvesting, Inc., now operating as WaHa, holds 13 US patents. In 2025 it closed an $8 million Series A-1 with a revealing cap table: the Berkeley Catalyst Fund, the Swedish climate fund Vestafund, and Mitsui Mining & Smelting, a 150-year-old Japanese metals company. The metals firm is the interesting name. Making metal-organic frameworks at scale is a wet-chemistry manufacturing problem, and the companies that already know how to precipitate metal salts by the ton have a head start no water startup can buy. WaHa says its field units, running in Abu Dhabi, Riyadh, West Texas, and Stockholm, have hit 99.998 percent mechanical reliability. A second Yaghi company, Atoco, has publicly targeted machines that pull 1,000 liters a day from air at 20 percent humidity or lower, at under 5 kilowatt-hours per liter.

Who cares? Roughly two billion people live where the air is too dry for a conventional condenser to be worth running. The sorbent approach is the first version of this technology that addresses that market instead of the humid one that already has cheap water. That is the whole commercial thesis in one sentence.

The tell: even the incumbents are hedging

Here is the detail that keeps this honest. WaHa’s own portfolio is not pure sorbent. Alongside the metal-organic framework patents sit grants for “heat pump-based water harvesting” and a refrigerator with an integrated water unit. Even the flagship module patent includes a condensation chamber downstream of the sponge. Sorption does not abolish refrigeration; it shrinks it, concentrating the water first so you only have to condense a small humid stream rather than chill the open sky.

And the field around WaHa is bifurcating in public. The largest cluster of atmospheric water patents after WaHa belongs to companies that mostly are not doing sorption at all. Genesis Systems, which holds several recent grants, spent decades as a robotic welding and factory-automation shop, filing patents on welding torch cleaners and robotic workcell doors, before it pivoted to building container-scale water machines for defense customers. Israel’s Watergen and the military-focused Exaeris each hold four. AWN Nanotech takes a third path entirely, a passive wall with a hydrophilic face and a hydrophobic back that wicks water through pores with no sorbent and no compressor.

So the patent record is telling two stories at once. One is a category growing fast enough to pull in a welding-robot company and a Japanese smelter. The other, underneath it, is a quiet materials swap: the fastest-growing slice of the fastest-growing category runs on a material that was invented to store hydrogen and ended up storing water. The Nobel committee gave the prize to the chemistry. The market is quietly deciding the application.

Method note

Counts come from roughly 9.3 million US utility patent grants sourced from USPTO bulk grant data, searched for titles describing atmospheric water generation or water harvesting from air, from January 2000 through mid-June 2026 (2026 is a partial year). Assignee tallies combine variant spellings for each company and reflect only patents whose assignee is recorded; some very recent 2026 grants have no assignee attached yet, so leaderboard counts are conservative. The sorbent-specific subset was identified by abstracts referencing metal-organic frameworks, desiccants, or hygroscopic sorbents in a harvesting context, and grew from 2 such grants in 2020 to 7 in 2025. Literature figures come from OpenAlex (357 million works); the citation count for the 2019 ACS Central Science paper is as recorded there. Company funding, field-trial locations, and device specifications are drawn from company statements and reporting by Berkeley News, Tom’s Hardware, and trade press, cited inline; treat vendor performance figures as manufacturer claims, not independent measurements.