Quiet Breakout: The Insulation Between Your EV’s Battery Cells Is Trying to Become the Anode Too

In 1931, a chemistry professor named Samuel Kistler at the College of the Pacific made a bet with a colleague that he could pull the liquid out of a wet gel without collapsing the solid left behind. He won by inventing supercritical drying. The result was a feathery silica solid he called an aerogel. Nature published it the same year. For most of the 20th century the material stayed a curiosity: NASA wrapped it around the Mars Pathfinder rover and stitched it into spacesuit linings; pipeline operators eventually used it on LNG tanks. Almost nobody else cared.

Last year, one specific chunk of that same 1931 invention sat between every prismatic cell in General Motors’ Ultium battery packs. It was sold by Aspen Aerogels under the name PyroThin, and it brought in $306.8 million in 2024, according to the company’s own filings: 68% of total revenue, and the reason Aspen turned a profit for the first time in its history.

Then, this year, two things happened that have not happened to a 94-year-old material before. First, GM throttled its EV production. Aspen’s Q3 2025 revenue fell to $73.0 million from $117.3 million a year earlier, and the company cancelled a planned $325 million factory in Statesboro, Georgia, walking away from a $670 million conditional commitment from the Department of Energy’s Loan Programs Office. The thermal-barrier business that the entire EV pivot rested on suddenly looked smaller. Second, and almost nobody outside a thin layer of materials chemists noticed: Aspen’s recent USPTO grants stopped being about the wall between cells and started being about the inside of the cells themselves.

That second shift is the story.

The wall

PyroThin is, technically, a silica-based aerogel sheet reinforced with fiber and opacifiers, dialed in for a single job: stop a cell that has gone into thermal runaway from cooking its neighbors. The relevant Aspen patent (US 12,355,050, granted July 8, 2025) specifies the part precisely. Thermal conductivity below 25 mW/m·K. Density under 0.3 g/cm³. Compressibility of 25%–50% at 50 kPa, with a resilience above 50%. That last spec is the one to notice: the sheet has to take repeated cell swelling without permanently deforming, then bounce back. The chemistry section drops in methyl triethoxysilane, tetraethyl orthosilicate, and guanidine hydroxide, which is the recipe for the wet silica sol that gets supercritically dried into the actual mat.

In plain English, the patent describes a material that is almost entirely air, but stiff enough to install on a robot line, fire-rated to survive a cell venting at 600 °C-plus, and compressible enough to absorb a cell that expands a couple of millimeters as it charges. Until recently, nobody outside Aspen and a few Chinese state labs had a manufacturing process that could deliver it at automotive scale.

The academic side caught up fast. Across 357 million OpenAlex-indexed papers, the term cluster “aerogel + thermal runaway + battery” returned exactly one paper in 2020. In 2024 it returned 12. In 2025 it returned 28, and 2026 is already at 13. Independent groups have shown that a 2 mm silica aerogel barrier between nickel-rich NMC cells blocks roughly 50% of transmitted heat and prevents propagation outright. A 1 mm sheet extends the propagation interval from 368 seconds to 1,294 seconds, roughly three and a half times the rescue window. Those numbers are what convinced GM and others to buy.

In September, Chemical & Engineering News noted that aerogel patenting actually overtook aerogel academic publishing in 2024. That is a rare inversion, and it usually means commercialization is now leading the field rather than following it.

The pivot inside the pivot

Here is the second move. While Aspen’s marketing has been about EV pack safety, the company has been quietly granted five US patents in the last four years for using the same supercritical-drying playbook to build the active electrode of future lithium cells:

• US 11,374,213 (June 2022): carbon-aerogel cathode for lithium-sulfur batteries.
• US 11,605,854 (March 2023): polyimide-derived carbon-aerogel cathode for lithium-air batteries. The claim is unusually specific: a binder-free, monolithic carbon foam with a “fibrillar morphology” and pores explicitly sized to accommodate lithium peroxide particles, the literal discharge product of a Li-air cell, forming inside the pore walls.
• US 11,648,521 (May 2023): carbon-aerogel electrode materials, broadly.
• US 12,272,814 (April 2025): nanoporous carbon structures for electrochemical devices.
• US 12,381,221 (August 2025): fibrous carbon aerogels coated with nano-thin silicon as lithium-ion anodes. The method runs a silica coating layer over a carbon-nanofiber skeleton, then reduces the silica to silicon by reacting it with magnesium turnings at 450–900 °C. The final product is silicon, alloyed into carbon, suspended on a porous skeleton. That porosity is the engineering trick: it gives the silicon room to expand as it lithiates without cracking the electrode.

Silicon expands roughly 300% when it absorbs lithium. That is why silicon anodes, theoretically capable of three times the capacity of graphite, have spent two decades failing in production cells. The whole game is finding a scaffold that holds the silicon close enough to the current collector to conduct, but loose enough to breathe. A carbon aerogel, 99% air and with a stiff carbon skeleton, is plausibly that scaffold.

In other words: the same supercritically dried nanopore network that is currently keeping GM’s cells from killing each other when one fails is also being engineered, by the same company, to live inside the cell as the part doing the lithium storage. Both jobs reduce to the same trick: controlling what happens in pores a few tens of nanometers across.

Who else is in the pool

This is not a single-company story. DuPont Safety & Construction has US patents on aramid-fibril papers loaded with aerogel powder, marketed as flame barriers for battery cells (US 11,509,016, US 11,578,461). Panasonic was granted a heat-insulating-sheet patent for battery use in January 2025. Hefei Gotion High-Tech, one of China’s larger battery makers, filed multiple aerogel-based composite-thermal-barrier patents through 2024–2025. Soochow MOFs has a metal-organic-framework-derived carbon-aerogel anode. Ford Global Technologies has a 2024 grant on high-temperature insulating foam for high-voltage battery protection in EVs. The supplier base for this niche has, in five years, grown from one US specialty firm to a dozen.

Who cares

If you are a corporate development scout at a battery OEM, the question on your desk is whether the silicon-anode space is about to consolidate around aerogel-scaffold chemistry. Sila Nanotechnologies, Group14, and Amprius have raised billions on different silicon strategies. None of them is a silica-aerogel company. If aerogel-scaffolded silicon actually cycles, the incumbent silicon-anode makers face a competitor whose factories already exist, and whose chemistry was being invented for a different problem entirely.

If you are an EV safety regulator, the relevant fact is that thermal-runaway propagation barriers have moved from optional to functionally mandatory between cells in U.S. pack designs in the last 36 months. China’s revised GB 38031 battery-safety standard, released in March 2025 and taking effect for new vehicles on July 1, 2026, eliminates the previous five-minute warning window entirely. The new requirement is no fire, no explosion, and no smoke in the cabin during or after a single-cell thermal runaway, for two hours. That is an aerogel-shaped requirement.

If you are a public-markets investor, the lesson is harsher. Aspen’s stock has spent 2025 cratering because a single customer in a single application defined the company’s revenue. The patents that have been granted to it since are a signal that the management knows this and is trying to convert one nanoporous-material franchise into three. Whether the silicon-anode work ships at scale is the question that decides whether Aspen is a fading auto supplier or a battery-materials company hiding inside one.

Either way, a bet that began in 1931 in Stockton, California, over whether a gel’s solid network could survive without its liquid, is, almost a century later, the structural detail that decides whether your EV’s pack survives an off-nominal Thursday. And quite possibly the detail that decides how much charge it holds in the first place.

Method

Patent counts and grant assignments come from a local mirror of USPTO full-text grant data (9.3 million issued utility patents), filtered to grants issued between January 2020 and September 2025. Assignee filings combine variant spellings. Academic counts are from OpenAlex (roughly 357 million scientific works); the “aerogel + thermal runaway + battery” tally is a full-text search of title and abstract for publication years 2020 through 2026 year-to-date. Revenue and customer-concentration figures come from Aspen Aerogels’ 8-K filings; the Q3 2025 number is from the company’s October 2025 earnings release. Independent thermal-barrier performance numbers (50% heat attenuation, propagation delay from 368 to 1,294 seconds) are from peer-reviewed studies in Energy, Process Safety and Environmental Protection, and Journal of Thermal Analysis and Calorimetry published in 2023–2025. Patent claim language was read directly from US 12,355,050, US 11,605,854, and US 12,381,221.