ABAT Stock: Can the Customers Still Buy? — FY2026

The 60-second read

Five things to know

1. Factory margins improved. Derived fourth-quarter gross profit reached $1.29 million, while the full year still recorded a gross loss.
2. Customer access is the near-term test. Black mass is a majority of revenue; the company seeks an exception to the domestic-allocation requirement.
3. Stock pay exceeded sales. FY2026 compensation expense was $46.48 million, and the common-share count rose sharply.
4. Cash came from financing. Outside capital more than offset operating and investing cash use; zero financial debt does not mean self-funding.
5. The mine is an option, not cash. Our valuation separates project success, financing ownership and per-share dilution.

Three scenarios, one probability-weighted number

Independent scenario judgments; probabilities sum to 100%. Values are assumption-driven estimates.

Signposts: what would change our mind

These are our research thresholds and review dates, not management guidance or confirmed earnings dates.

Price and ownership

The reference price is $2.47, the September 11 closing quote returned by Financial Modeling Prep. The annual filing became public on September 14 before a completed regular trading session. This packet does not attribute Friday’s decline to Monday’s annual report. The line chart provides context for the decision; it is not evidence that investors have already priced every new disclosure. An apparent earnings-calendar actual is also not a substitute for reading the financial statements.

There are two market-capitalization numbers worth keeping separate. FMP reports $336.94 million. Multiplying the same share price by the annual cover’s September 8 count of 142,577,758 shares produces $352.17 million. We use that later share count for the market-implied enterprise value, approximately $302.65 million after subtracting unrestricted June cash. Combining an old share count with new cash would make the stock appear cheaper without changing the actual ownership claim. Neither number is a fresh valuation of every outstanding option or warrant.

The price has been volatile, with the provider’s trailing range spanning $2.00 to $11.49. Those extremes measure trading history, not intrinsic value. A return to the high requires an economic explanation: profitable recycling growth, a dependable route to customers, or a mine whose financing preserves meaningful value for common shareholders. Saying that the shares were once more expensive supplies none of those things. Likewise, the decline alone does not demonstrate that a financing-dependent project is undervalued.

Our model starts with common shares and then adds the June unvested awards as a conservative ownership sensitivity. That produces 155.559 million shares, not a claim about GAAP diluted earnings per share. Awards may forfeit, vest later, or overlap with shares issued after quarter-end; the overlay intentionally favors caution. We separately discuss warrants because assuming exercise requires a price, vesting conditions and cash proceeds. We do not silently treat every warrant as current common stock.

The annual results

Annual revenue rose to $21.74 million, up 406.8% from a very small prior-year base. The rise is real: the company has moved beyond its first commercial sales and increased throughput at its Nevada recycling operation. But revenue is the first line of the ownership equation. Cost of goods sold still exceeded annual sales, leaving a $3.09 million gross loss before research, administration and exploration. Describing the annual business as profitable because the last quarter showed positive gross profit would cross both a period boundary and a cost boundary.

The fourth quarter deserves credit on its own terms. Subtracting the filed nine-month statement from the audited annual statement gives $8.233 million of revenue and $1.289 million of gross profit for the quarter ended June 30. Those are computed amounts from dollar-level statements. The annual report’s management discussion rounds the same picture to $8.2 million and $1.3 million. Revenue increased modestly from March while the cost line declined; the gross-margin improvement is therefore more informative than the large year-over-year growth percentage off an early-stage base.

There is an equally important limit. The derived fourth-quarter net loss was $19.96 million, including substantial stock compensation and development costs outside gross profit. We do not subtract rounded quarterly earnings per share to manufacture a fourth-quarter EPS figure. The annual audited loss per share is $0.58. FMP’s calendar row showing a $0.15 loss is not the annual figure, and its annual revenue estimates are thin, mixed-date observations rather than a clean consensus collected just before the release.

The August 20 release explicitly called its fourth-quarter results preliminary and said the financial review was incomplete. This September 14 episode uses the complete audited FY2026 report. It can compare the earlier operational update with the completed accounts, but it cannot describe today as the first time any fourth-quarter revenue number was disclosed. That distinction matters for both the freshness of the episode and any claimed stock-price reaction.

Recycling and the mine

There are two different economic tasks inside this company. Recycling has a physical facility, current customers, reported revenue and a cost of goods sold. Tonopah Flats is a development project with resources, reserves, a process design and a permitting path. Those are not interchangeable measures of progress. This packet uses the consolidated financial statements because the filing does not provide a separately reported segment profit table that can support a clean two-segment earnings split. We will not assign fictional revenue to the mine to make a chart look complete.

The recycling operation’s attraction is the possibility of extracting value from a domestic waste stream while supplying critical minerals. The company’s internally developed process and continuing improvements could create a useful commercial position. Yet the present customer mix is the weak link. Black mass supplies a majority of revenue and substantially all current black-mass customers are outside the United States. A policy intended to keep strategic materials domestic can therefore restrict the customer channel of a domestic producer. National importance and near-term shareholder profitability can move in different directions.

The annual report summarizes Tonopah’s September 2025 pre-feasibility study. It cites approximately $2.0 billion of initial capital, average annual lithium hydroxide monohydrate production of 30,000 tonnes, and an after-tax project NPV of $2.57 billion at an eight-percent discount rate. At ten percent, the reported NPV is $1.75 billion. These are project-model outputs under the study’s operating and commodity assumptions, not cash sitting in the company’s bank account and not a quoted acquisition offer. The study distinguishes 2.7 million tonnes of reserves from the larger 21.3 million-tonne resource estimate.

The financing question is therefore central. The initial project capital is many times the company’s present equity capitalization. Financing might involve grants, debt, partners, sales of economic interests or new common equity. A project can proceed successfully while existing shareholders retain less of its eventual cash flow than a headline NPV comparison suggests. Our valuation explicitly separates the probability of commercial success from the share of project economics retained. Both are judgments. Neither is a signed financing plan or a management forecast.

Where the revenue went

The annual gross loss improved by $7.49 million, which is a meaningful manufacturing achievement. It was overwhelmed by the increase in operating expenses. General and administrative expense reached $51.63 million, research and development reached $17.87 million, and exploration expense reached $2.11 million. In total, operating expenses rose to $71.62 million. The company lost $74.71 million before other income and expense. This is why the factory-margin story cannot carry the whole investment case.

Stock compensation explains much of the difference between economic losses and current cash outflow. FY2026 expense was $46.48 million, versus $14.65 million a year earlier. Most was recorded in administration, with additional amounts in research, exploration and cost of goods. It is legitimate to distinguish noncash compensation from a cheque that must clear tomorrow. It is not legitimate to call it free. Awards transfer a future claim on earnings and assets to employees, and large new awards can keep ownership dilution alive even after the initial cash-financing need has passed.

The company provides a narrower non-GAAP adjustment at the factory level. Its rounded annual cash cost of goods sold excludes approximately $3.7 million of depreciation and $1.1 million of stock compensation from GAAP cost of goods sold. That produces about $1.7 million of adjusted gross profit. It does not produce adjusted corporate operating profit. Research, administration and exploration remain outside that calculation. Our analysis preserves the scope of the adjustment rather than using a manufacturing subtotal to label the entire enterprise cash profitable.

Below operating loss, the annual statement includes interest income, financing-cost amortization and other income. Their combined benefit was about $1.33 million. There was no income-tax provision that rescued the result, and there is no minority-interest bridge needed to explain away the loss. The bigger story is within operating costs and the ownership denominator. The bridge is deliberately simple: it should show where the loss arose without suggesting that every noncash expense will disappear from future years.

Earnings quality

The financial statements received an unmodified audit opinion from KPMG. That is evidence about the statements as a whole, not a certification that every internal process is strong. Management concluded that internal control over financial reporting was ineffective at June 30. The described weaknesses include accounting expertise, segregation of duties and pervasive process-level control deficiencies. Remediation is under way. We treat that as an uncertainty that requires closer reconciliation and a wider valuation margin, not proof that the current audited numbers are fabricated.

The current annual error-correction and recovery-analysis checkboxes are not checked. Older filings nevertheless contain corrected comparative information. Historical series in this packet retain the actual primary statement source and column identity so that a reader can see which version was used. An old correction is not automatically a new restatement event. It does reinforce why neither a provider’s standardized field nor an isolated headline should be trusted without matching it to the report.

Receivables were $6.95 million at year-end, compared with $2.80 million a year earlier. Four customers represented 81 percent of the receivable balance. This is concentration in amounts owed, not a disclosed 81-percent revenue share. Revenue increased faster than year-end receivables over the full year, so the balance alone does not prove worsening collections. The days calculation uses quarter-end receivables divided by quarterly revenue and annualized days; it is a rough operating signal rather than an audited average-days measure. A few large invoices can move it sharply.

Inventory is small relative to reported sales, but it carries exposure to commodity prices and recovery assumptions. The cash-flow statement includes a noncash inventory write-down. A cash gross-margin presentation cannot eliminate the commercial risk of lower realized metal prices or a blocked customer route. We prefer a scorecard that keeps these distinctions visible: the operating improvement is real, cash burn is also real, and dilution can absorb part of any eventual improvement before it reaches each share.

Cash and financing

Unrestricted cash was $49.52 million at June 30, with another $0.80 million restricted under letters of credit. Financial notes payable had been eliminated; the small remaining lease liability should not be described as a conventional funded-debt balance. That is a much stronger immediate cash position than a year earlier. The source of the improvement matters, however. Financing supplied $75.61 million during FY2026 while operations consumed $24.19 million and investing consumed $13.58 million. The cash balance grew because outside capital exceeded the cash used by the business.

The largest financing source was the at-the-market common-stock program, supplemented by warrant exercises and employee share purchases. The cash-flow statement and equity statement describe related but different timing and classification: proceeds, issue costs, shares issued and receivables do not always fall into identical rows. We use the cash-flow statement for cash received and the common-share statement for ownership issued. The company’s quarter-end common shares increased from 97.40 million to 141.54 million, a 45.3% rise. The September cover count was higher again.

A standardized data trap would materially flatter free cash flow here. FMP’s annual cash-flow response reports zero capital expenditure. The primary filing reports $12.15 million of property and equipment acquisitions plus $1.43 million of mining-property cash purchases. Our ordinary free-cash-flow measure is operating cash flow less property-and-equipment cash purchases, a $36.34 million outflow. Including all investing cash flow produces a $37.77 million outflow. Both definitions are shown so a reader can reconcile the amount rather than inherit a provider’s missing capital-spending field.

Dividing year-end unrestricted cash by the past year’s operating-plus-investing burn yields about 15.7 months. This is a static historical comparison, not a solvency forecast or a promise of funding through a date. Construction spending could accelerate, an export restriction could reduce collections, and equity markets could supply additional money. The right forward question is how much cash must be raised under a viable operating plan and how much of the company current owners retain after it is raised.

What the price requires

Our fair-value anchor is $1.25 per share, with very high uncertainty and a SELL rating at the $2.47 reference price. The main operating DCF plus risked project option produces $1.235, rounded to this research anchor. Alternative terminal-sales and funding-reserved-book methods produce $1.43 and $1.57. These are cross-checks rather than an arithmetic blend; they share the same uncertain mine option. A 35-percent discount suggests a conditional entry around $0.81, only with a viable customer route and credible funding.

The corporate cost base must survive the factory-margin discussion. FY2026 operating expenses of $71.619 million less $45.363 million of nonfactory stock compensation give a $26.256 million cash-cost proxy. Other noncash costs remain in that conservative proxy. We retain it, grow it three percent annually, and add $5 million per year of cash-equivalent compensation for future services or new grants. That last figure is our assumption, about half the remaining $27.8 million award expense divided by 2.94 years. It is not a second charge for historical awards already in the share overlay. Zero and $10 million future-service budgets produce $1.44 and $1.02 of DCF equity value, respectively.

Our FY2027–FY2031 revenue assumptions are $30 million, $45 million, $65 million, $85 million and $110 million. Factory cash-contribution margins are 25, 30, 33, 35 and 37 percent. The first year is below the rounded derived Q4 cash gross margin of about 26.8 percent; later improvement requires scale and viable customer access. Corporate budgets of $32.04 million rising to $35.44 million prevent that factory contribution from masquerading as corporate profit. EBITDA remains negative through FY2030 and reaches $5.26 million in FY2031. We subtract taxes at 25 percent of positive EBIT, capital spending at five percent of revenue, and working capital at fifteen percent of incremental revenue; depreciation is four percent. None of this is management guidance.

At eighteen percent discount and eight times terminal EBITDA, operating enterprise value is negative $41.16 million. That is the modeled burden of continuing this plan before the project option, not a negative quoted sale price. Undiscounted economic cash uses total $85.88 million, compared with $49.52 million starting cash. The model crosses below zero during FY2028 and reaches a $36.36 million funding shortfall by FY2031, even without mine construction or a minimum-cash reserve. Treating future compensation as equity could reduce cash need while increasing dilution. Negative operating value recognizes the burden but does not establish that financing can be obtained; a failed funding plan can invalidate the continuation case.

The mine option uses the issuer’s $1.75 billion project NPV at ten percent, fifteen-percent commercial success and seventy-percent retained economics, giving $183.75 million. Initial project capital is already within study cash flows, so we do not subtract the $2 billion twice. The full corporate cost proxy is retained because the filing does not support a precise allocation of project and corporate overhead. Financing ownership is reflected in the retained-economics judgment; no unverified fundraising proceeds are added. Existing unvested awards are a separate conservative denominator overlay. The base DCF adds June cash to operating value and this option, then divides by 155.559 million shares.

The terminal-sales alternative capitalizes FY2031 sales at one times, discounts that terminal value, and includes the same five years of fully costed negative cash flows. It gives $1.43 per share; the multiple is our assumption, not an observed peer average. The static book check removes mining-property and intangible carrying values, reserves $51.14 million for the first two years of economic cash use, then adds the risked project option. It gives $1.57. Cash already in equity is not added twice. This book check does not fund later losses or verify asset recoveries, which is why it does not override the full DCF.

With other base inputs fixed, $2.47 requires about 30.7% mine success at seventy-percent retention, versus our fifteen percent. The reverse probability is model-dependent, not a measured market consensus. A punitive sensitivity adds all 19.182 million warrants with no proceeds and produces about $1.10; actual exercise terms and cash proceeds would matter. The large scenario range is more informative than false precision in the central estimate.

Wall Street context

The available analyst picture is thin. FMP returned one price-target news record: Maxim Group initiated with a Buy and a $6 target on April 28, 2026. That predates this annual filing and the effective date of the domestic-allocation order. The provider’s consensus-target endpoint also says $6, but one underlying observation should not be presented as a broad or freshly reaffirmed Street consensus. The grades and rating-distribution responses were empty. We will not invent a ratings stack, a new analyst update or a named firm’s view of today’s report.

The annual-estimates response shows two analysts contributing FY2027 revenue and EPS estimates. Average revenue is $31.712 million, with a wide low-to-high range. FY2028 revenue is based on one analyst and is lower than the FY2027 average. That uneven path illustrates how sparse the coverage is. It does not establish that the business must contract in FY2028, and it does not justify smoothing the numbers into a stronger growth curve while still calling it consensus. Our DCF is visibly separate from these provider estimates.

Some standardized expense estimates are internally implausible relative to the reported business scale. We retain the raw response for audit and exclude those fields from our valuation rather than dress them up as useful precision. The same discipline applies to the provider’s float count, which conflicts with the later common-share count on the annual report cover. A named data source is necessary for provenance; it does not remove the need for judgment about comparability and timing.

The disagreement with the old target is therefore substantive. A bullish analyst can reasonably emphasize strategic assets and the long-term domestic supply chain. Our lower anchor places more weight on financing ownership, stock compensation, customer access and the chance that the mine takes longer or costs more to commercialize. We would reassess on fresh, sourced evidence. We would not move the model merely because an old target offers a larger percentage upside.

Management and governance

Management earned credit for the manufacturing improvement. Revenue increased and the latest quarter delivered positive gross profit. The company also eliminated financial notes payable and raised the cash available to continue development. These achievements matter because an early-stage industrial business must first prove it can operate, collect revenue and survive the build-out. Dismissing all progress because the annual result is a loss would be as incomplete as celebrating growth without studying its funding.

The offset is capital discipline. Stock compensation more than tripled and the common-share base expanded sharply. Compensation can attract scarce technical talent, but shareholders need evidence that it buys durable cash economics rather than a recurring stream of awards. The annual report also discloses approximately $27.8 million of unrecognized award expense over a weighted remaining period of 2.94 years. That is not a forecast of all future stock compensation because new awards can be granted. It does make a simple one-time add-back narrative difficult to defend.

The latest saved proxy is a historical ownership snapshot. It reported Ryan Melsert with 2.805 million beneficial shares, or 2.36 percent, and directors and officers as a group at 2.80 percent. The annual report incorporates the forthcoming proxy for updated ownership detail. We do not recycle those older percentages as today’s ownership. The saved proxy establishes one vote per common share. Preferred series were authorized but had no shares outstanding at June 30; the historical special voting arrangement should not be described as currently outstanding control stock.

The questions we would ask are concrete. What permission or domestic offtake arrangement supports the next black-mass shipments? What realized price and cash collection terms accompany it? Which capital-spending obligations are committed rather than optional? What portion of Tonopah economics remains with present shareholders under the preferred financing plan? And what dated, independently tested evidence will demonstrate that the control weaknesses have been remediated? Answers to those questions would move the valuation more than another broad statement about the importance of American critical minerals.

What could change the outcome

The highest near-term risk is not that the recycling plant stops being technically interesting. It is that the business cannot sell enough output on workable terms. The government’s domestic-allocation requirement became effective after the fiscal year-end, so the June accounts cannot measure a full quarter of its effect. The company reports that it has requested an exception. We found no source establishing that it has received one. A domestic alternative could solve part of the problem, but the annual report warns that domestic demand and commercial terms may be less favorable.

Financing is the second major risk. A larger cash balance can buy time, while the costs of construction, engineering and corporate operations remain large. New capital may arrive at a price that transfers more ownership than our model assumes. Grants are conditional awards, not unrestricted equity value. A project-level partner can reduce the capital required from the company while also taking part of the upside. These are different mechanisms and should be modeled separately rather than hidden inside a single optimistic discount rate.

The strongest bull case is that the customer constraint proves temporary, the factory continues improving, and the mine attracts financing on terms that preserve substantial economics. The strategic value of a domestic supply chain could then support a much larger business than today’s revenue suggests. Our bull scenario allows that outcome. It is not dismissed simply because current profits are negative. The reason it does not set our anchor is the combination of execution, policy, commodity and ownership conditions that must hold together.

The bear case is that collection and customer access weaken while spending and share issuance continue, leaving the mine as a distant option and recycling unable to cover the corporate cost structure. Specialized assets can be hard to monetize under pressure. An unmodified audit does not prevent a commercial loss, and no default event is required for common shares to lose substantial value. Our downside scenario is severe because the operating business and the project option can deteriorate together. Scenario probabilities are research judgments, not statistical frequencies estimated from a historical sample.

Dated signposts

The useful catalyst is evidence that changes the economics. An exception decision, a credible domestic sales agreement or a financing plan can do that. A conference appearance or a calendar-estimated earnings date cannot do it by itself. We will distinguish the date on which information is expected from the date on which it is actually published and verified. No future earnings time is presented here as an issuer commitment unless the issuer has made that commitment.

The first scheduled checkpoint in our process is the next quarterly filing after the September quarter. November 30 is our review deadline for checking the latest available filing and customer-access disclosure; it is not a promised company release date. At that checkpoint, we want to see whether quarterly revenue stays above the June level and whether gross margin remains positive without a loss of cash discipline. The same review should reconcile receivables, cash spending, financing receipts and newly issued shares.

For the mine, a milestone is only useful if it advances feasibility, legal permission or financing. The company’s project NPV can remain large even while common shareholders’ expected share of it falls. We therefore track retained project economics and funding obligations alongside technical progress. A grant announcement should identify its conditions and timing; a partner announcement should identify the economic interest transferred. We do not treat the face value of an award as current cash available for any corporate purpose.

Sources and methods

The research dataset contains nine quarters from Q4 FY2024 through Q4 FY2026 and three annual cash-flow periods. June-quarter dollar results are computed from the annual statements less the preceding nine months. This does not create an audited standalone quarterly report, and per-share earnings are not computed by subtracting rounded EPS. The stated FY2026 first-three-quarter net losses sum to $53,416,734, while the March nine-month subtotal is $53,416,735. We retain that $1 source discrepancy; Q4 uses the filed cumulative subtotal. For the historical cash-flow series, we retain the actual acquisition-row wording because equipment, water rights, mining spending and grant reimbursements have changed classification. Operating cash flow plus total investing cash flow is the cleanest consistently broad cash-use comparison.

Every primary filing is saved with its URL, download time and SHA-256 digest. The normalized dataset keeps the selected statement tables, raw rows, numeric observations and column headers. FMP raw responses are retained even when a field is excluded from analysis. Computed ratios and valuation models are generated by the episode’s calculation script. No values are supplied from memory. Ratios involving negative earnings are withheld when they would imply a false economic interpretation, and unavailable segment or ratings information is labeled rather than filled with a synthetic series.

Definitions are deliberately narrow. Operating cash flow is the filed cash-flow subtotal. Free cash flow here means that subtotal less cash acquisitions of property and equipment; the broader investing-inclusive measure is shown separately. Enterprise value uses the reference share price and the later annual cover share count, then subtracts unrestricted cash; lease liabilities are disclosed separately. Gross margin uses GAAP gross profit divided by revenue. The award-adjusted denominator is a conservative valuation overlay and is not the denominator used for reported diluted EPS.

Project NPV belongs to a project model. Our probability and retained-economics factors transform that project value into an uncertain claim relevant to shareholders; they do not verify the technical study or guarantee financing. The operating DCF, sales multiple and book cross-check share that same project assumption, so their differences in terminal value and funding reserves do not remove overall uncertainty. The wider bear and bull scenarios are the more honest representation of that uncertainty. A 35-percent margin-of-safety requirement is our decision rule, not a forecast of how far the stock must fall.