Antora’s $550 million round landed in a patent field 54,623 families deep and about fifty years old. The only part of it still bending is thermal storage tied to data centers.
New money does not make a field new. Antora’s $550 million round landed in a patent landscape with 54,623 families in it, and only one corner of that landscape is still open.
Antora Energy closed a $550 million Series C at a $2.47 billion valuation on July 30, co-led by G2 Venture Partners and Eclipse. The product stores heat in blocks of solid carbon and gives it back as industrial process heat or electricity. It is a good company solving a real problem, and the round is a fair read of where industrial energy demand is going.
It also landed in a patent field with 54,623 extended families in it, filing continuously since the 1970s. That is the part worth sitting with. When new money arrives in an old field, the question is not whether the technology works. It is which specific claims are still available, and who already owns everything else.
Thermal and heat storage covers 54,623 extended patent families in Minesoft Origin as of August 4, 2026. Annual output went from 1,154 families in 2011 to 3,793 in 2025, with a local peak of 4,113 in 2023. That is roughly 3.3x over fourteen years.
Set that against the fields we have mapped recently. AI drug discovery grew about tenfold in a decade. Healthcare AI and medical imaging grew far more sharply than that. Those curves bend hard because the underlying capability did not exist before. Thermal storage does not bend, because storing heat in a hot solid is an idea with a very long history and a large installed base of prior art behind it.
This shape has a direct consequence for anyone filing here. In a bending field, you can be early and broad. In a flat field, breadth is already taken, and the value sits in narrow claims tied to a specific configuration, material system, or control method that the incumbents never needed.
Ranked by extended families, the leaders are industrial: Panasonic at 1,111, Denso at 603, Xian Thermal Power Research Institute at 593, Hitachi at 480, Mitsubishi Electric at 441, Xian Jiaotong University at 395, the Chinese Academy of Sciences at 381, Midea at 375, Toyota Central R&D at 362, Toshiba at 359. Siemens AG holds 326 and Siemens Energy 298. The German Aerospace Center holds 336, France’s CEA 318, BASF 257.
Notice the mix. Automotive thermal management, HVAC, power generation, materials chemistry, national labs. Most of these portfolios were not built for grid storage at all. They were built for cabin heating, waste heat recovery, and district energy, and they now sit across the path of anyone commercializing heat storage at scale.
The largest single bucket is not a corporation. 5,983 families are held by individual inventors, more than five times Panasonic’s total. In most technology landscapes that share is small. Here it reflects decades of independent work on solar thermal, water heaters, and building storage, and it means the prior art is scattered across thousands of unaffiliated filings rather than concentrated in a handful of portfolios you could review in a week.
The funded companies hold small, focused positions by comparison. Malta Inc has 337 families. Rondo Energy has 191. Antora Energy has 27. Electrified Thermal Solutions has 23. Fourth Power has 19.
Count is not quality, and this is the point where most patent commentary goes wrong. Twenty-seven families aimed precisely at a defensible configuration can be worth more than a thousand families of incremental heat exchanger geometry. What the count does tell you is what the money is buying. At a $2.47 billion valuation, investors are underwriting cost per kilowatt hour delivered, speed of deployment, and customer contracts. They are not underwriting patent depth, because the depth in this field belongs to Osaka, Nagoya, Xi’an and Munich.
Break the landscape into its main technical approaches and the crowding becomes concrete. Phase change materials account for 8,396 families. Solid storage media including graphite, carbon blocks, firebrick and concrete account for 5,091. Molten salt accounts for 3,685. Storage tied explicitly to industrial process heat and steam accounts for 1,264.
Thermophotovoltaic conversion, the piece that turns stored heat back into electricity efficiently and the part of Antora’s stack that is genuinely hard, sits at 1,930 families across the whole database, growing slowly from 69 in 2015 to 145 in 2025. That is a small, specialized, physics-heavy space where a handful of strong claims could matter a great deal. It is also a space where national labs and universities have been publishing for thirty years.
So the honest read on materials is this. If your pitch is that you store heat in a cheap solid, the novelty conversation with an examiner will be difficult, because someone stored heat in a cheap solid a long time ago and wrote it down. The claims that survive in a field this mature tend to be about the system: how you charge it from an intermittent grid, how you control discharge temperature, how you integrate with a specific industrial load.
There is exactly one part of this landscape that behaves like a young field.
Thermal storage tied to data centers accounts for 1,051 families. In 2015 there were 9. By 2020 there were 57. In 2025 there were 216, and 2026 is already at 135 with the year incomplete. That is a curve that did not exist a decade ago.
This is where the technical problem and the commercial problem meet. Data centers now compete for grid interconnection, they have punishing cooling loads, and their demand profile is flat while renewable supply is not. Heat storage sitting between a data center and the grid is a genuinely different engineering problem from heat storage sitting next to a cement kiln, and different problems produce different claims.
For anyone filing today, that distinction is the strategy. A claim about a carbon block at high temperature is a claim into fifty years of art. A claim about scheduling charge and discharge against a data center’s power purchase agreement, its cooling demand, and its backup power obligations is a claim into an eight-year-old field with roughly a thousand families in it. One of those is worth the drafting cost.
Geographically the split is what the manufacturing base predicts. China holds 19,675 of the 54,623 families. Japan holds 8,845. The United States holds 4,843, the EPO 4,031, Germany 3,604, Korea 2,658.
The Chinese share is heavily institutional. Xian Thermal Power Research Institute, Xian Jiaotong University, the Chinese Academy of Sciences and State Grid Corporation of China all appear in the top twenty. That is the same pattern we found in cybersecurity and in AI drug discovery: state-linked research institutions filing broadly and early in an area the state has designated as strategic.
For a US or European company, the practical implication is not that Chinese filings block you at home. Most of those families never leave China. It is that if your commercialization plan involves manufacturing in Asia or selling into that market, the ownership picture there looks nothing like the one in your domestic search report.
Three things follow from the data.
If you are a founder in thermal storage, stop trying to patent the storage medium and start patenting the integration. The medium is old. The control scheme, the charge strategy against a specific load, and the interface to a specific industrial or compute customer are not. Document them early, because the useful claims here are system claims and system claims need working examples to hold up.
If you are an investor, the patent count in the data room is not the diligence question. The question is whether the claims that exist read on the thing the company actually sells, and whether the incumbents with a thousand families each have anything pointed at it. That takes a landscape review, not a count.
If you are an acquirer, the interesting targets are not the largest portfolios. They are the small portfolios sitting inside the data center curve, filed in the last three years, before that space fills the way phase change materials did.
And if you are one of the incumbents on the leaderboard, the uncomfortable observation is that you have been filing in this space for thirty years and the venture money is flowing to companies with 27 families. Depth without a commercial thesis is a cost center. This is the moment to figure out which of your existing families read on the applications the market is now paying for.
The headline says $550 million into thermal batteries. The data says a mature field, incumbent depth, a startup layer that is small but focused, and one genuinely young corner where compute demand meets stored heat.
Both things are true at once. New money does not make a field new.
All patent counts in this piece are extended patent families from Minesoft Origin, searched August 4, 2026, using CPC classes F28D20, H02J15, Y02E60/14, F24H7 and F03G6 combined with thermal and heat storage text, grouped by extended family. Subfield counts use the same landscape with added keyword constraints. The data center and thermophotovoltaic counts use their own broader scopes and are labeled as such in the charts. This is landscape-level triage from bibliographic data. It is not a claim-scope, validity, or freedom-to-operate review, and it is not legal advice. Opinions on patentability or infringement come from qualified patent counsel.
Work with ipCapital Group
From invention to monetization, our team has guided 2,000+ engagements across the full IP lifecycle. Start with a free 30-minute discovery call.
Written by
Seth Cronin