Energy / Climate tech · Deep dive

Antora Energy

Thermal batteries that store renewable power in 1,800°C carbon blocks and discharge as industrial heat or TPV-converted electricity — $770M raised, a 5 GWh POET deployment, and Rondo still ahead on live commercial sites.

emerging

The question that decides it: Does Antora's TPV-plus-carbon-block architecture reach $/kWh parity with Rondo's brick-and-steam design at industrial scale before the OBBBA-era phase-outs unwind the tax credits its $770M investor base is underwriting?

My take ★★ Very interesting ★ Interesting − Maybe ✗ Not interesting

Snapshot

Antora Energy builds refrigerator-sized thermal batteries that heat solid carbon blocks to over 1,800°C with cheap renewable electricity and then discharge that energy as either process heat or, via thermophotovoltaic (TPV) cells, as electricity on demand. Founded in 2018 by two Stanford-trained engineers and one MIT thermal-storage PhD, the San Jose company raised a $550M Series C on July 30 2026 co-led by G2 Venture Partners and Eclipse — a step-up from the $150M Series B in February 2024 and the $50M Series A in February 2022 — bringing total equity funding to about $770M and total corporate plus project financing to roughly $1B. Its landmark reference project, a 50 MW / 5 GWh installation at POET’s Big Stone City biofuels plant in South Dakota, was commissioned in May 2026. Rondo Energy is still further along on live commercial dispatch, and the OBBBA rewrite of the IRA tax credits has changed the underwriting on every project Antora signs.

Founding story

Andrew Ponec’s path into thermal storage started with photovoltaics. As a Stanford undergrad in energy systems engineering he co-founded Dragonfly Systems, a power-electronics startup for utility-scale solar farms that SunPower acquired in 2014. He returned to Stanford for a materials-science PhD, then left — the conclusion he later gave interviewers was that photovoltaics were on a solved trajectory and the harder, more valuable problem was what industry does with the electrons.

He teamed with Justin Briggs, a Stanford applied-physics PhD who had spent roughly a decade in renewable-energy R&D, and David Bierman, an MIT mechanical-engineering PhD who had built solid-state converters and thermal storage prototypes as a graduate student and had already spun those out as Marigold Power in 2017. Marigold folded into Antora at incorporation in 2018. The founding thesis was not that another chemistry battery could out-compete lithium; it was that heat is roughly two-thirds of industrial energy use, that gas is the incumbent, and that nobody was building the box that would let a steel or cement plant run on wind and solar without a steam-cycle intermediary.

Antora went through the Activate Fellowship — the Cyclotron Road spinoff for hard-tech scientists — and took early cheques from Fifty Years. It has never been a YC company. The rest of the cap table then followed the industrial-decarbonization pattern: Breakthrough Energy Ventures and Lowercarbon co-led the 2022 Series A; Decarbonization Partners, BlackRock and Temasek’s joint climate vehicle, led the 2024 Series B; G2 and Eclipse led the 2026 Series C, with Salesforce Ventures, Ribbit, John Doerr and Liberty Mutual joining a cap table that now leans as much on infrastructure investors as on venture.

How it works

An Antora unit is a shipping-container-scale insulated box. Inside it, blocks of solid carbon — essentially graphite blocks, a material sold by the ton to the aluminium industry and priced at roughly $3.60/kWh of storage per Antora’s own figures — are wired to resistive heaters. Renewable electricity, taken during the cheapest hours of the day, flows through the heaters and lifts the blocks to about 1,800°C. Refractory and multi-layer insulation hold them there for hours or days.

Discharge takes one of two forms. For process heat, air is blown across the glowing carbon and delivered at temperatures well above what a heat pump or resistance boiler can reach — the range that steel, cement, glass, aluminium and heavy chemicals actually need. For electricity, TPV cells face the incandescent blocks: the visible-and-infrared photons streaming off the carbon hit multi-junction semiconductor cells that convert them directly to current, no turbine, no water, no moving parts. Antora has published TPV efficiency above 40% in its lab work — meaningful given a good combined-cycle gas turbine tops around 60% and a small steam turbine at industrial scale is usually below 40%.

The interesting design choice is the vertical integration. Antora runs what it says is the world’s largest TPV fab, a 2 MW/year line inside its Sunnyvale-area headquarters, later expanded across a three-building San Jose campus. Nobody was going to supply it TPV cells at scale, so it built them. That is the source of both the technical moat and the execution risk: TPV yields at industrial volumes are unproven and the semiconductor capex needed to move from 2 MW/year to tens of MW/year is real.

Product and business overview

The commercial product is sold as an offtake — the customer signs a long-term energy contract, Antora finances, builds and operates the plant, and delivers heat, electricity or both at a fixed price. That structure is the one that landed the POET Big Stone deal: POET does not own the batteries, it buys energy from them.

The physical stack breaks into: (1) the thermal core (carbon blocks, resistive heaters, refractory insulation); (2) the heat-delivery module (air handling, ducting, temperature control) that converts stored heat to process-usable output; (3) the TPV module (multi-junction cells, filters, cooling loop) that converts heat to electricity; (4) the plant balance-of-system (interconnection to the customer’s on-site solar or wind, controls, grid tie-back where relevant); and (5) the software layer (dispatch optimization to charge on cheap electrons and discharge into the customer’s load curve).

The most important business-level configuration is behind-the-meter. Antora openly frames the value as being able to take an industrial plant off gas without the eight-year wait for a grid interconnection. Front-of-meter deployments are possible but slower.

Business model and pricing

Antora does not publish a per-unit price. Public and semi-public data points to work backwards from:

• Graphite carbon at $3.60/kWh of storage capacity per the company’s own numbers; total installed all-in capex reportedly below $10/kWh — a large discount to lithium-ion at grid scale ($150-$250/kWh installed in 2025) but for a very different discharge profile.
• Offtake contracts denominated per MWh of delivered heat or electricity, structured to beat industrial gas plus expected carbon costs; Antora’s public marketing puts the target below the $/MMBtu delivered price of pipeline gas in most US industrial markets.
• Non-dilutive money that materially lowers the effective capex: an ARPA-E award of $14.5M in mid-2024 to accelerate the heat-and-power product, a California Energy Commission plus ARPA-E award over $4M in late 2023 to expand TPV production, and Antora appears in DoE Industrial Demonstrations Program cohorts (including Summit Materials cement projects). Total non-dilutive awards conservatively exceed $30M and function as a direct price cut on the first commercial units.

The revenue model is therefore infrastructure economics dressed up as software: long-duration offtake contracts underwritten by a mix of investor equity, project debt, IRA-era ITC/PTC eligibility (§48/§45Y for eligible electricity output, §45X for domestic manufacturing of the components that qualify), and DoE grants.

Traction over time

The Big Stone project is the material traction event. It is the largest single-site thermal-storage installation announced by any competitor, it is charging real electrons into real carbon blocks against a real long-term offtake with the largest US biofuels producer, and Antora built it in roughly a year. The counter-argument, kept honest: it is one project, and the majority of Antora’s $770M has gone into manufacturing and TPV rather than into a diversified reference book of live sites.

Market analysis

Industrial process heat is roughly two-thirds of industrial energy use and around 20% of global energy consumption per the IEA. The addressable global capex for industrial-heat decarbonization is variously quoted at $500B (360 Capital’s framing) and above; Nasdaq’s climate-tech coverage called it a trillion-dollar opportunity. Nearer-term market forecasts are much smaller — one third-party estimate puts the industrial heat decarbonization market at $9.2B in 2025 growing at ~9.4% CAGR — the honest read is that the true TAM is enormous but the near-term SAM is limited by permitting, interconnection and gas price.

Two structural forces move the market in Antora’s favour. One, industrial customers cannot get grid capacity: interconnection queues in most US ISOs exceed five years, and behind-the-meter thermal batteries collapse that timeline to months. Two, wind and solar keep getting cheaper on a per-MWh basis, so the arbitrage between the cheap hours (charge) and the delivered gas price (discharge) widens each year the customer waits.

Two forces cut the other way. First, US natural-gas prices remain low by global standards, which raises the payback period on any gas-displacement project. Second, the OBBBA of July 2025 has begun the phase-out of §48E and §45Y for solar and wind that begin construction after July 4 2026 and are placed in service after Dec 31 2027, and has tightened §45X eligibility for battery modules — the underwriting of every Antora project financed on those credits is now materially harder.

Competitive intel

The named competitors are in the frontmatter. The important distinctions:

Rondo Energy is the sharp-edged rival. Rondo’s brick-and-steam architecture is physically simpler, cheaper to certify, and further along commercially — Calgren live since 2023, a 100 MWh unit for a California fuel producer live October 2025, Covestro’s Brunsbüttel chemical plant broke ground January 2026. Rondo has raised less than Antora but has more live GWh discharging into industrial processes today. On steam-only, gas-displacement retrofits, Rondo wins deals Antora will not price out of.

Electrified Thermal Solutions attacks the very-high-temperature end (steel, cement) where Antora’s storage temperature is the differentiator. Smaller balance sheet, but a distinct cost curve.

Fourth Power is the head-to-head TPV rival, but is aimed at grid electricity rather than industrial heat, and is roughly one to two funding rounds behind.

Kraftblock, Hyme, Brenmiller are Europe-first or public-comp benchmarks rather than direct US wins.

The honest take: on steam and moderate-temp heat, Rondo is winning today. On very-high-temp process heat plus round-trip electricity, Antora’s TPV is a real technical wedge — but the round-trip efficiency of TPV-plus-storage against a decent industrial steam turbine is not yet settled at production scale.

History and evolution

2018: incorporation, Marigold folded in, Activate fellowship. 2020-21: quiet build; first-of-a-kind lab prototypes at ultra-high temperature. February 2022: Series A signals the industrial-decarb thesis has arrived. Late 2022 through 2023: R&D moves into first customer pilots; September 2023 Wellhead Electric 5 MWh unit near Fresno; San Jose TPV fab opens; Antora sets efficiency records for TPV above 40%. February 2024: $150M Series B and a shift from pilot to manufacturing; DoE brings Antora into the Industrial Demonstrations Program cohorts including Summit Materials cement projects. Mid-2024: ARPA-E $14.5M scale-up award. Late 2024 through 2025: San Jose campus doubled; second California facility added; hiring accelerates. July 2025: OBBBA is signed and rewires the credit stack every Antora project depends on. May 2026: POET Big Stone City comes online — 50 MW, 5 GWh, roughly a year from construction to commissioning. July 2026: $550M Series C.

The stumbles are mostly execution risk, not visible failures — the founders have kept the science on-schedule and the fabrication ramp on-narrative. The two honest scars: Antora is later to real commercial dispatch than Rondo, and its TPV product remains the piece of the stack most exposed to yield and semiconductor-capex risk.

What people say

The case for. Contrary Research and Canary Media, in successive coverage, credit Antora with (a) an unusually credible founding team for hard-tech thermal storage, (b) the only proven approach to sub-$10/kWh installed storage capex, and (c) a two-product architecture — heat and electricity — that lets it play in more customer segments than a pure-steam competitor. Latitude Media’s post-Series B coverage framed the $150M as “enough to actually build a factory,” an unusual endorsement of the manufacturing thesis. Investors describe Antora as the archetype of the industrial-scale climate company the OBBBA era still funds because it is domestic manufacturing.

The complaints. The recurring negative in trade press is timing: Antora spent 2023-2025 building manufacturing while Rondo was accumulating live GWh. howtostoreelectricity.com’s independent teardown notes that Antora’s TPV fab still runs at 2 MW/year — enough for only a handful of modules annually — and that scaling to tens of MW/year will demand advanced semiconductor capex and yield work that no one has done at that temperature. Analysts point out that TPV round-trip efficiency, while impressive in the lab, is not yet obviously better than a well-run industrial steam cycle when you count the losses in the whole delivery chain. And the OBBBA-era phase-outs of §48E/§45Y for wind and solar that supply the charging power, plus the tightening of §45X eligibility, have made the offtake economics harder to underwrite — a real risk to a company that has raised $770M against those credits’ continued existence.

Outlook: the open question

The answer conditions are these. For Antora to be right, three things have to be true by the end of 2028: (1) it needs at least two more Big-Stone-scale projects live, on schedule and on cost, not just announced; (2) its all-in installed $/kWh at production volumes has to close the gap to Rondo’s simpler brick architecture, or its TPV round-trip efficiency has to demonstrably beat a good industrial steam turbine on a full-chain basis in a customer’s actual accounts, not a press release; (3) the offtake economics have to survive the OBBBA phase-outs of §48E/§45Y and the tightening of §45X — meaning either the underlying gas-displacement arbitrage has to be so wide that credits are gravy, or the projects have to be structured to lock in credits under the transition rules before the July 2026 start-of-construction cliff.

For Antora to be wrong, the failure paths are just as concrete. Rondo continues to accumulate live commercial GWh at simpler unit economics and Antora spends 2027-28 raising more capital against a manufacturing ramp rather than a revenue ramp. The TPV product turns out to be a lab efficiency achievement that does not economically translate at plant scale. Or the OBBBA-era rollback makes Antora’s per-project IRR too thin to attract the project-finance capital its offtake model depends on, and the company has to fund construction off its own balance sheet — a $770M cap table that will not fund many more $150M sites without a mark-down.

The single question that decides it: does Antora’s TPV-plus-carbon-block architecture reach $/kWh parity with Rondo’s brick-and-refractory design at industrial scale before the IRA §45X and §48E credits its $770M cap table depends on are unwound? If yes, Antora is the anchor tenant of a decade of US industrial reindustrialization. If no, it is the more ambitious but slower loser of a market Rondo pockets while Antora is still ramping TPV yield.

How to attack it

The specific wedge is a retrofit-first, brick-only, TPV-free product priced explicitly against Rondo for steam and hot-air process heat below 1,200°C. Antora’s TPV is the differentiator when the customer needs electricity or ultra-high temperature; for the ~60% of industrial heat load that is steam and moderate-temp air, TPV is dead weight in the capex and Rondo has proven the simpler architecture wins deals. A well-funded attacker with $75M-$150M could ship a firebrick-and-heater box, license or commodity-source refractory, and sell to the same food-processing, biofuels, pulp-and-paper and chemicals customers Antora and Rondo both chase — undercutting Antora’s structural cost by dropping the TPV line and undercutting Rondo by starting on a lower cost base with none of the legacy pilot-era engineering.

The exploitable weaknesses in Antora’s current position: (1) its TPV fab still runs at 2 MW/year and every dollar tied up in scaling it is a dollar not deployed at customer sites — the vertical integration that is a moat at scale is a drag at this stage; (2) offtake economics leaning on §48E/§45Y for the wind and solar that charge each site expose Antora to project-finance withdrawal if the OBBBA transition rules get further tightened; (3) go-to-market is thin — most of Antora’s public references are POET, Wellhead, Summit Materials, Shell LOI — so a fast-moving competitor with a broker channel into industrial EPC firms could sign three deals in the time Antora signs one; (4) the founding team is deep on science but the operational depth for running twenty simultaneous multi-hundred-million-dollar construction projects is unproven; (5) product pricing is per-project and opaque, which is fine for the first ten sites but a channel disadvantage against a competitor publishing a per-MMBtu number that a customer CFO can drop into a model.

Adjacent-segment play

The obvious adjacent-segment play is data-center backup and shaping. Antora already frames the Series C narrative around data-center demand, and TPV-plus-storage sized at hundreds of MW behind a hyperscaler campus is a plausible use of the exact same technology, sold to a different buyer (hyperscaler procurement rather than industrial plant manager) at a different price point (data-center-grade reliability commanding a large premium over $/MMBtu of displaced gas). The competitor set becomes different — Bloom Energy fuel cells, on-site gas turbines, grid-connected lithium — but the physical wedge (behind-the-meter, no grid queue, multi-day storage) is even sharper for hyperscalers than for chemical plants.

A second adjacent segment is district heat in cold-climate cities, particularly in northern Europe where the gas-to-electricity price ratio is inverted versus the US and Kraftblock and Hyme are already selling. Antora’s carbon block would be over-specified for the temperatures needed, but the operating economics are dramatically better than in the US Gulf Coast; the play is licensing rather than direct build.

A third: TPV cells sold as a merchant product to third-party thermal-storage or waste-heat-recovery companies. Antora’s 2 MW fab is the world’s largest TPV production line; if the module business at scale is slower than plan, spinning TPV out as a component supplier to a fragmented waste-heat-recovery market is a real optionality that is worth more than zero on the cap table. The wedge does generalise — the physical asset (a fab making high-efficiency TPV cells) is portable across buyer segments in a way pure project developers are not.

Sources and further reading

• Antora Energy Raises $150M to Slash Industrial Emissions — Antora, Feb 2024
• Antora Closes $550M Series C — Antora, Jul 2026
• Antora and POET Commission 5 GWh Thermal Battery Project — Antora / POET, May 2026
• Antora Energy Business Breakdown & Founding Story — Contrary Research
• Bill Gates, Chris Sacca invest $50 million in Antora Energy — CNBC, Feb 2022
• Antora snags $550M for heat batteries to run data centers, factories — Canary Media, Jul 2026
• Antora Energy Deep Dive 2026 — howtostoreelectricity.com, 2026
• Antora Energy Doubles San Jose Manufacturing Footprint — TipRanks
• Rondo Energy vs Antora Energy — Cheap clean energy could unleash thermal storage — Canary Media
• One Big Beautiful Bill Act: New Law Disrupts Clean Energy Investment — Latham & Watkins, Jul 2025
• ARPA-E Investor Update Vol. 22: Antora Energy’s Thermal Batteries — ARPA-E