In June 2009, a drill rig on the Krafla caldera in northeast Iceland plunged its bit 2.1 kilometers into the ground and hit something the crew was not expecting: a pocket of molten rock, about 900 °C, oozing up from below. The drill string seized and snapped. The engineers did what engineers do when their equipment is stuck inside a volcano. They cemented a perforated steel casing into place, backed away, and let the well heat up.

For the next two years, that hole funneled superheated steam to the surface at over 450 °C, the hottest geothermal well ever measured. A ruptured valve eventually forced the team to plug it. If the valve had held, IDDP researchers calculated the single borehole would have generated 36 megawatts, roughly ten times the output of a conventional geothermal well of the same depth.

Then, for more than a decade, nobody went back.

Look at the US patent database now and you will see that the long quiet is over. Twenty-one US patents with the word magma in the title have been granted since January 2023. In the 23 years before that, the total was thirteen, and most of them were about house plants, lava lamps, and stator windings with brand names like “Magmate.” The modern pile is about one thing: running drill rigs, casings, and chemical reactors inside active magma reservoirs.

Sixteen of the twenty-one new grants belong to a single Florida-registered entity called EnhancedGEO Holdings, LLC, with a handful of the most recent filings not yet assigned in the public record but sharing the same titles and the same claim structures. An affiliate, Magma Power LLC, issues the press releases. Between them they have built a wall of intellectual property around a specific physical act: putting a pipe into liquid rock and getting useful heat back out.

The trick is quenching the rock before it eats the drill

Normally, a drill bit meeting magma is a short story. Steel softens around 1,000 °C, and drilling mud boils long before it gets there. The EnhancedGEO patents approach the problem by changing the rock instead of the drill. The 2025 grant Detecting entry into and drilling through a magma reservoir (US 12,291,965) describes a controller that, the moment the borehole instruments detect molten material at the bit, switches the rig into a “magma-drilling mode”: flow rate on the drilling fluid jumps, drilling speed drops, and the bit begins moving in a reciprocating motion. The fluid floods the interface and thermally shocks the magma into a solid glass plug, essentially a skin of vitrified obsidian, which the bit can then chew through like any other rock.

Once the borehole is extended into the magma body, a related patent (US 12,297,711, Casing a wellbore in magma) covers the trick of lowering a sealed steel casing into the quenched fluid column so it sinks gently onto the target depth without contacting the solidified glass. A third (US 12,305,486) specifies the follow-up: a fluid conduit inside the casing circulates a working fluid between the surface and the heat-exchange zone at the bottom, with a pressure generator maintaining at least 1,000 PSI of back pressure to keep the fluid in a supercritical state. Steam comes up. Electricity comes out.

Lab-scale confirmation arrived on June 7, 2025, when Jeffrey Karson’s Syracuse Lava Project, which since 2009 has specialized in melting crates of basalt in industrial furnaces to study how lava flows in the wild, co-ran a test with Magma Power. Together they drilled a 17-inch borehole, three inches in diameter, into 1,200 °C synthetic lava. The casing held. The plug-quench-drill sequence worked. A university lab, for the first time, had the equipment and the procedure to bore through molten rock on purpose.

The real prize is chemistry, not kilowatt-hours

If the story stopped at superhot geothermal electricity, that would already be a big deal: a typical supercritical well can carry five to ten times the thermal energy per kilogram that conventional geothermal fluid does, which collapses the pipe count and the surface footprint. Magma Power publicly targets 1 cent per kilowatt-hour. Take that with skepticism — it comes from the vendor, not an independent LCOE — but superhot rock does inherently rearrange the cost stack.

The more interesting claim sits in a cluster of thermochemistry patents the same entity has been quietly assembling. System and method for magma-driven thermochemical processes (US 11,918,967, issued March 2024) and Molten-salt mediated thermochemical reactions using geothermal energy (US 12,319,573, issued June 2025) describe dropping a reaction chamber down the wellbore and running known high-temperature chemistry inside it, using the magma itself as the heat source.

The cycles named in the claims are familiar to anyone who read the nuclear-hydrogen literature of the 2000s. The sulfur-iodine cycle: iodine plus sulfur dioxide plus water in, hydrogen and oxygen out, needing about 850 °C to close. The copper-chloride cycle: copper and hydrochloric acid feeds, closing around 500 °C. Both were designed for fourth-generation nuclear reactors, but the reactors never arrived at commercial scale. The EnhancedGEO patents propose running those same cycles downhole, with a magma reservoir standing in for the reactor core. Other claims describe reacting the resulting hydrogen with carbon dioxide to produce liquid and gas hydrocarbons — synthetic fuels, made with geological heat and captured CO₂.

This is the quiet part. The press releases talk about 10-gigawatt AI data centers. The patents talk about converting a caldera into a hydrogen rig.

Iceland is about to find out who’s right

None of this is settled physics at field scale. No one has yet sustained a commercial wellbore in a magma body. Which is why the scientific counterpart matters: the Krafla Magma Testbed, a Landsvirkjun-backed nonprofit, has funding for two wells and plans to drill KMT-1 in 2026 into the same reservoir the 2009 accident hit — except this time on purpose, with instrumented casing, and the specific goal of placing temperature and pressure sensors directly in the melt. KMT-2 follows in 2028 for energy extraction experiments. If the well survives a full test cycle, the physics that EnhancedGEO’s patent wall presupposes becomes, for the first time, something independent researchers can measure.

Meanwhile, a different adjacent-possible is coming from the opposite direction. Quaise Energy, an MIT spinout, holds a small US patent family (the most recent, US 12,326,085, Basement rock hybrid drilling) built on using gyrotron-generated millimeter-wave beams to vaporize granite. The logic is to dodge magma entirely by getting 20 kilometers down into dry superhot basement rock. Magma and millimeter waves are chasing the same prize from opposite edges of the problem. The oil-patch drilling industry, which already owns the world’s directional-drilling and high-temperature casing expertise, is the shared tailwind.

For an R&D director in energy, the signal is that superhot geothermal is no longer a Reykjavík-only research curiosity. Someone in Florida has spent four years assembling a patent estate around it. A Syracuse geology professor just proved the drill sequence works on a furnace-scale analog. Iceland is about to put a steel pipe through the accidental well’s successor. The technology still has to prove it works for more than two years at a time without a valve failure. But it now has paperwork, prior art, and a test schedule.

Gas turbines, for most of the last century, have been limited by how hot the first-stage blades can tolerate before they creep. That ceiling is around 1,600 °C and it took a hundred years of metallurgy to get there. A magma well asks the opposite question: how do you handle a heat source that is already hotter than you can use, and stable on a scale of centuries? The patents being granted in 2025 are the first serious attempt at an answer.

Method. US grant counts come from USPTO-sourced bulk grant data current through April 14, 2026; we filtered on the word “magma” in the title and excluded filings related to plant varieties, trademarks, and unrelated acronyms. Assignee counts combine variant spellings (e.g., “EnhancedGEO Holdings, LLC” and “ENHANCEDGEO HOLDINGS, LLC”). Claim language for the patents discussed here was read directly from published USPTO full text. Scientific background on the 2009 Krafla event draws on the peer-reviewed Iceland Deep Drilling Project literature (Elders et al., Geothermics, 2014) and subsequent work on IDDP-2 at Reykjanes (Friðleifsson et al., Scientific Drilling, 2017). Syracuse University lab-scale drilling details are from Magma Power LLC’s June 2025 press release and Jeffrey Karson’s Lava Project at Syracuse. KMT drilling dates and well specifications are as reported by ThinkGeoEnergy and Landsvirkjun. Cost claims attributed to Magma Power LLC are the company’s own and have not been independently verified.