Form Energy

Phase A — Understand the business

Lens 1 · Company Overview

Form Energy sells multi-day (up to 100-hour) grid-scale energy storage built on a reversible-rusting iron-air battery: the cell stores energy by chemically converting iron to rust (discharge) and back to iron (charge), using only iron, water, and ambient air. That is the whole thesis in one sentence — it trades round-trip efficiency for radical cost and duration.

• Founded 2017 as an MIT spinout, Somerville MA; now HQ'd in Somerville with its factory in Weirton, WV.
• Product: modular battery "blocks" sized by duration, not by stacking cells — you add iron mass to add hours, so the marginal cost of the 50th hour is iron ore, not electronics. Targets the 24–150-hour window where lithium is too expensive (you'd need 100× the cells) and hydrogen too inefficient/immature.
• Customers: regulated and cooperative electric utilities plus, newly, hyperscaler data centers — Great River Energy, Xcel Energy, Georgia Power, PacifiCorp (Berkshire/Buffett), and now Google.
• Suppliers: ArcelorMittal (iron materials, also a strategic investor); broadly earth-abundant, North-America-dominant supply chain.
• Competitors: Eos Energy (zinc), ESS Inc (iron-flow), Antora Energy (thermal), Hydrostor (compressed air), Energy Vault, Ambri (liquid metal), and incumbent pumped hydro.
• Contract structure: project-by-project utility procurement, increasingly embedded in utility Integrated Resource Plans (PacifiCorp put 3,073 MW of iron-air in its 2025 IRP) and in bespoke hyperscaler deals (Google ~$1B). Revenue is lumpy capital-equipment sales tied to multi-year deployment schedules (the Google system phases in 2028–2031) — not recurring.

Plain-terms model: Form is a capital-equipment manufacturer selling the cheapest stored-electron on the grid for the jobs lithium can't afford to do. Its product is "firmness" — the ability to ride through multi-day wind/solar lulls — sold mostly to utilities that can rate-base it.

Lens 2 · Supply Chain

Upstream → Form → end customer, named at each node:

• Raw input: iron ore + water + air. Iron-ore content of a cell costs ~$0.10/kWh, electrolyte ~$0.01/kWh — i.e. ~$0.11/kWh of raw material. This is the structural cost story — the bill of materials is rust.
• Strategic iron supplier: ArcelorMittal — world-leading iron-ore producer, invested via Series D ($200M lead) and Series E ($450M lead) and agreed to non-exclusively supply jointly-developed iron materials. This is the single most important supply-chain relationship: it secures the input and aligns a steel major behind the technology.
• Cell/system manufacturing: Form Factory 1, Weirton WV — built on the carcass of the former Weirton Steel mill (poetic: rust-belt steel → rust battery). Broke ground May 2023, commercial production started ~2024–2025, >400 employees, a new line targeting up to 20 GWh/yr by 2027, backed by a $150M DOE award plus state incentives.
• EPC / grid integration: delivered with utility partners and, for the Google project, Xcel Energy as the deploying utility. Form also has a GE Vernova collaboration (GE Vernova is also a Series F investor) — a route to grid-integration and balance-of-plant scale.
• End customers (buyers): utilities (Great River, Xcel, Georgia Power, PacifiCorp) and hyperscalers (Google).

Chokepoints / single-source dependencies:

1. One factory. Form Factory 1 is the entire supply of cells today — a single-site concentration risk until a second factory exists.
2. Iron-electrode process IP is proprietary; the material is abundant but the reversible-rusting electrode is the hard part (USPTO filings on iron-electrode additives). The moat is process, not ore.
3. Domestic-content is a feature, not a risk — earth-abundant, North-America-sourced supply chain insulates Form from the China-graphite/lithium chokepoints that dog lithium-ion.

Verdict on this lens: the chain is unusually short and domestic — its vulnerability is concentration (one factory, one anchor iron partner), not geopolitics.

Lens 3 · Competitive Advantages (moats)

• Cost moat (the big one). System-level target ~$20/kWh — roughly one-tenth of lithium-ion at the duration the product serves. No competing chemistry credibly claims a multi-day cost this low; the BOM is rust.
• Duration moat. Adding hours = adding iron mass, so cost scales ~linearly and flatly with duration — lithium's cost scales with cells, so it becomes absurd past ~10h. Form owns the 100-hour point on the curve in a way lithium structurally cannot reach.
• Process IP / know-how. The reversible iron-air electrode is hard to make cycle durably (the historical failure mode of iron-air); Form's patent estate and a multi-year head start are the real barrier, since iron and air are free.
• Anchor-partner moat. ArcelorMittal (iron + capital), GE Vernova (grid integration + capital), T. Rowe Price/Coatue/Temasek/GIC/Breakthrough (deep capital). A steel major upstream and a grid-equipment major downstream is a hard syndicate to replicate.
• Safety/siting moat. No thermal runaway, UL9540A certified, no rare earths — easier permitting and siting than lithium near data centers and towns.

Bargaining power: over suppliers — strong (iron ore is a global commodity with many sellers; ArcelorMittal is a partner not a gatekeeper). Over customers — moderate and rising: Form is effectively the only merchant supplier of true multi-day iron-air at scale, so a utility wanting 100-hour firmness has few alternatives — but the customer can always choose a different solution to the same problem (gas peaker, pumped hydro, overbuild + curtail), which caps pricing power.

The honest moat caveat: the moat protects the cost-per-kWh-of-capacity, not the cost-per-kWh-delivered. At sub-40% RTE, two-thirds of moat value is contingent on the grid valuing capacity/firmness over throughput (see Lens 13).

Lens 4 · Segments

segments.csv is empty (header-only) and Form is private, so there is no audited revenue-by-segment breakout — n/a, not disclosed. What can be said structurally:

• By product: effectively single-product (iron-air multi-day systems). No second commercial line disclosed.
• By end-market: historically regulated/cooperative utilities (pilots → IRP-scale procurement); a new and now-dominant hyperscaler data-center channel opened with Google (~$1B) in Feb 2026. The data-center channel is the inflection — it pulls demand forward of the slow utility-procurement cycle.
• By geography: US-only disclosed (WV factory; MN, NY, GA, VA, and Western US via PacifiCorp deployments).
• Pipeline trend (the only "segment" number that exists): ~750 MW / 75 GWh under active development as of April 2026, up ~375% from >200 MW / 20 GWh in Oct 2025. That is the single most important growth datapoint in the file — pipeline more than tripled in ~6 months, driven by the data-center channel.

---

Phase B — Measure performance

Lens 5 · Funding & Valuation Trajectory (+private swap — replaces "Earnings Result")

No earnings exist. The scoreboard is the cap raise:

• Total raised: ~$1.2–1.3B across ~10 rounds.
• Round history (disclosed):
• Series D — ~$200M, Aug 2021, led by ArcelorMittal.
• Series E — $450M, Sep 2022, led by ArcelorMittal — the largest single round.
• Series F — $405M, Oct 2024, led by T. Rowe Price, with GE Vernova joining + TPG Rise Climate, Breakthrough Energy Ventures, Coatue, Temasek, GIC, Capricorn, EIP, NGP, The Engine.
• Plus earlier A–C and an undisclosed early-2025/2026 round.
• Valuation — CONFLICT, surfaced not resolved:
• Tracxn lists a "current valuation of $1.2B" — this reads like total-raised mislabeled as valuation and should be treated skeptically.
• The Series F (Oct 2024) is widely reported at ~$3.5B post-money, and secondary-market activity pushed an implied ~$3.3B market cap in late-2025/early-2026.
• Resolution: do not pick one. Most-credible read: ~$3.3–3.5B post-Series-F, pre-final-round.
• Burn signals: $1B+ raised + a $760M-class factory + a single product still pre-mass-revenue ⇒ heavy cash burn, capex-led. The reported plan for a $300–500M "final" private round before a 2027 IPO is itself a burn/runway tell — they need one more tank of fuel to reach the tradeable milestone.

Lens 5b · Traction & Unit Economics (+private add)

• Contracted demand (the real traction metric): ~750 MW / 75 GWh pipeline (Apr 2026); Google ~$1B for a 300 MW / 30 GWh Minnesota system (with Xcel); PacifiCorp 3,073 MW in its 2025 IRP. The Google deal alone is a ~$1B revenue event spread over a 2028–2031 deployment.
• Unit economics signal: target ~$20/kWh system cost vs raw material ~$0.11/kWh — the gap is manufacturing, integration, and margin; gross-margin is n/a — not disclosed (private), but the BOM headroom is enormous if yields and cycle-life hold.
• Revenue run-rate / ARR: n/a — private, not disclosed. Equipment sales are lumpy and back-loaded to deployment years, so even post-IPO this won't look like SaaS.

Lens 6 · Founder/Operator Signal (+private swap — replaces "Earnings Calls")

No earnings calls. Reading the founders' public posture (interviews, conference talks):

• CEO Mateo Jaramillo consistently frames Form as solving the "multi-day duration gap" lithium can't, and is deliberately unhurried on IPO timing — told Latitude Media there's "no strict timeline… but it's certainly on the horizon". Tone = patient-capital, infrastructure-builder, not hype-cycle.
• Recurring message: cost and domestic supply chain over efficiency — they lead with the sub-40% RTE rather than hide it, reframing it as the price of being 10× cheaper. That candor is a credibility signal.
• The Google $1B deal (Feb 2026) is the de-facto "earnings beat" — a top-tier, technically-sophisticated buyer underwriting the technology at scale shifted the narrative from "promising pilot" to "bankable infrastructure".

Lens 7 · Cap Table & Secondary Marks (+private swap — replaces "Comps")

Syndicate quality (the IPO-proximity tell):

• Crossover / public-market funds present: T. Rowe Price (led Series F), Coatue, Temasek, GIC, Fidelity-class long-only appetite implied by secondary demand. A T. Rowe-led round is a textbook IPO-proximity signal — crossover funds enter when a private is ~2–3 years from public.
• Strategics: ArcelorMittal (iron), GE Vernova (grid equipment) — both operating partners, not just financial.
• Climate/deep-tech tier-1: Breakthrough Energy Ventures (Bill Gates), TPG Rise Climate, Capricorn, Energy Impact Partners, The Engine (MIT), NGP, Prelude. Angel backers reportedly include Bill Gates and Jeff Bezos.
• Secondary marks: shares quoted ~$14.81 on Hiive/Forge; secondary demand implied a ~$3.3B cap late-2025/early-2026.

Mechanism/peer comps (private LDES — no clean public multiples; this is the right comp axis):

P/E, EV/Sales, ROE: n/a — Form is private; peers are pre-profit or distressed, so no meaningful multiple set exists. Not sourced → not fabricated.

Lens 8 · Funding & Product Catalysts (+private framing of "Stock-Price Catalysts")

No tradeable stock, so the "catalysts that moved the mark" are funding + commercial milestones:

• 2020 — Great River Energy pilot (Cambridge, MN; 1.5MW/100h) — first commercial deployment, validated the tech.
• 2021 — Series D + ArcelorMittal; the $20/kWh reveal reframed the category ("substantial breakthrough").
• 2022 — $450M Series E (ArcelorMittal) — scale capital.
• 2023 — Weirton groundbreaking; Xcel coal-plant-site deals.
• 2024 — $405M Series F (T. Rowe) + GE Vernova + $150M DOE award; commercial production at Form Factory 1.
• 2025 — PacifiCorp 3,073 MW in its IRP; pipeline >200 MW.
• Feb 2026 — Google ~$1B / 300MW / 30GWh with Xcel (largest battery by energy capacity ever announced); pipeline jumps to ~750 MW / 75 GWh; IPO chatter for 2027.

Pattern: the mark is driven by anchor validation events (a steel major, then a hyperscaler) and policy/factory milestones — not by quarterly numbers. The next mark-mover is the final private round and any first-large-system commissioning.

---

Phase C — Judge people & books

Lens 9 · Management

• Mateo Jaramillo — Co-founder & CEO. Built and led Tesla's stationary storage business from scratch (founded the program; commercial lead on $100M+ development and $500M+ production contracts); earlier COO/founding team at Gaia Power Technologies. Harvard A.B. (Economics) + Yale Divinity M.A.. Read: a commercial founder who has actually shipped grid storage at scale — rare and exactly the profile this stage needs.
• Yet-Ming Chiang — Co-founder, Chief Scientist. MIT Kyocera Professor of Materials Science; serial deep-tech founder (A123 Systems, 24M, Desktop Metal). The scientific credibility anchor. A123 is a double-edged reference (breakthrough chemistry that later went bankrupt — see Lens 13).
• Other founders: Marco Ferrara (SM/PhD MIT), William Woodford (PhD MIT), Ted Wiley (energy-storage operator) — a materials-science + operations founding bench.
• Track record: delivered the iron-air tech from lab to a commercial factory + a $1B order in ~9 years — genuinely built something hard.
• Skin in the game / insider ownership: n/a — private, not disclosed; founders presumably hold meaningful equity but no insider-transactions.csv exists.
• Capital allocation: raised ~$1.2B and put it into one focused bet (the iron-air factory) rather than diversifying — disciplined, but concentrated. No buybacks/M&A (early-stage).
• Red flags: none specific surfaced. Generic founder-archetype risk: a brilliant-chemistry founder (Chiang) whose prior marquee company (A123) went bankrupt despite great tech — the technology working is necessary but not sufficient (Lens 13).
• Archetype: founder-led, mission-driven, patient-capital infrastructure builder. Implication: long time-horizon, capex-heavy, dilution-tolerant — fine for a private, something public-market investors will need to underwrite carefully.

Lens 10 · Forensic Red Flags

No audited financials exist (private; financials/segments/customers/capex CSVs all empty ). Standard income-statement/balance-sheet forensics are n/a — not disclosable. What a forensic eye flags structurally:

• Pre-revenue capex burn vs. raise cadence — the need for a "final" $300–500M round before IPO says cash runway is the live constraint; watch for a down-round or flat-round as the tell if commissioning slips.
• Order-to-revenue lag — the Google "$1B" is a contract value deploying 2028–2031, not booked revenue; any public S-1 will show a large backlog against thin recognized revenue. Don't conflate the two.
• Cycle-life / warranty reserve — iron-air's historical failure mode is electrode degradation; the real accounting risk post-IPO is warranty/performance-guarantee reserves on 20-year utility contracts. Unverifiable now.

Regulatory findings (required sub-section):

• SEC (EDGAR LR + AAER): none — Form Energy has no CIK; it is private and not required to file.
• Item 3 Legal Proceedings: n/a — no 10-K exists.
• Non-SEC (FTC/DOJ/FDA/etc.) web sweep: no material enforcement, consent decree, settlement, or penalty surfaced against Form Energy in searches.
• Conclusion: No material regulatory or legal findings — verified via the pre-fetched SEC EDGAR EFTS sweep (no CIK / 0 findings) and web search as of 2026-06-17. No 10-K Item 3 exists (private).

Lens C-extra · Science & Exclusivity (de-risking sub-lens for a deep-tech private)

• Mechanism validation: reversible iron-air is scientifically real and now commercially demonstrated (Great River pilot, Form Factory 1 production, UL9540A) — the question is durability-at-scale and yield, not whether it works.
• IP estate: iron-electrode additive + reversible iron-air patents (USPTO 10,374,261; 8,758,948) — the process IP is the moat since the inputs are free.
• The hard truth: the science is validated; the un-validated variable is sub-40% RTE economics holding up under real grid dispatch and policy (Lens 13).

---

Phase D — Project & stress-test

Lens 11 · IPO-Readiness & Path-to-Tradeable (+private swap — replaces "Forward Projection")

(No EPS model — private, no financials. No Brier forecast logged: breadth/watchlist loop + a private name with no scoreable EPS line. The scoreable question would be the IPO-timing binary, noted below for a future /thesis pass.)

• Stage: late-stage, commercial-scale, pre-IPO. Factory live, $1B anchor order in hand, crossover funds (T. Rowe) on the cap table.
• Stated path: a $300–500M "final" private round, then a targeted 2027 IPO — though management is explicitly non-committal on timing.
• Milestones that unlock an S-1:

1. Close the final private round (the fuel-to-IPO) — near-term, watch 2026.
2. Commission the first large-scale system(s) and show real-world cycle-life/performance — the Great-River 100h system and early MN/NY/GA/VA deployments.
3. Convert the Google + PacifiCorp + IRP pipeline into recognized revenue ramp (deployments phase 2028–2031, so revenue is back-loaded — an IPO would price largely on backlog, not trailing revenue).

• Estimated window: 2027 base case, with real risk of slippage to 2028+ if market conditions or commissioning timelines move (management's own hedging supports this).
• Implied valuation entry point: secondary marks ~$3.3B / ~$14.81/share; the "final round" will reset this.
• Be-early payoff: this is a dossier-warm, watch-for-the-round name. Note: I could not write back to private-watch.json — the file is not present on disk; flag for the central wave-marking step to create/seed the entry (stage: late-stage pre-IPO; ipo_readiness: high-but-unscheduled; catalyst: final private round → 2027 IPO).

Future scoreable forecast (for a later /thesis pass, not logged here): "Form Energy files an S-1 / goes public by 2027-12-31, p≈0.45".

Lens 12 · Bull vs Bear

Bull case. Form owns the structurally cheapest multi-day storage on Earth at the exact moment the grid acquires its first truly inelastic, 24/7, multi-day load: AI data centers. The Google $1B deal isn't a pilot — it's the most sophisticated power buyer on the planet underwriting iron-air as bankable infrastructure, and it dragged the pipeline up 375% in six months. The cost curve ($20/kWh, BOM of rust) is unmatchable by lithium in the 100-hour window; the supply chain is domestic and China-proof; the syndicate (ArcelorMittal upstream, GE Vernova downstream, T. Rowe for the exit) is best-in-class. If multi-day firmness becomes a procured grid product — and IRPs + data-center deals say it's becoming one — Form is the default supplier of a brand-new category, with a 2027 IPO into a market starved for "clean firm power" stories.

Bear case (permanent-impairment risks).

1. Sub-40% round-trip efficiency is not a footnote — it's the business model's load-bearing wall. At <40% RTE you burn 2.5 MWh to deliver 1 MWh and cycle maybe ~25×/year. The entire economic case requires charging electricity to be nearly free (curtailed wind/solar at $0–10/MWh). If that surplus doesn't materialize where/when needed, the asset is a stranded capacity payment.
2. The bet is on policy, not physics. Iron-air only pencils if regulators pay for capacity/firmness rather than throughput, and if subsidies (DOE awards, 45X-class manufacturing credits) persist. The OBBBA (Jul 2025) is phasing out IRA clean-energy credits — 45X manufacturing is relatively protected but the demand-side project credits are being compressed. A policy regime that rewards cheap throughput over firm capacity guts the thesis.
3. Cheaper/better mousetraps. Lithium keeps falling and stretching to 8–12h; if 12–16h lithium + overbuild covers most "multi-day" needs at >85% RTE, Form's 100-hour niche shrinks. Sodium-ion and competing LDES (Eos zinc, Hydrostor CAES) crowd the same slot.

Pre-mortem (18 months out, thesis broke): The final private round priced flat or down because commissioning of the first large systems slipped and revealed worse-than-modeled real-world cycle-life; the IPO window pushed to 2028; meanwhile OBBBA's credit compression made utilities re-rate firm-capacity value downward, and a cluster of 12-hour lithium projects won procurements Form expected. The Google deal stayed a single marquee logo rather than the first of many.

Are the marks too high? ~$3.3–3.5B on pre-meaningful-revenue is a venture bet on a category, not a multiple on a business — defensible for patient capital, rich for anyone needing near-term liquidity.

Contrarian view (what the market refuses to see): Bears obsess over RTE as if Form were competing with lithium on arbitrage. It isn't. Form competes with gas peakers and overbuild-and-curtail for firmness, where the comparison isn't efficiency but $/kW of dependable capacity over multi-day stress — and on that axis, rust at $20/kWh is the cheapest insurance the grid can buy. The real question is governance, not physics: will markets pay for firmness? Data centers, which can't tolerate downtime, are answering yes with checkbooks before regulators do.

Lens 13 · Devil's Advocate (short-seller)

Dismantling the bull case:

• The efficiency tax is permanent and structural. No amount of scale fixes sub-40% RTE — it's thermodynamic. Every bull deck waves this away; it is the whole ballgame. A 40% RTE asset cycling ~25×/year has to be dirt cheap and heavily subsidized to clear, and "cheap to build" ≠ "cheap to operate".
• Revenue is a backlog mirage. "$1B Google deal" deploys 2028–2031. The PacifiCorp 3,073 MW is an IRP line item, not a signed PO. An S-1 will show enormous "pipeline" against tiny recognized revenue — exactly the setup that de-rates hard when growth-equity sentiment turns.
• The most dangerous competitor bulls underrate: lithium itself + sodium-ion. Lithium's relentless cost decline keeps eating "duration." If 12–16h lithium at 85%+ RTE covers the bulk of real grid needs, the true addressable multi-day niche is far smaller than the TAM decks claim.
• The A123 ghost. Co-founder Chiang's prior breakthrough-chemistry company (A123 Systems) had genuinely great tech and still went bankrupt — proof that in grid hardware, manufacturing yield, warranty exposure, and capital intensity kill more companies than bad science does.
• Policy is the load-bearing assumption. Strip the DOE award and the firm-capacity regulatory tailwind and the model wobbles; OBBBA shows the policy ground is moving against clean-energy demand credits even if 45X manufacturing survives.
• What must hold for today's ~$3.3B mark: (a) multi-day firmness becomes a paid, procured product at scale; (b) first large systems commission on-time at modeled cycle-life; (c) the final round prices up and the IPO lands by ~2027; (d) lithium doesn't stretch into the niche. If growth disappoints 20–30%, or commissioning slips, the down-round/IPO-delay scenario is the base case, not the tail.
• Single permanent-impairment scenario (most plausible): real-world cycle-life/RTE comes in materially below lab numbers on the first utility-scale systems, warranty reserves balloon, and the firm-capacity payment regime fails to materialize fast enough — Form becomes a subscale, capital-starved manufacturer of a product the grid admires but won't pay enough for. Plausibility: moderate — the science works; the market design and execution are the real risk.

Lens 14 · Management Questions (ordered by information value)

1. At sub-40% round-trip efficiency, what specific market mechanism or contract structure makes the unit economics work without a permanent subsidy — and how much of your pipeline is contracted on capacity/firmness payments vs. energy arbitrage?
2. Of the ~750 MW / 75 GWh pipeline, how much is signed, binding PO vs. IRP line-items / LOIs / MOUs — and what's the contracted-revenue figure?
3. What were the real-world round-trip efficiency and cycle-life on the first commissioned 100-hour system (Great River / Cambridge) vs. your lab model?
4. What is your warranty/performance-guarantee reserve methodology for 20-year utility contracts, and how is it sized?
5. How large is the "final" private round, at what valuation, and is it structured to be the last money before an IPO?
6. What is your current monthly burn and runway, and what milestone does this next round have to reach?
7. How exposed is your model to OBBBA / IRA credit changes — specifically, which credits (45X manufacturing, ITC) does your customer pricing assume, and what's the downside if they're cut?
8. Beyond Google, how many other hyperscaler / data-center deals are in active negotiation, and is this a repeatable channel or a single marquee logo?
9. What is Form Factory 1's actual achieved yield and output today vs. the 20 GWh/yr 2027 target — and when does Factory 2 start, given single-site concentration risk?
10. How do you defend the niche if 12–16-hour lithium (or sodium-ion) at >85% RTE covers most of what utilities call "multi-day"?
11. What is gross margin at scale on a system, and what's the bridge from ~$0.11/kWh BOM to the $20/kWh system price?
12. How much founder/employee ownership survives to IPO after ~$1.2B raised, and what's the post-money cap-table dilution picture?
13. What's the ArcelorMittal supply agreement's pricing/exclusivity structure, and what happens to your iron cost if steel/iron markets spike?
14. What's the GE Vernova relationship beyond capital — does it give you a balance-of-plant / grid-integration scale advantage rivals lack?
15. What is the single scenario you most fear could permanently impair the business, and what are you doing about it today?

---