← All thesesEOSEUnattractiveHigh convictionThursday 5 November 2026עברית

Eos Energy Enterprises

This is the archived full research. Read the story

A Pittsburgh maker of zinc batteries that store power for the grid over four to twelve hours, still selling them for less than they cost to make.

Our viewUnattractive High conviction
The scores
The profile · 1-4
Growth3/4 ↓

Sales rose sevenfold in 2025 to outside buyers, but June's quarter was 80% related party and outside sales fell.

Business quality2/4 →

US content rules favour it, but US-made lithium-ion earns the same credits and already won four to eight hours.

Financial position1/4 ↓

Cash for about three quarters at June's burn, $948M of debt, and the margin turn not due until after that.

Execution record2/4 →

Missed 2025's revenue expectation, a late capacity milestone, a narrowed guide, and four finance chiefs in fifteen months.

Per share capture1/4 ↓

Shares up sevenfold since 2021, about 644M fully diluted, and growth shared with Cerberus through Frontier.

Potential2/4 →

A large US market, but the eight-hour slice is unmeasured and half the backlog is a related-party agreement nobody can read.

Risk profile2/4 ↓

Two customers were 97.7% of June's revenue, one controlled by the lender, with a class action pending.

Valuation1/4 →

About 3.8 times sales, 501% of a three-stage value near $0.36 at an 8% return: priced for perfection, beyond the 20% margin.

Evidence quality3/4 →

Cash, debt, buyers and Frontier's terms are filed; performance, backlog and the margin path are management's.

The bet · 2-3 years
Priced for success0.4×

If it works 0.4× · if it fails you keep up to 0%.

The Verdict

We rate Eos Energy Unattractive. The business is weak: US tax rules favour its battery, but each one sold costs more to make than it brings in, and in September 83% of sales went to a venture its own lender controls. The price is stretched, still paying for sales near $0.9B a year, about three times the last twelve months. As a bet it is Fragile: even a full factory is worth about half last month's price, and if margins do not turn before cash runs short, shareholders could be left with nothing.

The long caseIf US grids need storage running eight hours and more, and the rules keep Chinese cells out, a fireproof US-made battery wins those orders.

How we weighed it

The price against our value

Against it

Price at 483% of our value

The price sits far above our value and above the bull case.

The business

Against it

Weak

A weak business: each battery is priced below what it costs to make.

Growth against what the price assumes

Cuts both ways

Growth rests on the lender's venture; outside sales barely moved.

What could go wrong

Against it

The margin may never turn, cash may run short and the lender is also the customer.

Against competitors

Cuts both ways

Rivals are cheaper, so the high price is its own.

How its group is behaving

Cuts both ways

Power names move together on AI power news.

The latest news and report

Cuts both ways

The last report improved the margin, but the forecast slipped.

The evidence under this - 7 claims
C04facthigh confidence
Q3 2026 gross loss was $44.2M on $82.7M of revenue, a 53% negative margin; adjusted gross margin was negative 44%, against negative 62% in Q2, and adjusted EBITDA negative $63.8M against $71.4M. The nine months lost $137.4M at the gross line on $208.5M of revenue.
  1. 2026-11-05editedQ3 2026 results read from the 10-Q and release: Q3 gross loss and adjusted margin
historical · Q3 2026written 2026-10-06
C21inferencemedium confidence
At $2.14 the company value is about $1.55B on filed figures (365.8M shares, $1,040M of debt principal, less $271.2M of cash), 5.8 times trailing revenue of $266.5M and 5.0 times the $312M midpoint of the lower half of the 2026 guide; the vendor's figure is 4.5 times. At the filed rivals' median of 1.7 times sales, $1.55B needs about $0.91B of revenue, between the 2027 and 2028 consensus, and that before the gross margin has turned positive.
derivedArithmetic on filed figures (C05, C06, C08) at the pack's close (C19), against the pack's cohort median (C22) and consensus (C20).
against this

Counting the preferred and penny warrant as equity raises the figure to about $2.4B, and the cohort is two companies, one near insolvency, so the median is a thin yardstick (C07, C22, C25).

  1. 2026-11-05editedre-struck at the 4 November close on the Q3 balance sheet
current state · 2026-11-04written 2026-10-06
C10facthigh confidence
Eos put $112.6M into Frontier Power USA, the net proceeds of July's $75.0M direct offering and August's rights offering, which sold 6,885,218 of 27,367,171 units against a roughly $150M target. Eos holds 112.6M of 312.6M units, about 36%; Cerberus holds 150.0M, including 50.0M founder units for no cash, and appoints four of seven managers; in liquidation Cerberus's and Hudson Bay's cash comes back before Eos's.
current state · 2026-08-04written 2026-10-06
C03facthigh confidence
Q3 2026 revenue was $82.7M, of which $68.4M was related-party product revenue from projects financed by a Cerberus affiliate and contributed to Frontier Power USA, about 83% of the quarter; product revenue from everyone else was $14.3M, against $12.7M in Q2 and about $56.7M in Q1.
  1. 2026-11-05editedQ3 2026 results read from the 10-Q and release: Q3 related-party and outside product revenue
historical · Q3 2026written 2026-10-06
C49estimatemedium confidence
At the first half's rate of about $131M a quarter, the $271.2M of unrestricted cash at September, which already holds the $87.0M DOE advance, lasts about two quarters, to around the spring of 2027, before the margin turn management promises would have time to show. On the Q3 call management said cash used has tracked the operating loss with no working-capital drain, capital spending moderates from the fourth quarter, and it will give 2027 figures on the Q4 call.
derivedC05's first-half outflow carried forward against the 30 September unrestricted cash; the third quarter's own outflow is distorted by the Frontier investment and equity raised (S35).
against this

Customer deposits ran $51.0M in Q2, 45X credits come in as cash, and Cerberus's $105M revolver is available at its discretion, so the outflow could be smaller (S01).

  1. 2026-11-05editedQ3 2026 results read from the 10-Q and release: cash runway re-struck on the Q3 balance
  2. 2026-11-05editedQ3 call read: management says cash burn tracks the loss, capital spending falls from Q4, and declined a 2027 figure.
forward · 2026-11-05written 2026-10-06
C15management claimlow confidence
Management says scrap, labour, materials and Thorn Hill's ramp give a path to over 72 points of adjusted gross margin improvement over the twelve months from August 2026, assuming the plan and expected volumes, which would take it from negative 62% to about positive 10%. On the Q3 call the chief financial officer said the third and fourth quarters together are well over half of the way to the 72 points, with materials cost down and Thorn Hill cutting conversion cost.
against this

Seven straight quarters of improvement still left adjusted gross margin at negative 62% in Q2; Line 2 was 1% of production, and Line 1 goes down for its move into early 2027 (S05, S14).

  1. 2026-11-05editedQ3 call read: the chief financial officer says Q3 and Q4 together are well over half of the way to the 72 points.
forward · 2026-11-05written 2026-10-06
C46inferencemedium confidence
US-made content is the moat Eos sells on, and it is shared: the tax rules reward any battery with little Chinese content, and Fluence already ships US-made lithium-ion modules on Tennessee cells, so the rules protect Eos from imported cells more than from US lithium-ion.
derivedThe statute's content tests (C28) read against Fluence's domestic-content offering (C23) and the US price premium tariffs created (C27).
against this

A zinc system needs no lithium, cobalt or Chinese cells at all, so it clears the 75% threshold of 2030 more easily than a lithium-ion supply chain still rebuilding (C28, C01).

current state · 2026-10-06written 2026-10-06
01

The company

Eos Energy builds the Z3, a battery block that stores electricity in zinc and salt water instead of lithium. Utilities and developers place rows of them beside solar farms, charge them when power is cheap and release it over four to twelve hours.

The usual alternative is lithium-ion from Tesla or a Chinese maker. Eos builds near Pittsburgh, and its water-based design does not burn. The cost is efficiency: it returns roughly 78 cents of each unit of power stored, by its own count, against more for lithium-ion.

The evidence under this - 4 claims
C01facthigh confidence
Eos makes the Z3, a battery module built on its Znyth aqueous zinc-halide chemistry, assembled into DC storage systems for 3 to 12 hours of discharge; it is made in Pennsylvania at Turtle Creek and at Thorn Hill in Warrendale, which began commercial production in June 2026. One segment; 787 employees at the end of 2025.
current state · 2026-06-30written 2026-10-06
C16management claimmedium confidence
Line 2 made 17% of Q3 battery output on a partial shift, up from about 1% in Q2, and reached first-pass yield parity with Line 1 within three months; Line 1 disassembly at Turtle Creek began in late October and the line is expected back near the end of the first quarter of 2027, giving Thorn Hill about 4 GWh of nameplate capacity on two lines.
against this

The DOE underwrote an 8 GWh four-line plan by 2027 in December 2024; the plan is now two lines and about 4 GWh, and Lines 3 and 4 need a new facility (S31, C34).

  1. 2026-11-05editedQ3 2026 results read from the 10-Q and release: Line 2 ramp and the Line 1 move
current state · 2026-11-05written 2026-10-06
C32management claimlow confidence
Management says the fleet averages about 78% round-trip efficiency, with the best cycles above 90%, and has run cycles from 2.5 to 14 hours.
against this

The only published test, written by Eos, averaged 75.2% on the earlier design, and no third party has published Z3 data (C31).

current state · 2026-08-07written 2026-10-06
C04facthigh confidence
Q3 2026 gross loss was $44.2M on $82.7M of revenue, a 53% negative margin; adjusted gross margin was negative 44%, against negative 62% in Q2, and adjusted EBITDA negative $63.8M against $71.4M. The nine months lost $137.4M at the gross line on $208.5M of revenue.
  1. 2026-11-05editedQ3 2026 results read from the 10-Q and release: Q3 gross loss and adjusted margin
historical · Q3 2026written 2026-10-06

How it makes money

Eos earns money by selling battery systems, booked when they are delivered, plus a sliver of service. In the first half of 2026, 55% came from outside buyers and 44% from one project sold to a related party.

It also collects a US manufacturing tax credit on every battery it makes. Even counting that credit, each dollar of sales cost about $1.71 to produce in June, so more sales have so far meant bigger losses.

The evidence under this - 4 claims
C42facthigh confidence
First-half 2026 revenue was $69.4M of product to outside customers, $55.0M of product to the related party and $1.3M of service; by destination, $96.5M was the US and $28.9M the UK.
historical · H1 2026written 2026-10-06
C03facthigh confidence
Q3 2026 revenue was $82.7M, of which $68.4M was related-party product revenue from projects financed by a Cerberus affiliate and contributed to Frontier Power USA, about 83% of the quarter; product revenue from everyone else was $14.3M, against $12.7M in Q2 and about $56.7M in Q1.
  1. 2026-11-05editedQ3 2026 results read from the 10-Q and release: Q3 related-party and outside product revenue
historical · Q3 2026written 2026-10-06
C04facthigh confidence
Q3 2026 gross loss was $44.2M on $82.7M of revenue, a 53% negative margin; adjusted gross margin was negative 44%, against negative 62% in Q2, and adjusted EBITDA negative $63.8M against $71.4M. The nine months lost $137.4M at the gross line on $208.5M of revenue.
  1. 2026-11-05editedQ3 2026 results read from the 10-Q and release: Q3 gross loss and adjusted margin
historical · Q3 2026written 2026-10-06
C28facthigh confidence
US tax law denies the storage investment credit to projects begun after 2025 with material assistance from a prohibited foreign entity, requiring 55% non-prohibited content in 2026 rising to 75% after 2029, adds a 10-point bonus for domestic content, and pays manufacturers $35 per kWh of cells and $10 per kWh of modules under § 45X until it phases out after 2032.
current state · 2026-10-06written 2026-10-06
Where the money comes from, stream by stream
StreamShare of sales (H1 2026)How it earns it
Product sold to outside customers55.2%Z3 storage systems for utilities and developers, booked when delivered.
Product sold to a related party43.8%A project financed by a Cerberus affiliate and now owned by Frontier Power USA, booked in June.
Service1.0%Maintenance and support on installed systems, booked over time.

Who runs it

Joe Mastrangelo, who ran GE's gas-turbine business, has led Eos since 2019. Four people have held the finance job in about fifteen months, and the sales chief leaves on 20 October.

Insiders own under 2%; Cerberus controls about 35% through preferred shares and warrants. Bonuses pay on orders, batteries shipped and a profit measure, and paid nothing for 2025. The chief executive bought shares with his own money in March.

The evidence under this - 4 claims
C37facthigh confidence
Joe Mastrangelo, formerly chief executive of GE's gas power business, has run Eos since 2019; the finance chief Eric Javidi was terminated without cause in May 2025 after about twelve weeks, Nathan Kroeker served as interim until Alessandro Lagi joined on 8 June 2026, and Kroeker, the commercial chief, leaves on 20 October 2026.
current state · 2026-09-16written 2026-10-06
C38facthigh confidence
Directors and officers owned 1.73% at 20 March 2026; the chief executive held 2.72M shares including 1.24M options. Cerberus held 31.98% then, BlackRock 7.6% and Vanguard 5.07% in later filings.
current state · 2026-03-20written 2026-10-06
C39facthigh confidence
The chief executive was paid $4.85M for 2025; the bonus paid on orders, cubes delivered and adjusted EBITDA and paid nothing, and long-term awards are half time-vesting shares, a quarter relative return against the Russell 2000 and a quarter undisclosed company metrics, none of which vested for 2025; 25% of votes cast opposed the 2026 pay vote.
historical · FY 2025written 2026-10-06
C40facthigh confidence
The chief executive bought 60,000 shares at $5.75 and 23,900 at $6.58 in March 2026 outside any plan; his and the accounting chief's September sales were made under 10b5-1 plans to cover tax on vested performance shares, and a new director bought 15,000 shares in September.
historical · March to September 2026written 2026-10-06
02

Why it wins, and who could beat it

Why customers come back

Eos's protection is the law, not the product. US tax rules cut off the main storage credit for projects with too many Chinese parts and pay extra for American content, and a zinc battery built in Pennsylvania passes easily. Safety helps too: a lithium-ion fire at Moss Landing in California damaged about 55% of some 100,000 battery modules.

The weakness is that the protection is shared. Fluence already sells American-made lithium-ion batteries that earn the same credits, so the rules mainly shield Eos from imports, not from rivals at home.

The evidence under this - 5 claims
C28facthigh confidence
US tax law denies the storage investment credit to projects begun after 2025 with material assistance from a prohibited foreign entity, requiring 55% non-prohibited content in 2026 rising to 75% after 2029, adds a 10-point bonus for domestic content, and pays manufacturers $35 per kWh of cells and $10 per kWh of modules under § 45X until it phases out after 2032.
current state · 2026-10-06written 2026-10-06
C46inferencemedium confidence
US-made content is the moat Eos sells on, and it is shared: the tax rules reward any battery with little Chinese content, and Fluence already ships US-made lithium-ion modules on Tennessee cells, so the rules protect Eos from imported cells more than from US lithium-ion.
derivedThe statute's content tests (C28) read against Fluence's domestic-content offering (C23) and the US price premium tariffs created (C27).
against this

A zinc system needs no lithium, cobalt or Chinese cells at all, so it clears the 75% threshold of 2030 more easily than a lithium-ion supply chain still rebuilding (C28, C01).

current state · 2026-10-06written 2026-10-06
C33facthigh confidence
The EPA says about 55% of roughly 100,000 lithium-ion modules were damaged in the January 2025 fire at Vistra's 300 MW Moss Landing site, and expects removal and demolition to finish in mid-2027.
historical · January 2025 to September 2026written 2026-10-06
C23facthigh confidence
Fluence had 6.8 GW and 17.8 GWh of storage deployed and 9.1 GW of contracted backlog at 30 September 2025, and builds modules in Utah with cells made in Tennessee for a US domestic-content offering.
current state · 2025-09-30written 2026-10-06
C27facthigh confidence
Lazard puts the capital cost of a US 100 MW, 4-hour standalone system at $270 to $365 per kWh in July 2026 and says tariffs on imported lithium-ion cells reversed earlier cost declines; BloombergNEF put the average stationary pack at $70/kWh in 2025, and Ember about $125/kWh for a whole 4-hour-plus system outside the US and China.
current state · 2026-07-01written 2026-10-06

The threat

Lithium-ion is winning. Tesla installed 22.3 GWh of storage in six months, and Wood Mackenzie, an industry researcher, says lithium-ion has taken the four-to-eight-hour market Eos sells into and should keep 85% of all storage through 2034. The other listed non-lithium maker Eos names, ESS Tech, has warned it may not survive.

The evidence under this - 4 claims
C24facthigh confidence
Tesla deployed 22.3 GWh of energy storage in the first half of 2026.
historical · H1 2026written 2026-10-06
C26facthigh confidence
Wood Mackenzie says lithium-ion has captured the four-to-eight-hour storage market on cost and supply chain and expects it to hold 85% share through 2034; long-duration storage was 6% of global installations in 2025, a 4-hour lithium-ion project costs $107/kWh in China, and Eos is one of three long-duration companies to have raised over $1B.
current state · 2026-03-09written 2026-10-06
C23facthigh confidence
Fluence had 6.8 GW and 17.8 GWh of storage deployed and 9.1 GW of contracted backlog at 30 September 2025, and builds modules in Utah with cells made in Tennessee for a US domestic-content offering.
current state · 2025-09-30written 2026-10-06
C25facthigh confidence
ESS Tech, the listed iron-flow rival Eos names, reported substantial doubt about continuing as a going concern, $10.8M of cash at 30 June 2026, and that its near-term revenue will come mainly from sodium-ion products.
current state · 2026-06-30written 2026-10-06

The rivals: TSLA, FLNC, GWH, and 1 more that is private, in-house or listed abroad.

Who else does this, rival by rival
TickerRivalHow far alongPriced against it
TSLATesla MegapackPriced against it: Shown, not countedOn saleShown, not counted
FLNCFluence, with US-made modules on Tennessee cellsPriced against it: YesOn saleYes
PrivateForm Energy's iron-air batteryPriced against it: PrivateOn salePrivate
GWHESS Tech's iron-flow Energy WarehousePriced against it: YesOn saleYes
03

How big it could get

How big this could get

American batteries are being added fast: 43.6 GW were running by the end of 2025, with 26 GW more planned for 2027, and about 500 GWh expected from 2026 to 2031. Eos's two Pennsylvania lines are built for about 4 GWh a year.

What nobody measures is the part Eos is built for: storage that runs eight hours or longer. Worldwide, that long-duration slice was only 6% of installations in 2025.

The evidence under this - 4 claims
C29facthigh confidence
The EIA counts 43.6 GW of US battery storage at the end of 2025 and 8.3 GW added in the first half of 2026, with 26 GW planned for 2027; Wood Mackenzie expects half a terawatt-hour of US storage installed from 2026 to 2031.
current state · 2026-08-07written 2026-10-06
C26facthigh confidence
Wood Mackenzie says lithium-ion has captured the four-to-eight-hour storage market on cost and supply chain and expects it to hold 85% share through 2034; long-duration storage was 6% of global installations in 2025, a 4-hour lithium-ion project costs $107/kWh in China, and Eos is one of three long-duration companies to have raised over $1B.
current state · 2026-03-09written 2026-10-06
C47inferencemedium confidence
Eos's distinct ground is eight hours and longer, where a lithium-ion system's cost keeps rising with every added hour; from four to eight hours the outside measurers say lithium-ion has already won, and the government calls anything under ten hours short duration.
derivedWood Mackenzie's reading of the 4-to-8-hour market (C26), Ember's falling savings per kWh beyond four hours (C27), the DOE taxonomy (C30) and Eos's own pipeline mix (C12).
against this

Management says over half its pipeline is eight hours or longer and cites Virginia's 10-hour carve-out (C12, C35), so the longer-duration market may be larger than the outside data show.

current state · 2026-10-06written 2026-10-06
C16management claimmedium confidence
Line 2 made 17% of Q3 battery output on a partial shift, up from about 1% in Q2, and reached first-pass yield parity with Line 1 within three months; Line 1 disassembly at Turtle Creek began in late October and the line is expected back near the end of the first quarter of 2027, giving Thorn Hill about 4 GWh of nameplate capacity on two lines.
against this

The DOE underwrote an 8 GWh four-line plan by 2027 in December 2024; the plan is now two lines and about 4 GWh, and Lines 3 and 4 need a new facility (S31, C34).

  1. 2026-11-05editedQ3 2026 results read from the 10-Q and release: Line 2 ramp and the Line 1 move
current state · 2026-11-05written 2026-10-06

Who has signed

The biggest signed order is 2 GWh from Frontier Power USA, a venture Eos owns 36% of and Cerberus runs; its price and length are not public. Beyond it sit a 750 MWh supply deal in central Europe and a 100 MWh Google-linked project due online in 2030.

The evidence under this - 4 claims
C11facthigh confidence
Frontier Power USA and Eos signed a 2 GWh purchase commitment and capacity reservation agreement dated 11 May 2026, which the company calls firm; its price, term, deposit and cancellation terms are not filed. The related-party project and Frontier together were 49% of backlog volume at 30 June 2026.
current state · 2026-06-30written 2026-10-06
C10facthigh confidence
Eos put $112.6M into Frontier Power USA, the net proceeds of July's $75.0M direct offering and August's rights offering, which sold 6,885,218 of 27,367,171 units against a roughly $150M target. Eos holds 112.6M of 312.6M units, about 36%; Cerberus holds 150.0M, including 50.0M founder units for no cash, and appoints four of seven managers; in liquidation Cerberus's and Hudson Bay's cash comes back before Eos's.
current state · 2026-08-04written 2026-10-06
C44management claimlow confidence
Management cites a strategic partnership under the Golden Dome defence programme, a 750 MWh master supply agreement with CAPAC for Germany, Austria and Switzerland, and a $100M purchase order for Blanquilla under Frontier's 2 GWh agreement.
against this

None of the three counterparties' own filings or releases was found, and a master supply agreement fixes terms rather than volumes (S01).

current state · 2026-08-07written 2026-10-06
C36facthigh confidence
The first project under MN8's supply agreement pairs 10 MW and 100 MWh of Eos storage at ten hours with 70 MW and 280 MWh of lithium-ion beside a solar farm serving Google's data centres; the Eos part is due to follow in 2030.
current state · 2026-09-02written 2026-10-06
04

The results

The numbers

Is it growing?Fast

Sales grew 320% over the twelve months to Sep 2026, slower than the 632% in the year to Dec 2025.

Dec 2021: $4.60MDec 2022: $17.9MDec 2023: $16.4MDec 2024: $15.6MDec 2025: $114M12 mo to Sep 2026: $266M$4.60M$266M
Is it profitable?Losing

Operating margin is −121%, up from −934% in the year to Dec 2023.

Dec 2021: −29×Dec 2022: −12×Dec 2023: −934%Dec 2024: −11×Dec 2025: −227%12 mo to Sep 2026: −121%−29×−121%
Does growth pay off?Pays off

Running costs take 49¢ of each sales dollar, down from 2019¢ in the year to Dec 2021.

Dec 2021: 2019¢Dec 2022: 479¢Dec 2023: 486¢Dec 2024: 589¢Dec 2025: 101¢12 mo to Sep 2026: 49¢2019¢49¢
How steady is it?Swingy

Yearly sales growth has ranged from −8.6% to +20× over the last 5 years, with sales falling in 2 of them.

Dec 2021: +20×Dec 2022: +290%Dec 2023: −8.6%Dec 2024: −4.7%Dec 2025: +632%+20×+632%
How long can it keep going?Short

It used $443M of cash over the last twelve months, so the $271M it holds lasts about 0.6 years at that pace.

Dec 2021: −$132MDec 2022: −$217MDec 2023: −$174MDec 2024: −$187MDec 2025: −$265M12 mo to Sep 2026: −$443M−$132M−$443M
Is the balance sheet safe?Strained

It owes $385M more than the cash it holds, and its operations bring in no cash to repay it.

Dec 2021: −$4.38MDec 2022: −$164MDec 2023: −$138MDec 2024: −$243MDec 2025: −$245M12 mo to Sep 2026: −$385M−$4.38M−$385M
Are shares being diluted?Diluting

The share count rose 336% in 2.8 years, and pay in shares costs 9.1% of sales.

Dec 2023: 127MDec 2024: 212MDec 2025: 261M12 mo to Sep 2026: 553M127M553M
All the numbers

Filed figures from the company's 10-K and 10-Q, read 2026-11-10. Fiscal years end in December.

Dec 2021Dec 2022Dec 2023Dec 2024Dec 202512 mo to Sep 2026
Size and growth
Sales$4.60M$17.9M$16.4M$15.6M$114M$266M
Sales growth+20×+290%−8.6%−4.7%+632%+320%
Profit
Gross margin−911%−755%−448%−534%−126%−72%
Operating margin−29×−12×−934%−11×−227%−121%
Net margin−27×−13×−14×−44×−849%89%
What growth costs, per sales dollar
Research and development417%103%114%146%25%15%
General and admin935%338%328%---
Pay in shares327%77%86%120%22%9.1%
Capital spending339%112%179%212%47%49%
Cash
Cash from operations−$116M−$197M−$145M−$154M−$211M−$312M
Free cash flow−$132M−$217M−$174M−$187M−$265M−$443M
Balance sheet
Cash and short-term investments$105M$17.1M$69.5M$74.3M$568M$271M
Debt$109M$181M$207M$317M$813M$656M
Net cash−$4.38M−$164M−$138M−$243M−$245M−$385M
Returns and shares
Shares (diluted)--127M212M261M553M

What growth costs

Running costs are being spread thinner: research, selling and office costs fell from about 99 cents per sales dollar in 2025 to 57 cents over the last year. That helps only once the batteries themselves sell for more than they cost, and they do not yet.

The evidence under this - 2 claims
C43facthigh confidence
Research, selling and administration cost $113.6M against $114.2M of 2025 revenue, about 99 cents per sales dollar, and $122.9M against $214.2M over the year to June 2026, about 57 cents, while the gross line still lost money on every sale.
historical · FY 2025 and the year to June 2026written 2026-10-06
C04facthigh confidence
Q3 2026 gross loss was $44.2M on $82.7M of revenue, a 53% negative margin; adjusted gross margin was negative 44%, against negative 62% in Q2, and adjusted EBITDA negative $63.8M against $71.4M. The nine months lost $137.4M at the gross line on $208.5M of revenue.
  1. 2026-11-05editedQ3 2026 results read from the 10-Q and release: Q3 gross loss and adjusted margin
historical · Q3 2026written 2026-10-06

The last report (5 Nov)

Quarter ended September 2026, filed figures.

20.3%
Growth from the quarter before
against 20.7% the quarter before that
-96¢
Operating profit
per dollar of sales

In Q3 2026 Eos reported record sales of $82.7M. Again most came from a project financed by Cerberus and handed to Frontier: $68.4M, or 83%. Sales to everyone else were $14.3M, up from $12.7M.

It still lost 53 cents at the gross line on each sales dollar, better than 71 cents last quarter. It expects the lower half of its $300M to $350M sales forecast, as moving a production line cuts fourth-quarter capacity. Orders on its books fell to $765M. On the call it said cash used tracks its losses; it gave no 2027 figures.

What it did to the thesisThe margin improved, but growth still rests on the lender's venture: outside sales barely moved, the forecast slipped and the battery still sells below cost.

The quarter in detail

Sales growth from the previous quarter, filed figures.

Earnings per share $-0.17 against $-0.24 expected; $-0.17 as filed, with no adjustments.

The evidence under this - 6 claims
C03facthigh confidence
Q3 2026 revenue was $82.7M, of which $68.4M was related-party product revenue from projects financed by a Cerberus affiliate and contributed to Frontier Power USA, about 83% of the quarter; product revenue from everyone else was $14.3M, against $12.7M in Q2 and about $56.7M in Q1.
  1. 2026-11-05editedQ3 2026 results read from the 10-Q and release: Q3 related-party and outside product revenue
historical · Q3 2026written 2026-10-06
C04facthigh confidence
Q3 2026 gross loss was $44.2M on $82.7M of revenue, a 53% negative margin; adjusted gross margin was negative 44%, against negative 62% in Q2, and adjusted EBITDA negative $63.8M against $71.4M. The nine months lost $137.4M at the gross line on $208.5M of revenue.
  1. 2026-11-05editedQ3 2026 results read from the 10-Q and release: Q3 gross loss and adjusted margin
historical · Q3 2026written 2026-10-06
C14management claimmedium confidence
Management now expects 2026 revenue in the lower half of its $300M to $350M range, saying the move of Line 1 into Thorn Hill from late October will temporarily lower capacity and delay some shipments; nine-month revenue is $208.5M, so the lower half needs $91.5M to $116.5M in the fourth quarter.
against this

First-half revenue was $125.7M, so even the low end needs $174M in six months, more than any half-year Eos has reported, and Q4 2025 came in $35.8M under consensus (S01, C17).

  1. 2026-11-05editedQ3 2026 results read from the 10-Q and release: the outlook moved to the lower half of the range
forward · 2026-11-05written 2026-10-06
C12management claimmedium confidence
Management reports backlog of $765M and 3.2 GWh at 30 September 2026, down 5% in the quarter as record shipments were only partly replaced by six new orders, including a $100M order that converted from an existing reservation, and a $24.9B pipeline.
against this

Remaining performance obligations in the 10-Q were $34.0M at 30 June, down from $45.8M in December, and half the backlog's volume is the lender's venture (S01, C11).

  1. 2026-11-05editedQ3 2026 results read from the 10-Q and release: backlog at 30 September
current state · 2026-09-30written 2026-10-06
C15management claimlow confidence
Management says scrap, labour, materials and Thorn Hill's ramp give a path to over 72 points of adjusted gross margin improvement over the twelve months from August 2026, assuming the plan and expected volumes, which would take it from negative 62% to about positive 10%. On the Q3 call the chief financial officer said the third and fourth quarters together are well over half of the way to the 72 points, with materials cost down and Thorn Hill cutting conversion cost.
against this

Seven straight quarters of improvement still left adjusted gross margin at negative 62% in Q2; Line 2 was 1% of production, and Line 1 goes down for its move into early 2027 (S05, S14).

  1. 2026-11-05editedQ3 call read: the chief financial officer says Q3 and Q4 together are well over half of the way to the 72 points.
forward · 2026-11-05written 2026-10-06
C49estimatemedium confidence
At the first half's rate of about $131M a quarter, the $271.2M of unrestricted cash at September, which already holds the $87.0M DOE advance, lasts about two quarters, to around the spring of 2027, before the margin turn management promises would have time to show. On the Q3 call management said cash used has tracked the operating loss with no working-capital drain, capital spending moderates from the fourth quarter, and it will give 2027 figures on the Q4 call.
derivedC05's first-half outflow carried forward against the 30 September unrestricted cash; the third quarter's own outflow is distorted by the Frontier investment and equity raised (S35).
against this

Customer deposits ran $51.0M in Q2, 45X credits come in as cash, and Cerberus's $105M revolver is available at its discretion, so the outflow could be smaller (S01).

  1. 2026-11-05editedQ3 2026 results read from the 10-Q and release: cash runway re-struck on the Q3 balance
  2. 2026-11-05editedQ3 call read: management says cash burn tracks the loss, capital spending falls from Q4, and declined a 2027 figure.
forward · 2026-11-05written 2026-10-06
What management did with the money

Eos has never paid a dividend or bought back shares. It funded losses by selling shares at prices from 87 cents to $12.78, by borrowing, and by paying $564.6M to retire $200M of notes early. The money it raised in July and August went straight into Frontier.

Insiders in 2026Q3: 41 purchases and 25 sales, from their Form 4 filings.

The evidence under this - 2 claims
C41facthigh confidence
Eos has never paid a dividend or bought back stock; it raised $458.2M selling shares at $12.78 in November 2025 and $600M of convertible notes, paid $564.6M to settle $200M of its 2030 notes early with a $63.5M inducement charge, and sold shares at $5.481 in July 2026.
historical · 2025 to August 2026written 2026-10-06
C10facthigh confidence
Eos put $112.6M into Frontier Power USA, the net proceeds of July's $75.0M direct offering and August's rights offering, which sold 6,885,218 of 27,367,171 units against a roughly $150M target. Eos holds 112.6M of 312.6M units, about 36%; Cerberus holds 150.0M, including 50.0M founder units for no cash, and appoints four of seven managers; in liquidation Cerberus's and Hudson Bay's cash comes back before Eos's.
current state · 2026-08-04written 2026-10-06
05

The price, and where we disagree

In short. Priced for perfection. The price is about 501% of what we think the business is worth, at an 8% yearly return.

Share price
$2.14
close,
Company value for each $1 of sales
4.5
rivals' middle 1.6 · its 6-year normal 55.0

The shares have fallen about 35% in a month, to $2.14, after two disappointing reports. Nearly 40% of the shares available to trade are sold short, a bet by others that they fall further.

Three dated tests are close. From late December Cerberus and Hudson Bay can sell shares they hold. The line move into Thorn Hill runs to the end of the first quarter of 2027. And the fourth-quarter report, around March, shows whether outside sales and the margin recover.

The evidence under this - 5 claims
C19facthigh confidence
The shares closed at $2.14 on 4 November 2026, 89% under the $19.86 high and 35% above the 52-week low of $1.59, down 35% since the 5 October close of $3.27; 39% of the float was short at the 15 October settlement, and four analysts rate it a buy and seven a hold.
  1. 2026-11-05editedre-struck at the 4 November close after the 35% fall
current state · 2026-11-04written 2026-10-06
C17facthigh confidence
Q4 2025 revenue of $58.0M against a $93.75M consensus sent the shares down 39.4% on 26 February 2026, the capacity milestone was reached five weeks late, and a securities class action filed on 6 March 2026 alleges misleading statements about manufacturing capabilities and the financial outlook.
historical · Q4 2025 to March 2026written 2026-10-06
C45facthigh confidence
From 31 December 2026 Hudson Bay can exchange its 50M Frontier units for up to 9.1M Eos shares at $5.481, and Cerberus's and Hudson Bay's securities are registered for resale, with Cerberus's lock-up ending on 21 December 2026.
current state · 2026-08-31written 2026-10-06
C14management claimmedium confidence
Management now expects 2026 revenue in the lower half of its $300M to $350M range, saying the move of Line 1 into Thorn Hill from late October will temporarily lower capacity and delay some shipments; nine-month revenue is $208.5M, so the lower half needs $91.5M to $116.5M in the fourth quarter.
against this

First-half revenue was $125.7M, so even the low end needs $174M in six months, more than any half-year Eos has reported, and Q4 2025 came in $35.8M under consensus (S01, C17).

  1. 2026-11-05editedQ3 2026 results read from the 10-Q and release: the outlook moved to the lower half of the range
forward · 2026-11-05written 2026-10-06
C15management claimlow confidence
Management says scrap, labour, materials and Thorn Hill's ramp give a path to over 72 points of adjusted gross margin improvement over the twelve months from August 2026, assuming the plan and expected volumes, which would take it from negative 62% to about positive 10%. On the Q3 call the chief financial officer said the third and fourth quarters together are well over half of the way to the 72 points, with materials cost down and Thorn Hill cutting conversion cost.
against this

Seven straight quarters of improvement still left adjusted gross margin at negative 62% in Q2; Line 2 was 1% of production, and Line 1 goes down for its move into early 2027 (S05, S14).

  1. 2026-11-05editedQ3 call read: the chief financial officer says Q3 and Q4 together are well over half of the way to the 72 points.
forward · 2026-11-05written 2026-10-06
The betPriced for success
If it works
0.4×
from today's value, valued like the cheaper of its two rivals
If it fails
keep up to 0%
of today's market value: at most, with what is left counted at book value
How we value it
Method
Cash the business will produce, with any shortfall raised in new shares
Required return
8%
Long-run growth
4%
Moat
No moat
BearBaseBull
Sales growth, years 1 to 330%, 30%, 19%60%, 60%, 37.9%75%, 75%, 47.4%
Normal operating margin5.1%6.4%7.1%
Return on new capital after year ten8%8%11%
Value a share$0.17$0.36$0.63
Price
$1.74
Priced on
6 Nov
Price as a share of the base value
483%
Yearly return the price offers on the base case
5.1%
Risk check: valued at 12%, the rate for a business not yet making a profit
$0.17

Where each figure comes from

FigureSourceWhere exactly
Sales, the last twelve monthsFiled reportQuarterly report (10-Q) for the period to 30 Sep 2026, filed 5 Nov 2026; the four quarters to that date.
Operating margin todayFiled reportQuarterly report (10-Q) for the period to 30 Sep 2026, filed 5 Nov 2026; the four quarters to that date.
Tax rateOur judgementOur reading of the evidence.
Sales per dollar of new capitalFiled reportAnnual report (10-K) for the period to 31 Dec 2025, filed 26 Feb 2026: the year's sales over its operating capital.
SharesData vendorFMP's count of listed shares, since the filed diluted count is on another basis (a split since, a second class or a depositary ratio).
Cash and investmentsFiled reportQuarterly report (10-Q) for the period to 30 Sep 2026, filed 5 Nov 2026; the balance sheet on that date.
DebtFiled reportQuarterly report (10-Q) for the period to 30 Sep 2026, filed 5 Nov 2026; the balance sheet on that date.
Sales growth, years 1 to 3Analysts' forecastsFMP's consensus for the years to 31 Dec 2026, 31 Dec 2027 and 31 Dec 2028, as it stood when the value was struck on 10 Nov 2026: the average of 6 to 8 analysts' revenue estimates, 60%, 60%, 37.9%, each year held to at most 60%. FMP publishes how many analysts, not who they are.
Normal operating marginAnalysts' forecastsFMP's consensus operating margin for the year to 31 Dec 2028, as it stood when the value was struck on 10 Nov 2026, the average of 8 analysts' estimates, less stock pay: 6.4% in the base case. FMP publishes how many analysts, not who they are.
Return on new capital after year tenOur method's ruleSet by our reading of the moat: the 8% rate with none, 12% with an average one, 15% with a strong one; the bear case earns the rate.
The bear and bull casesOur method's ruleOur method's spread: the bear case takes half the growth and a fifth off the margin, the bull a quarter more growth and a tenth more margin.
PriceMarket priceFMP's closing price on 6 Nov 2026.

What each case rests on

Bear caseGrowth runs at half the analysts' path, 30%, 30% and 19%, the margin settles a fifth lower at 5.1%, and new capital earns only the 8% rate: the downturn test.

Base caseRevenue grows 60%, 60% and 38% over three years, the analysts' path (C52), held to at most 60% a year, then fades toward 4%, closing 30% of the gap each year; the margin moves to 6.4%, the analysts' year-three figure after stock pay.

Bull caseGrowth runs a quarter above the analysts' path, 75%, 75% and 47%, the margin reaches 7.1%, and new capital earns 11%.

What moves the base value

LowerHigher
Growth in years 1 to 3, a quarter lower or higher$0.30$0.42
The normal margin, 15% lower or higher$0.30$0.42
Capital per dollar of growth, a quarter more or less$0.32$0.39
Return on new capital after year ten, at the rate or one step up$0.36$0.47
The rate, a point higher or lower$0.26$0.53
Long-run growth a point lower$0.35-

Value a share as growth and margin change

Normal margin ↓ · growth →50%75%100%125%150%
5.1%$0.17$0.23$0.28$0.33$0.38
5.8%$0.20$0.26$0.32$0.38$0.43
6.4%$0.23$0.30$0.36$0.42$0.48
7%$0.26$0.33$0.40$0.47$0.53
7.7%$0.29$0.36$0.44$0.52$0.58

Growth is shown as a share of the base case's. Green: the price is at or under 80% of that value. Red: at or over 120%.

How we read the multiple

Counting its debt at face value and every share in issue, Eos is valued at about $1.55B, against $266M of sales over the last year. It has never sold a battery for more than it cost to make.

For the price to work:

  1. Its listed rivals are valued at about 1.7 times their yearly sales.
  2. At that rate, $1.55B needs about $0.91B of yearly sales.
  3. That is between analysts' 2027 and 2028 forecasts.
  4. And the battery must first sell for more than it costs, which it has never done.

Against its own history. It has rarely had enough sales for its own past to be a yardstick, so a lower figure than before says the hope has shrunk, not that it is cheap.

The evidence under this - 5 claims
C21inferencemedium confidence
At $2.14 the company value is about $1.55B on filed figures (365.8M shares, $1,040M of debt principal, less $271.2M of cash), 5.8 times trailing revenue of $266.5M and 5.0 times the $312M midpoint of the lower half of the 2026 guide; the vendor's figure is 4.5 times. At the filed rivals' median of 1.7 times sales, $1.55B needs about $0.91B of revenue, between the 2027 and 2028 consensus, and that before the gross margin has turned positive.
derivedArithmetic on filed figures (C05, C06, C08) at the pack's close (C19), against the pack's cohort median (C22) and consensus (C20).
against this

Counting the preferred and penny warrant as equity raises the figure to about $2.4B, and the cohort is two companies, one near insolvency, so the median is a thin yardstick (C07, C22, C25).

  1. 2026-11-05editedre-struck at the 4 November close on the Q3 balance sheet
current state · 2026-11-04written 2026-10-06
C20estimatemedium confidence
Analysts expect revenue of $305M in 2026 on four estimates, $572M in 2027 and $1.06B in 2028 on six each.
derivedThe pack's analyst consensus, read 5 November 2026 (S37), before the Q3 print is in it.
against this

Management's own 2026 range is $300M to $350M and the low end needs a record half (C14); 2028's figure is four times Eos's best annualised output.

  1. 2026-11-05editedconsensus re-read on 5 November, before analysts reflect the print
forwardwritten 2026-10-06
C22facthigh confidence
On company value to sales Eos is the dearest of its filed cohort at 4.5 times on the vendor's figures, against Fluence at 0.5 and ESS Tech at 2.8, a median of 1.65; Tesla, whose storage arm is a small part of it, is printed beside them at 14.7.
  1. 2026-11-05editedre-struck at the 4 November close
current state · 2026-11-04written 2026-10-06
C05facthigh confidence
Cash, equivalents and restricted cash were $344.1M at 30 September 2026, of which $271.2M was unrestricted, against $624.6M at the end of 2025; the nine months used $261.6M in operations and $219.0M in investing, including the investment in Frontier Power USA, and raised $200.1M, including the $87.0M DOE advance.
  1. 2026-11-05editedQ3 2026 results read from the 10-Q and release: cash at 30 September
current state · 2026-09-30written 2026-10-06
C04facthigh confidence
Q3 2026 gross loss was $44.2M on $82.7M of revenue, a 53% negative margin; adjusted gross margin was negative 44%, against negative 62% in Q2, and adjusted EBITDA negative $63.8M against $71.4M. The nine months lost $137.4M at the gross line on $208.5M of revenue.
  1. 2026-11-05editedQ3 2026 results read from the 10-Q and release: Q3 gross loss and adjusted margin
historical · Q3 2026written 2026-10-06
Priced against its rivals

The competitors its own annual report names, each confirmed against the rival's filing. Measured on company value for each $1 of sales, cheapest first:

TickerCompanycompany value for each $1 of sales
FLNCFluence Energy0.5×
GWHESS Tech2.8×
EOSEEos Energy Enterprises4.5×
TSLATesla not in the median14.7×

The middle of the other 2 is 1.6. Tesla, Inc. competes with it in one product line inside a much bigger company, so its price says little about this business and it is not counted in the middle. Eos Energy Enterprises, Inc. is dearest of 3.

Rivals' figures come from a data provider, so read the spread of the group rather than any single pair.

The rivals are cheaper because they are a larger integrator and a struggling flow-battery maker; the higher price is Eos's own, not the group's.

Private or foreign rivals that cannot be priced: Panasonic, Samsung Electronics Co. Ltd., LG Chem Ltd., BYD, Sungrow, Contemporary Amperex Technology Co. Limited, Enervenue, Ambri, Form Energy.

The bet: what it pays if it works

This is a bet that Eos's new Pennsylvania factory runs full and sells its batteries at a profit. Even then, valued like its listed rivals, the company would be worth only a little more than today, and much less at the lower rival's rate.

If the margin never turns, the cash runs out and the lenders, owed far more than the company owns, take the plants. Shareholders would keep little or nothing.

Revenue in two to three yearsGWh delivered a year neededValued like its peersLike the cheapest
$573.00Mabout 2 at 100% kept0.8×0.2×
$950.00M (the big win)about 4 at 100% kept1.3×0.4×
$1.07Babout 5 at 100% kept1.5×0.4×

To pay 2× (a Fair bet) it would need about $4.76B of revenue (about 20 GWh delivered a year); to pay 5× (a Cheap bet), about $11.89B of revenue (about 50 GWh delivered a year), at the same multiple.

The steps it has to take

  • Done Thorn Hill Line 2 in commercial production
  • Done Frontier Power USA closed
  • Done First DOE advance under Tranche 2
  • Slipped 2025 capacity milestone due 2025-12-31
  • In progress 2026 revenue of $300M to $350M due 2026-12-31
  • Not started Line 1 moved to Thorn Hill due 2027-03-31
  • In progress Adjusted gross margin positive due 2027-08-31
The detail behind the label
Rating
Expensive
the big win's payoff against the 2× and 5× lines
At their midpoint
1.3×
the same success, valued between the two
Cheap bet
not on price
the big win pays five times only below the company's debt
The evidence under this - 6 claims
C16management claimmedium confidence
Line 2 made 17% of Q3 battery output on a partial shift, up from about 1% in Q2, and reached first-pass yield parity with Line 1 within three months; Line 1 disassembly at Turtle Creek began in late October and the line is expected back near the end of the first quarter of 2027, giving Thorn Hill about 4 GWh of nameplate capacity on two lines.
against this

The DOE underwrote an 8 GWh four-line plan by 2027 in December 2024; the plan is now two lines and about 4 GWh, and Lines 3 and 4 need a new facility (S31, C34).

  1. 2026-11-05editedQ3 2026 results read from the 10-Q and release: Line 2 ramp and the Line 1 move
current state · 2026-11-05written 2026-10-06
C34facthigh confidence
The DOE loan guarantee is for up to $303.5M, $277.5M of principal plus capitalised interest, funding 80% of eligible line costs; about $178M was drawn after a $87.0M advance on 10 September 2026, and the last two tranches need a new facility the DOE approves.
current state · 2026-09-10written 2026-10-06
C12management claimmedium confidence
Management reports backlog of $765M and 3.2 GWh at 30 September 2026, down 5% in the quarter as record shipments were only partly replaced by six new orders, including a $100M order that converted from an existing reservation, and a $24.9B pipeline.
against this

Remaining performance obligations in the 10-Q were $34.0M at 30 June, down from $45.8M in December, and half the backlog's volume is the lender's venture (S01, C11).

  1. 2026-11-05editedQ3 2026 results read from the 10-Q and release: backlog at 30 September
current state · 2026-09-30written 2026-10-06
C50facthigh confidence
Total assets were $906.8M at 30 June 2026, including $177.5M of plant, against $1,224.3M of liabilities and $713.2M of preferred stock in mezzanine, leaving a $1,030.9M shareholders' deficit.
current state · 2026-06-30written 2026-10-06
C06facthigh confidence
Debt principal was $1,040.1M at 30 September 2026 against a carrying value of $655.8M: Cerberus's term loan at $205.0M paying 7% in kind, $600M of 1.75% notes converting at about $16.16 due 2031, $50M of 6.75% notes converting at about $5.06 due 2030, and $185.1M of DOE loan; a springing maturity could bring the Cerberus and DOE loans due on 14 March 2030. Only a $15M minimum-liquidity covenant applies until the EBITDA and revenue tests return on 31 March 2027.
  1. 2026-11-05editedQ3 2026 results read from the 10-Q and release: debt principal at 30 September
current state · 2026-09-30written 2026-10-06
C49estimatemedium confidence
At the first half's rate of about $131M a quarter, the $271.2M of unrestricted cash at September, which already holds the $87.0M DOE advance, lasts about two quarters, to around the spring of 2027, before the margin turn management promises would have time to show. On the Q3 call management said cash used has tracked the operating loss with no working-capital drain, capital spending moderates from the fourth quarter, and it will give 2027 figures on the Q4 call.
derivedC05's first-half outflow carried forward against the 30 September unrestricted cash; the third quarter's own outflow is distorted by the Frontier investment and equity raised (S35).
against this

Customer deposits ran $51.0M in Q2, 45X credits come in as cash, and Cerberus's $105M revolver is available at its discretion, so the outflow could be smaller (S01).

  1. 2026-11-05editedQ3 2026 results read from the 10-Q and release: cash runway re-struck on the Q3 balance
  2. 2026-11-05editedQ3 call read: management says cash burn tracks the loss, capital spending falls from Q4, and declined a 2027 figure.
forward · 2026-11-05written 2026-10-06

Where we disagree with the market

The market believes

A US-made, fireproof battery maker about to scale, with orders up 25% in a quarter and margins about to turn as the new factory ramps.

We believe

Outside demand is thinner than the orders on its books suggest, and the margin turn is a promise, not yet a result.

The market reads the orders on Eos's books and the projects it is bidding for. The filings show half the order volume is the lender's venture, and the contracts Eos must report under accounting rules were $34.0M at June.

We would be wrong if third-quarter sales to outside buyers top $40M while the margin improves, or if Frontier's order is filed with a price and term. A large short position shows we are not alone.

What analysts say, and the evidence
Analysts positive
36%
of 11, 1 Nov
Four months earlier
36%
of 11, 1 Jul
The evidence under this - 6 claims
C12management claimmedium confidence
Management reports backlog of $765M and 3.2 GWh at 30 September 2026, down 5% in the quarter as record shipments were only partly replaced by six new orders, including a $100M order that converted from an existing reservation, and a $24.9B pipeline.
against this

Remaining performance obligations in the 10-Q were $34.0M at 30 June, down from $45.8M in December, and half the backlog's volume is the lender's venture (S01, C11).

  1. 2026-11-05editedQ3 2026 results read from the 10-Q and release: backlog at 30 September
current state · 2026-09-30written 2026-10-06
C13facthigh confidence
Remaining performance obligations were $34.0M at 30 June 2026 against $45.8M at the end of 2025; the filed figure leaves out contracts satisfied within a year.
current state · 2026-06-30written 2026-10-06
C03facthigh confidence
Q3 2026 revenue was $82.7M, of which $68.4M was related-party product revenue from projects financed by a Cerberus affiliate and contributed to Frontier Power USA, about 83% of the quarter; product revenue from everyone else was $14.3M, against $12.7M in Q2 and about $56.7M in Q1.
  1. 2026-11-05editedQ3 2026 results read from the 10-Q and release: Q3 related-party and outside product revenue
historical · Q3 2026written 2026-10-06
C21inferencemedium confidence
At $2.14 the company value is about $1.55B on filed figures (365.8M shares, $1,040M of debt principal, less $271.2M of cash), 5.8 times trailing revenue of $266.5M and 5.0 times the $312M midpoint of the lower half of the 2026 guide; the vendor's figure is 4.5 times. At the filed rivals' median of 1.7 times sales, $1.55B needs about $0.91B of revenue, between the 2027 and 2028 consensus, and that before the gross margin has turned positive.
derivedArithmetic on filed figures (C05, C06, C08) at the pack's close (C19), against the pack's cohort median (C22) and consensus (C20).
against this

Counting the preferred and penny warrant as equity raises the figure to about $2.4B, and the cohort is two companies, one near insolvency, so the median is a thin yardstick (C07, C22, C25).

  1. 2026-11-05editedre-struck at the 4 November close on the Q3 balance sheet
current state · 2026-11-04written 2026-10-06
C11facthigh confidence
Frontier Power USA and Eos signed a 2 GWh purchase commitment and capacity reservation agreement dated 11 May 2026, which the company calls firm; its price, term, deposit and cancellation terms are not filed. The related-party project and Frontier together were 49% of backlog volume at 30 June 2026.
current state · 2026-06-30written 2026-10-06
C19facthigh confidence
The shares closed at $2.14 on 4 November 2026, 89% under the $19.86 high and 35% above the 52-week low of $1.59, down 35% since the 5 October close of $3.27; 39% of the float was short at the 15 October settlement, and four analysts rate it a buy and seven a hold.
  1. 2026-11-05editedre-struck at the 4 November close after the 35% fall
current state · 2026-11-04written 2026-10-06
06

What could go wrong, and how we would know

The risks

  • The margin never turns Seven quarters of improvement still leave each sales dollar losing 62 cents; the promised 72-point swing depends on a factory move that has not happened.
  • The cash runs short At the first half's pace the cash lasts to about mid-2027, and only a $15M floor stands between it and the lenders.
  • The lender is the customer Cerberus lends to Eos, owns about 35% of it, and controls the venture that bought most of June's output.
  • More new shares Shares in issue rose sevenfold since 2021 and could near 644M if every warrant and preferred share converts.
  • Lithium-ion keeps getting cheaper Outside forecasters expect lithium-ion to hold 85% of storage; any easing of the US rules on Chinese parts removes Eos's edge.
  • A lawsuit Shareholders sued in March 2026, alleging misleading statements about the factory and the outlook.
The evidence under this - 6 claims
C15management claimlow confidence
Management says scrap, labour, materials and Thorn Hill's ramp give a path to over 72 points of adjusted gross margin improvement over the twelve months from August 2026, assuming the plan and expected volumes, which would take it from negative 62% to about positive 10%. On the Q3 call the chief financial officer said the third and fourth quarters together are well over half of the way to the 72 points, with materials cost down and Thorn Hill cutting conversion cost.
against this

Seven straight quarters of improvement still left adjusted gross margin at negative 62% in Q2; Line 2 was 1% of production, and Line 1 goes down for its move into early 2027 (S05, S14).

  1. 2026-11-05editedQ3 call read: the chief financial officer says Q3 and Q4 together are well over half of the way to the 72 points.
forward · 2026-11-05written 2026-10-06
C49estimatemedium confidence
At the first half's rate of about $131M a quarter, the $271.2M of unrestricted cash at September, which already holds the $87.0M DOE advance, lasts about two quarters, to around the spring of 2027, before the margin turn management promises would have time to show. On the Q3 call management said cash used has tracked the operating loss with no working-capital drain, capital spending moderates from the fourth quarter, and it will give 2027 figures on the Q4 call.
derivedC05's first-half outflow carried forward against the 30 September unrestricted cash; the third quarter's own outflow is distorted by the Frontier investment and equity raised (S35).
against this

Customer deposits ran $51.0M in Q2, 45X credits come in as cash, and Cerberus's $105M revolver is available at its discretion, so the outflow could be smaller (S01).

  1. 2026-11-05editedQ3 2026 results read from the 10-Q and release: cash runway re-struck on the Q3 balance
  2. 2026-11-05editedQ3 call read: management says cash burn tracks the loss, capital spending falls from Q4, and declined a 2027 figure.
forward · 2026-11-05written 2026-10-06
C10facthigh confidence
Eos put $112.6M into Frontier Power USA, the net proceeds of July's $75.0M direct offering and August's rights offering, which sold 6,885,218 of 27,367,171 units against a roughly $150M target. Eos holds 112.6M of 312.6M units, about 36%; Cerberus holds 150.0M, including 50.0M founder units for no cash, and appoints four of seven managers; in liquidation Cerberus's and Hudson Bay's cash comes back before Eos's.
current state · 2026-08-04written 2026-10-06
C08facthigh confidence
The share count rose from 51.8M in February 2021 to 339.4M in February 2026 and 364.2M on 3 August 2026, about seven times; authorised shares were raised from 600M to 800M on 3 June 2026 so the 2031 notes could settle in stock.
historical · February 2021 to August 2026written 2026-10-06
C26facthigh confidence
Wood Mackenzie says lithium-ion has captured the four-to-eight-hour storage market on cost and supply chain and expects it to hold 85% share through 2034; long-duration storage was 6% of global installations in 2025, a 4-hour lithium-ion project costs $107/kWh in China, and Eos is one of three long-duration companies to have raised over $1B.
current state · 2026-03-09written 2026-10-06
C17facthigh confidence
Q4 2025 revenue of $58.0M against a $93.75M consensus sent the shares down 39.4% on 26 February 2026, the capacity milestone was reached five weeks late, and a securities class action filed on 6 March 2026 alleges misleading statements about manufacturing capabilities and the financial outlook.
historical · Q4 2025 to March 2026written 2026-10-06
What we still do not know
  • What price and length Frontier's 2 GWh order carries, and whether Frontier can walk away.
  • How the Z3 performs in a utility's own hands: Dominion's Virginia test has not reported.
  • How much American storage is built to run eight hours or longer, the slice Eos is made for.
  • How much cash was left after September's government loan advance and the third quarter's spending.
The evidence under this - 4 claims
C11facthigh confidence
Frontier Power USA and Eos signed a 2 GWh purchase commitment and capacity reservation agreement dated 11 May 2026, which the company calls firm; its price, term, deposit and cancellation terms are not filed. The related-party project and Frontier together were 49% of backlog volume at 30 June 2026.
current state · 2026-06-30written 2026-10-06
C31facthigh confidence
No independent measurement of the Z3 is published: the only regulator-hosted data is an Eos-written report on the older Gen 2.3 at San Diego Gas & Electric, averaging 75.2% round-trip efficiency over 27 cycles with DC efficiency of 59% to 78%, and Dominion's 16 MWh Eos system at Darbytown, approved by Virginia's commission in May 2024, has reported nothing yet.
current state · 2026-10-06written 2026-10-06
C47inferencemedium confidence
Eos's distinct ground is eight hours and longer, where a lithium-ion system's cost keeps rising with every added hour; from four to eight hours the outside measurers say lithium-ion has already won, and the government calls anything under ten hours short duration.
derivedWood Mackenzie's reading of the 4-to-8-hour market (C26), Ember's falling savings per kWh beyond four hours (C27), the DOE taxonomy (C30) and Eos's own pipeline mix (C12).
against this

Management says over half its pipeline is eight hours or longer and cites Virginia's 10-hour carve-out (C12, C35), so the longer-duration market may be larger than the outside data show.

current state · 2026-10-06written 2026-10-06
C49estimatemedium confidence
At the first half's rate of about $131M a quarter, the $271.2M of unrestricted cash at September, which already holds the $87.0M DOE advance, lasts about two quarters, to around the spring of 2027, before the margin turn management promises would have time to show. On the Q3 call management said cash used has tracked the operating loss with no working-capital drain, capital spending moderates from the fourth quarter, and it will give 2027 figures on the Q4 call.
derivedC05's first-half outflow carried forward against the 30 September unrestricted cash; the third quarter's own outflow is distorted by the Frontier investment and equity raised (S35).
against this

Customer deposits ran $51.0M in Q2, 45X credits come in as cash, and Cerberus's $105M revolver is available at its discretion, so the outflow could be smaller (S01).

  1. 2026-11-05editedQ3 2026 results read from the 10-Q and release: cash runway re-struck on the Q3 balance
  2. 2026-11-05editedQ3 call read: management says cash burn tracks the loss, capital spending falls from Q4, and declined a 2027 figure.
forward · 2026-11-05written 2026-10-06

What would change our view

  1. Unattractive→NeutralThe price falls back within 20% of fair value

    It needs:

    • ○The share price falls to about $0.43, back within 20% of what we think the business is worth.about ≈$0.92 today, 57% below the last close
  2. Unattractive→AttractiveThe price falls 20% under fair value

    It needs:

    • ○The share price falls to about $0.29, 80% of what we think the business is worth.about ≈$0.84 today, 61% below the last close
  3. Unattractive→NeutralOutside buyers at a positive margin, without a new raise

    It needs all of:

    • ○Sales to outside buyers pass $60M in a single quarter.checked at the next results, due in Mar 2027
    • ○The batteries sell for more than they cost to make, for a full quarter, before the one-off charges management excludes.checked at the next results, due in Mar 2027
    • ○No new shares or convertible notes are sold in that quarter or the one before.watched as it happens
We would also be wrong if
  • Frontier's 2 GWh order is filed with its price, length and deposit.
  • Dominion's Virginia test reports the battery returning under 70% of the power stored.

What is coming

  • 31 December 2026 Cerberus lock-up ends; Hudson Bay exchange opens: Whether Cerberus and Hudson Bay sell registered shares into a market already 42% short.
Every dated event ahead
  • 31 December 2026 Cerberus lock-up ends; Hudson Bay exchange opens: Whether Cerberus and Hudson Bay sell registered shares into a market already 42% short.
  • March 2027 Q4 and full-year 2026 results: Whether 2026 revenue lands in the lower half of its range, outside product revenue passes $40M, and how much of the $271M unrestricted cash is left.
  • March 2027 Line 1 moved to Thorn Hill: Whether output holds through the move and the 10% to 15% cost cut starts, or the move costs volume as the guide's top end did.
  • 31 March 2027 Revenue and EBITDA covenants return: Whether the Cerberus and DOE loans need another amendment once the deferred tests apply.
Since we wrote this
Written
6 Oct
Price the thesis read
$2.14
4 Nov
  • 11 NovThesis reviewUnattractive, high conviction: the view holds even in our best caseConviction re-graded on the new anchors: how robust the view is, not whether anything is still open.
  • 9 NovThesis reviewStill Unattractive: priced at 501% of a value near $0.35Re-valued in three stages at an 8% return, the method now drawn on every thesis: the price sits beyond the 20% margin of safety.
  • 5 NovEarningsOutside sales stayed at $14.3M; margin better, outlook trimmed; Unattractive holdsRecord revenue was 83% from the lender-linked venture and the full-year outlook moved to the lower half, while the gross margin gain left it still deep below zero.
  • 19 OctNewsEos's preliminary quarter: sales up a fifth, losses narrower, backlog down, guide trimmed. View heldEos preliminary third-quarter 2026 results.
  • 6 OctThesis writtenUnattractive, medium convictionWritten two months after the June-quarter print, with the Frontier venture closed, the shares near their 52-week low and Q3 results due on 4 November.

Also checked 5 times with nothing new to act on.

Sources 40

Where this came from

Sources - 40

Every claim above cites one of these, or states the reasoning that derived it. The quote is frozen at the date it was read: a link can rot, and an agent fetch returns a reading of a page rather than the page. Backs counts the whole claim ledger, so it can name a claim this page does not surface.

  1. S01primarycompany filing2026-08-05
    the income statement with 45X credits inside cost of goods, revenue by type (related-party product) and by country, customer concentration, the balance sheet, cash flow, the debt table with principal and carrying value, covenants, the Series B preferred and warrant notes, remaining performance obligations, the class action, and subsequent events (the direct offering, the rights offering, the FPUSA contribution)
    backsC06C07C08C10C13C17C18C41C42C43C50C51
  2. S02primarycompany release2026-08-05
    the related-party project at about 80% of revenue and FPUSA at 49% of backlog volume, backlog of $807M and 3.4 GWh, the $24.6B pipeline, the guide narrowed to $300M to $350M, adjusted gross margin and adjusted EBITDA, the Blanquilla order, the CAPAC agreement and the Golden Dome award
    backsC11C44
  3. S03primarycompany filing2026-02-26
    Item 1 on the Z3 product, the Znyth chemistry and the two Pennsylvania plants, the Competition section naming lithium-ion and long-duration rivals, the 45X and DOE loan terms, customer concentration, employees, the going-concern conclusion lifted, the induced conversion and the maturity table
    backsC01C02C08C13C34C41C43C51
  4. S04primarycompany release2026-02-26
    Q4 revenue of $58.0M and the year's $114.2M, management's disappointment at missing its revenue expectation, the late capacity milestone, the 2026 guide of $300M to $400M and year-end backlog and cash
    backsC02C17
  5. S05primarytranscript2026-08-07
    Eos Q2 2026 earnings call transcript (FMP)
    the 72-point margin path, Turtle Creek's 1.5 GWh annualised June rate and Line 2 at 1% of production, fleet round-trip efficiency, the pipeline's duration and data-centre mix, the 36% FPUSA stake, the CEO's case on PJM and Virginia's 10-hour carve-out, and the analysts' questions on margin timing and the guide
    these initiatives provide what we believe is a clear path to over 72 points of adjusted gross margin improvement over the next 12 months
    backsC15C32C35C44
  6. S06primarycompany filing2026-08-06
    the unit table (Eos 112.6M Class B, Cerberus 150.0M including 50.0M founder units, Hudson Bay 50.0M), four of seven managers to Cerberus, the liquidation waterfall, the 2 GWh purchase commitment and capacity reservation agreement dated 11 May 2026, Hudson Bay's exchange right and the DOE consent moving the lines to Thorn Hill
    backsC10C11C45
  7. S07primarycompany filing2026-05-13
    the venture's design, the roughly $150M rights-offering target and the firm 2 GWh capacity reservation as described at announcement
    backsC10C11
  8. S08primarycompany filing2026-09-14
    the $87.0M advance funded on 10 September, about $178M drawn in total, and Thorn Hill at about 4 GWh with two lines
    backsC34
  9. S09primarycompany filing2026-08-31
    Cerberus beneficially owning 195,847,758 shares, 35.0%, all through warrants and preferred, and the lock-up to 21 December 2026
    backsC07C45
  10. S10primarycompany filing2026-04-14
    officers and their histories, the beneficial ownership table, the summary compensation table, the 2025 bonus metrics and payout, the long-term award mix and the say-on-pay vote
    backsC08C37C38C39
  11. S11primarycompany filing2026-08-25
    Buczkowski appointed chief commercial officer and Kroeker leaving on 20 October 2026, with his separation terms
    backsC37
  12. S12primarycompany filing2026-04-30
    Lagi named chief financial officer from 8 June 2026, after Javidi's termination in May 2025 and Kroeker's interim year
    backsC37
  13. S13primarycompany filing2026-09-16
    the chief executive's discretionary purchases of 60,000 and 23,900 shares in March 2026, the September sales under 10b5-1 plans to cover tax on vested performance shares, and a new director's purchase of 15,000 shares
    backsC40
  14. S14primarycompany release2026-08-27
    the line move from the fourth quarter of 2026 to early 2027, about 250 employees affected, a 10% to 15% cut in conversion cost from 2027 and Thorn Hill at about 4 GWh once both lines run
    The move is expected to begin in the fourth quarter of 2026 and be completed in early 2027, subject to customary lender approvals.
    backsC01
  15. S15primarycompany release2026-09-02
    the first project under the MN8 supply agreement: 100 MWh of Eos at ten hours beside 280 MWh of lithium-ion, the Eos part following in 2030
    10 MW/100 MWh of Eos’ Z3™ zinc-based long-duration energy storage, and 70 MW/280 MWh of lithium-ion energy storage
    backsC36
  16. S16primarycompany filing2026-07-23
    6,885,218 units subscribed of 27,367,171 offered at $5.481
    backsC10
  17. S17secondaryvendor data2026-10-06
    Research pack (FMP market tier, FINRA short interest, filed competitor cohort)
    the $3.27 close of 5 October, the 52-week range and 252-day return with its decomposition, short interest at the 15 September settlement, the analyst split and revenue consensus, the reactions to each print, the company value and the EV/Sales cohort, and the next earnings date
    backsC17
  18. S18primarygovernment data2026-08-07
    43.6 GW of US battery capacity at the end of 2025, 8.3 GW added in the first half of 2026 and the planned additions for 2026 to 2028, in power rather than energy
    Operators currently plan to add 26 GW in 2027 and 14 GW in 2028.
    backsC29
  19. S19primaryindustry data2026-03-24
    18.9 GW installed in the US in 2025, half a terawatt-hour forecast for 2026 to 2031, and utility-scale system prices up 23% in a year
    The report projects that the U.S. will install half a TWh of storage between 2026-2031
    backsC29
  20. S20primaryindustry data2026-03-09
    lithium-ion holding 85% share through 2034, long duration at 6% of 2025 installations, a 4-hour lithium-ion project at $107/kWh in China, and Eos among three long-duration firms to raise over $1B
    Lithium-ion batteries have captured the economically critical four to eight-hour storage market through superior cost and supply chain advantages
    backsC26
  21. S21primaryindustry data2026-07-01
    a US 100 MW, 4-hour standalone system at $270 to $365 per kWh of capital cost, and storage costs rising again on tariffs and foreign-entity rules
    the materialization of tariffs on lithium-ion battery imports has curtailed access to the low-cost Chinese cell supply
    backsC27
  22. S22primaryindustry data2025-12-19
    the stationary lithium-ion pack price and China's stationary-cell output at over double global installations
    The average pack price for stationary storage systems dropped to $70/kWh, 45% lower than in 2024.
    backsC27
  23. S23primaryregulator2025-09-29
    the scale of the Moss Landing fire and, in the September 2026 release, removal and demolition running to mid-2027
    The facility contained approximately 100,000 lithium-ion battery modules, of which about 55 percent were damaged in the incident.
    backsC33
  24. S24primaryregulator2025-07-04
    the non-prohibited-foreign-entity cost share an energy storage project needs, 55% in 2026 rising to 75% after 2029, and § 48E's exclusion of storage with prohibited-foreign-entity material assistance and its 10-point domestic content bonus
    in the case of energy storage technology the construction of which begins- ... during calendar year 2026, 55 percent
    backsC28
  25. S25primaryregulator2025-07-04
    $35 per kWh for a battery cell and $10 per kWh for a module, phasing down from 2030 to nothing after 2032
    in the case of a battery cell , an amount equal to the product of- $35, multiplied by
    backsC28
  26. S26primarycompany filing2025-11-25
    6.8 GW and 17.8 GWh deployed, 9.1 GW of contracted backlog, and US-made modules in Utah using Tennessee cells for its domestic content offering
    backsC23
  27. S27primarycompany filing2026-07-23
    22.3 GWh of energy storage deployed in the first half of 2026
    backsC24
  28. S28primarycompany filing2026-08-11
    substantial doubt about continuing as a going concern, $10.8M of cash and the turn to sodium-ion products
    backsC25
  29. S29primaryregulator2024-05-06
    Dominion's Darbytown pilot approved, pairing Form Energy's 494 MWh iron-air system with Eos's 16 MWh zinc system, and the order's statement that a pilot guarantees no later approval
    the Eos Energy Zinc Hybrid BESS is a 4 MW /16 MWh AC system
    backsC31
  30. S30primarygovernment data2023-10-01
    a Gen 2.3 demonstration at San Diego Gas & Electric with 59% to 78% DC efficiency; written by Eos and not approved by the Commission
    Out of the 27 data points, or cycles, depicted, the Average Round Trip Efficiency was found to be 75.2 percent
    backsC31
  31. S31primaryregulator2024-12-03
    the guarantee's closing, its purpose of financing production lines and the 8 GWh four-line plan the lender underwrote
    All four lines of the project are expected to manufacture 8 GWh of storage capacity annually by 2027
    backsC34
  32. S32primarygovernment data2023-07-27
    the department's taxonomy, in which short duration runs up to 10 hours and long duration starts above it
    10-15 GW Annual manufacturing & deployment capacity needed by 2035
    backsC30
  33. S33primaryindustry data2025-12-11
    about $125/kWh of total capital cost for a 4-hour-plus system outside the US and China, and smaller savings per kWh beyond four hours
    total equipment costs are 10-15% cheaper for four-hour projects
    backsC27
  34. S34secondarynews2026-08-12
    Form Energy's $750M equity raise and its statement that backlog has quadrupled this year
    securing $750 million in new equity investments
    read, but no claim rests on it
  35. S35primarycompany filing2026-11-05
    the Q3 income statement with related-party revenue, the debt table at principal, share count at 30 September, customer concentration, remaining performance obligations, cash flow and the Frontier Power USA investment
    backsC03C04C05C06C14C18
  36. S36primarycompany release2026-11-05
    revenue of $82.7M, adjusted gross margin of negative 44%, the outlook moved to the lower half of $300M to $350M, backlog of $765M and 3.2 GWh, the Line 1 move into Thorn Hill, Line 2 yield and the DOE advance
    backsC03C04C12C14C16
  37. S37secondaryvendor data2026-11-05
    Research pack (FMP market tier, FINRA short interest, filed competitor cohort), 5 November
    the $2.14 close of 4 November, the 52-week range, short interest at the 15 October settlement, the analyst split and revenue consensus, and the company value to sales of Eos and its cohort
    backsC19C22
  38. S38primarytranscript2026-11-05
    Eos Energy Q3 2026 earnings call transcript (FMP), held November 5, 2026
    Margin path over half covered by Q3 and Q4, cash tracking EBITDA with capex moderating, no 2027 guide until the Q4 call, NYSERDA first round under a third of the announced amount, Google-MN8 and a 3.5 GW near-term pipeline.
    If you look at the fourth quarter combined with the third quarter, we are over -- well over the 50% of track to get to the 72 points.
    backsC15
  39. S39primarycompany filing2026-11-09
    trailing revenue, operating profit, stock pay, tax, debt, cash and the diluted share count
    backsC53
  40. S40secondaryvendor data2026-11-09
    the analysts' average revenue and operating profit for the next three fiscal years, and the close
    backsC52C53
Rendered from content/EOSE/thesis.json · thesis reviewed 2026-11-05. This page explains a company; it does not say what to pay, when to buy or how much.