Plug Power

Phase A — Understand the business

Lens 1 · Company Overview

Plug Power is a vertically integrated hydrogen company: it builds the kit that makes hydrogen (electrolyzers, liquefiers), the kit that uses it (PEM fuel cells — GenDrive forklifts, GenSure stationary power), and increasingly it produces and delivers the hydrogen molecule itself (liquid-hydrogen plants in Tennessee, Georgia, Louisiana; Texas under construction). The original franchise is material handling — running electric forklift fleets on hydrogen for "large to mid-sized fleet, multi-shift operations in high volume manufacturing and high throughput distribution centers". The "GenKey" turn-key model bundles the units, the on-site hydrogen infrastructure, and the fuel into one contract.

The business has one reportable segment — "the design, development and sale of hydrogen products and solutions" — managed by go-to-market sales channel, with the CEO as CODM measuring the whole company on consolidated net loss. So there is no clean segment-profit cut; the disaggregation below is by product/service line, not by reportable segment.

FY2025 revenue mix (net revenue, $709.9M total):

• Electrolyzers $187.8M (the growth engine — up from $135.5M FY24, $82.6M FY23)
• Fuel delivered to customers & related equipment $133.4M (up from $97.9M FY24)
• Power purchase agreements (PPAs) $107.6M (up from $77.8M)
• Cryogenic equipment & liquefiers $95.7M (down from $111.5M FY24, $231.7M FY23 — the cryo business has shrunk by more than half off the peak)
• Services on fuel cells & infrastructure $94.5M (up from $52.2M)
• Fuel cell systems $54.0M (down from $52.1M FY24, $181.2M FY23 — the legacy forklift-unit business is a fraction of its former self)
• Hydrogen infrastructure $26.8M (collapsed from $183.6M FY23)
• Engineered equipment $6.9M; Other $3.4M

The story the mix tells: the company's identity has rotated away from selling fuel-cell forklift hardware and hydrogen-site installs (the 2023 model) toward electrolyzers + recurring fuel/PPA/service. That is a deliberate, healthier mix shift — recurring lines are growing while one-time hardware shrinks.

Customers: Walmart = 24.2% of FY2025 consolidated revenue (net of a $29.2M warrant charge that reduces revenue); the second-largest customer = 14.3%. Top-2 ≈ 38.5%. Amazon is the other anchor — a Jan-2025 deal for ~10,950 tons/yr of green hydrogen, with a warrant for up to ~16M Plug shares vesting against $2.1B of spend over seven years. Backlog/orders ≈ $724.1M as of 2025-12-31, with delivery terms spanning 90 days to 10 years.

Lens 2 · Supply Chain

Map: upstream inputs → Plug → end customer.

• Upstream / inputs: PEM membranes and catalysts (platinum-group metals — a cost and single-source risk), power electronics, balance-of-plant components, and — for its own hydrogen production — renewable electricity (the Graham, TX plant is powered by an adjacent wind farm ) and electrolyzer stacks (largely in-house). The 10-K flags tariff/trade-policy exposure on "key components" explicitly — component sourcing is internationalized and tariff-sensitive.
• Plug's own conversion nodes: electrolyzer + liquefaction plants in Tennessee, Georgia, Louisiana (operating) and Graham, Texas (under construction, DOE-loan-backed). These turn power + water into liquid hydrogen.
• Distribution: Plug's own cryogenic trailers and hydrogen-delivery logistics fleet (the "fuel delivered" line).
• End customers / OEM channel: sold "through our direct sales force, and by leveraging relationships with original equipment manufacturers (OEMs) and their dealer networks". End buyers: Walmart, Amazon (material handling); refineries, chemical/steel/fertilizer producers and commercial refueling stations (electrolyzers).

Chokepoints / single-source dependencies: (1) platinum-group-metal catalyst supply; (2) third-party liquid-hydrogen supply arrangements that supplement Plug's own production — when Plug's plants underproduce, it buys merchant hydrogen at a loss, which is the structural reason the "fuel delivered" line ran a (85.9)% gross margin in FY2025 (see Lens 4/5); (3) financing counterparties are effectively part of this company's supply chain — sale/leaseback financiers (with $328.0M of restricted cash pledged as of Q1-26) and the ATM/SEPA equity agents are as load-bearing as any component supplier.

Lens 3 · Competitive Advantages (moats)

What's genuinely defensible:

• Installed base + GenKey lock-in. Plug runs the largest deployed hydrogen material-handling fleet in North America (Walmart 9,500 trucks; Amazon 30,000-forklift-equivalent fuel supply ). Once a distribution center is built around Plug's on-site hydrogen infrastructure and fueling, switching is expensive and operationally disruptive. That's a real switching-cost moat in the one vertical where hydrogen forklifts already beat batteries on multi-shift uptime.
• Vertical integration breadth. Few competitors span electrolyzer → liquefaction → delivery → fuel cell. It is the most complete hydrogen value-chain footprint among the pure-plays.
• DOE loan guarantee ($1.66B, closed Jan 2025). A federal financing backstop competitors don't have.

Where the "moat" is weak (be honest):

• Bargaining power is inverted. Plug needs Walmart/Amazon far more than they need Plug — hence Plug pays them warrants (revenue-reducing) to keep the business. A 24.2%-of-revenue customer you have to bribe with equity is the opposite of pricing power.
• No cost moat. Gross margin was (34.1)% in FY2025 — the core product still costs more to make/deliver than it sells for. A moat that doesn't produce a positive gross margin isn't yet protecting profit, only revenue.
• The end market itself is contested by batteries — the 10-K names "advanced batteries" as the direct competitive threat. For most use cases batteries are winning; hydrogen wins only in the multi-shift / fast-refuel / cold-storage niche.

Net: a genuine switching-cost moat inside a narrow material-handling niche, plus a vertical-integration and federal-financing edge — but no cost moat and inverted customer bargaining power. The moat protects the franchise; it does not yet protect a profit.

Lens 4 · Segments (by product line + geography)

One reportable segment, so this is the product-line and geography cut, all ``.

Gross profit/(loss) by line, FY2025 (and the trend that matters):

The signal: every loss-making line improved its margin sharply YoY, and Services flipped positive (+25.5%). The two structural sinkholes remain PPAs and fuel-delivery — both still bleed because Plug is fulfilling legacy long-term fuel/PPA contracts at uneconomic prices and buying merchant hydrogen to plug supply gaps. The thesis hinges on whether its own production (Texas/Georgia/Louisiana) can backfill those lines and drag the fuel margin toward zero.

Geography, FY2025: US $535.9M (76%, up from $428.2M FY24), Other foreign $174.0M (24%, down from $200.6M). The business is re-concentrating onto the US (DOE-backed domestic production + material-handling base), while Europe/RoW shrinks.

---

Phase B — Measure performance

Lens 5 · Earnings Result (latest print: Q1 2026, reported 2026-05-11)

• Revenue $163.5M, +22.3% YoY (vs $133.7M Q1-25) — beat consensus; management said it "exceeded its expectations on revenue".
• GAAP gross margin (13.2)%, vs (55.3)% in Q1-25 — a ~42pt YoY improvement (management framed it as "71% margin improvement").
• Adjusted EPS −$0.08 vs −$0.10 consensus — beat by 2c.
• GAAP net loss attributable to Plug $(245.3)M, worse than Q1-25's $(196.7)M. The deterioration is non-operating: "other segment items, net" swung to $(137.5)M from $(23.5)M, driven mainly by a $54.6M non-cash mark-to-market loss on the $7.75 warrant liability (it rose from $52.3M to $107.0M as the stock climbed from $1.97 to $2.26). Higher stock price → bigger warrant liability → bigger GAAP loss. This is the single most important thing to understand about Plug's GAAP P&L: a rising share price manufactures GAAP losses.
• Operating lines all improved: R&D $(12.1)M (down from $(17.4)M), SG&A $(70.2)M (down from $(80.8)M) — cost-out ("Project Quantum Leap") is real and showing in the operating expense base.
• Balance sheet / liquidity: unrestricted cash $223.2M (down from $368.5M at 2025-12-31), current restricted cash $183.7M, working capital $734.1M, accumulated deficit $8.5B. Operating cash burn at the FY2025 run-rate was $(535.8)M/yr; financing activities turned negative in Q1-26 ($(31.7)M) for the first time as Plug bought back finance obligations rather than raising.
• Market reaction: the stock was +43% over the 30 days around the print — the market rewarded the margin trajectory and read the deepened GAAP loss as the cosmetic warrant artifact it is.

Unusual vs. its own history: the removal of the going-concern qualifier (see Lens 10) and the first negative financing-cash quarter both mark a genuine inflection from "survive at all costs" to "optimize." Revenue still skews 2H-weighted, so Q1 is seasonally the weakest — the +22% off a weak base is encouraging but not yet proof of the full-year guide.

Lens 6 · Earnings Calls (sentiment trend)

No transcripts on the research shelf (transcripts/ empty), so ``.

Management's narrative has shifted, across the last several calls, from "we have a going-concern problem / we must raise capital" (early-mid 2025) to a specific, dated profitability roadmap (late 2025 → Q1-26 call):

• Exit 2026 EBITDAS-positive; exit 2027 operating-income-positive; exit 2028 overall-profitable.
• Gross-margin breakeven by end of 2026; the structural target is ~$200M revenue per quarter as the level at which the cost base is covered.
• "Project Quantum Leap" is the recurring phrase — operations optimization, workforce cuts, facility consolidation, selective price increases, working-capital reduction.
• What they stopped saying: the constant capital-raise/going-concern language of 2024–early-2025 is gone, replaced by "infrastructure optimization" (asset monetization) and "strengthen the balance sheet."

Tone trajectory: from defensive/existential → cautiously operational. The risk is the classic Plug pattern — credible-sounding dated targets that have historically slipped. (Plug has a multi-year history of missing its own revenue and timeline guidance.) Treat the 4Q26 EBITDAS-positive promise as the falsification line, not a base case.

Lens 7 · Comps

Peer set = the listed fuel-cell / hydrogen pure-plays. Multiples are `` with date, or n/a.

P/E is meaningless for the whole set (all loss-making except, arguably, on Bloom's adjusted basis). The instructed columns EV/EBIT, dividend yield, and 5-yr avg ROE are n/a / not meaningful for these companies (negative EBIT, no dividends, deeply negative ROE).

The single most important comp read: Bloom Energy (+219% YTD, ~32x sales) and even FuelCell (+135%) have been massively re-rated on the AI-datacenter power thesis — fuel cells as on-site baseload for power-starved data centers. Plug has been left out of that trade. Plug's GenSure stationary line explicitly targets "data center markets", yet the market does not credit Plug as an AI-power play. That is either the asymmetric opportunity (Plug is the cheap optionality on the same theme) or the market's correct judgment that Plug's balance sheet and forklift-centric mix disqualify it. This gap is the crux of any bull case.

Lens 8 · Stock-Price Catalysts (>5% moves, last ~5 years) — ``

Pattern of what actually moves PLUG:

• Dilution / financing events (down): every large ATM raise, the 2023–2025 equity offerings, the registered direct offerings — Plug sells off hard on dilution. Share count went from ~945.8M weighted-avg (Q1-25) to 1,395.1M outstanding (May-26) — ~47% in a year. Dilution is the dominant bear catalyst.
• Going-concern flag (Mar 2024) (down hard) → removal of going-concern (FY2025 10-K, Mar 2026) (up) — solvency headlines are the biggest swing factor.
• DOE loan close (Jan 2025) (up).
• Quarterly margin trajectory (Q1-26: +43%/30d) (up) — the market now trades Plug on gross-margin progress toward breakeven, not on revenue beats per se.
• Reverse-split / Nasdaq-delisting headlines (early 2026) (volatile) — the sub-$1 episode and the proxy fight over authorized-share increase vs. reverse split.
• Short-squeeze dynamics — 329M shares short, 23.78% of float, 3.36 days to cover. Sharp rallies (the +43%) are partly squeeze-driven; this is a high-beta, heavily-shorted retail name where moves overshoot in both directions.

What the market reacts to, ranked: solvency/dilution > gross-margin-toward-breakeven > policy (DOE/IRA) > revenue. Earnings beats alone don't hold if the cap table keeps growing.

---

Phase C — Judge people & books

Lens 9 · Management

• CEO transition (the headline): founder-era CEO Andy Marsh stepped down in March 2026 to become Executive Chairman; Jose Luis Crespo (previously head of global sales) is the new CEO. This matters: Marsh built Plug and also presided over $8.5B of accumulated deficit and a multi-year record of missed guidance and serial dilution. Handing the wheel to a sales-driven operator at the moment the strategy pivots to "cost-out + hit revenue scale" is logical, but it's an unproven CEO at the most fragile moment in the company's history.
• Track record: Marsh's legacy is genuinely two-sided — he created the category leader and the largest deployed hydrogen fleet, and he never got it within sight of profitability across ~15 years, funding the gap with relentless equity issuance. Capital-allocation history is the weakest part of the file: cumulative ~$8.5B of shareholder capital consumed, multiple impairments ($785.4M in FY2025, $949.3M FY2024, $269.5M FY2023 ) — i.e., the company has written off >$2.0B of prior investment in three years. That is value destruction at scale.
• Skin in the game / insider ownership: insider-transactions.csv not on the shelf; insider ownership is low (typical of a serially-diluted large-cap former-growth name) — n/a on the precise figure.
• Archetype: transitioning from founder-visionary (Marsh) to professional operator (Crespo) — appropriate for the "execute the cost-down and survive" phase, but it removes the founder who could rally capital, right when capital access is the binding constraint.
• Red flags (management): the warrant-to-customers structure (paying Walmart/Amazon in equity), the 2024 + 2025 restructuring plans (serial — suggests prior over-build), and a long history of optimistic dated targets. No related-party self-dealing or comp scandal surfaced.

Lens 10 · Forensic Red Flags

Forensic lens. Every figure labeled.

• Going concern — the material positive. No "substantial doubt / going concern" language appears in either the FY2025 10-K or the Q1 2026 10-Q (verified by full-text search of both filings on disk). Plug carried a going-concern warning in its FY2023 10-K (Mar 2024); its removal in the FY2025 10-K is a genuine, auditor-blessed improvement and the single most important forensic data point. Liquidity is still tight, but the auditors no longer flag survival doubt.
• GAAP loss inflated by non-cash warrant marks. The $7.75 warrant liability ($107.0M, +$54.6M in Q1-26) re-prices upward as the stock rises. This makes GAAP net loss a poor read of operating progress — but it is also a real future dilution claim (185.4M shares at $7.75, exercisable to Mar 2028). Don't dismiss it as "just non-cash": it's non-cash now, dilutive later.
• Revenue quality — warrant charges reduce revenue. Walmart's 24.2% includes a $29.2M warrant charge netted against revenue. Reported revenue is after paying customers in equity — the "true" gross take is being diluted at the top line.
• Contract-loss accruals & service reserves. The FY2025 P&L includes a +$24.6M benefit from loss-contract releases (vs a $(48.5)M provision in FY24). Reversing prior loss accruals flatters the YoY margin improvement — watch that this is genuine cost reduction, not reserve release. Management itself warns "if elevated service costs persist, the Company will adjust its estimated future service costs and increase its contract loss accrual".
• Impairments. $785.4M FY2025, $949.3M FY2024 — recurring large impairments signal chronic over-investment relative to realizable value; goodwill/long-lived-asset risk remains.
• Cash vs. earnings divergence: operating cash burn $(535.8)M FY2025 vs net loss $(1,631.6)M — the gap is the ~$785M impairment + warrant/fair-value non-cash items. So cash burn is "only" ~$536M, materially better than the headline loss — but still ~$536M against $223M unrestricted cash. Restricted cash ($328.0M pledged to sale/leaseback financiers as of Q1-26) is not freely available — the usable liquidity is thinner than the working-capital number suggests.
• Dilution machinery (the core structural flag): ATM with B. Riley/Yorkville ($944.1M gross capacity remaining) + SEPA with Yorkville (up to $1.0B) as of Q1-26. Plus 431.6M additional dilutive securities outstanding. The company has pre-authorized roughly $1.9B of equity issuance capacity — i.e., the dilution is structural and ongoing, not episodic.

Regulatory findings (required):

• SEC: No Litigation Releases, no AAERs naming Plug Power, verified via SEC EDGAR EFTS (LR + AAER) for 2021-06-23 → 2026-06-23.
• 10-K Item 3 (Legal Proceedings): disclosed at p.46; legal matters described as "handled in the ordinary course of business," with the standard reasonably-possible-loss language and no single material proceeding quantified beyond ordinary course in the surfaced text. (Plug has faced shareholder securities litigation historically; nothing material/new surfaced in this run — flag for a targeted check on a refresh.)
• Non-SEC enforcement: web search for FTC/DOJ/FDA/settlement/penalty surfaced no material regulatory enforcement action. The relevant "regulatory" risk for Plug is policy, not enforcement (see Lens 12).
• Conclusion: No material regulatory or accounting-enforcement findings — verified via SEC EDGAR EFTS (LR, AAER), web search, and 10-K Item 3 as of 2026-06-23.

---

Phase D — Project & stress-test

Lens 11 · Forward Projection (EPS, next 3 fiscal years)

Note: Plug is loss-making at the EPS line and will be through the projection window on GAAP; the operative metric management itself uses is the path to gross-margin breakeven → EBITDAS-positive (4Q26) → operating-income-positive (exit 2027) → net profit (exit 2028). EPS estimates below are `` with arithmetic; consensus EPS lines are mostly n/a at the per-year granularity.

Inputs (labeled):

• FY2025 revenue $709.9M; Q1-26 $163.5M, +22.3%.
• Management FY2026 revenue guide: +13–15% → ~$802–816M.
• FY2025 gross margin (34.1)% → Q1-26 (13.2)%; mgmt targets gross-margin breakeven by end-2026.
• Share count: ~1,395M and rising via $1.9B of pre-authorized issuance capacity; assume continued dilution to ~1.7–2.0B over the window.

Base case:

• FY2026 rev ~$810M, gross margin improves to ~(5)% blended (front-half negative, exiting near breakeven); opex flat-to-down on Quantum Leap; operating cash burn narrows to ~$350–400M. GAAP EPS ~$(0.55)–$(0.70); cash-EPS far smaller. EBITDAS roughly breakeven in 4Q26 if the guide holds.
• FY2027 rev ~$950M–$1.0B, gross margin positive low-single-digits, operating loss narrowing toward breakeven by year-end; GAAP EPS ~$(0.25)–$(0.40).
• FY2028 rev ~$1.15–1.25B, approaching net breakeven per the roadmap; GAAP EPS ~$(0.05)–$(0.15), i.e. flirting with the company's first profitable quarter exiting 2028.

Bull path: AI-datacenter stationary demand inflects the GenSure/electrolyzer mix, revenue compounds >25%, gross margin positive by mid-2027, EBITDAS-positive holds 4Q26 → re-rating toward peer multiples. Bear path: the FY2026 guide slips (Plug's historical pattern), fuel/PPA margins stay deeply negative, IRA/45V credits are cut, and the company keeps issuing equity at depressed prices → EPS losses persist and per-share value keeps eroding via dilution even if the enterprise improves.

Forecast NOT logged (forecast.ts create skipped per --watchlist rule — no committed base case in unattended mode). The scoreable line to log on promotion: "PLUG exits 4Q26 EBITDAS-positive (mgmt definition)" — p ≈ 0.35.

Lens 12 · Bull vs Bear

Bull case. Plug is a genuine cost-down turnaround that the market is pricing as a melting ice cube. Three legs: (1) margin trajectory is real and steep — (99.4)% → (34.1)% → (13.2)% gross margin in five quarters, Services already positive, and a credible path to breakeven by end-2026. (2) The going-concern flag is gone and burn has more than halved ($1.1B → $536M op-cash burn) — the existential risk that capped the multiple has eased. (3) It's the un-re-rated AI-power option — Bloom is +219% and FuelCell +135% YTD on data-center demand, while Plug, which sells the same stationary fuel cells and is the only vertically-integrated hydrogen producer with a $1.66B DOE backstop, trades at ~6–7x sales. If Plug lands even one marquee data-center power contract, the gap closes violently (amplified by 23.8% short interest). Contrarian view the market refuses to see: Plug's GAAP losses are increasingly a warrant-mark artifact of a rising stock, not operating deterioration — the operating business is healing while the headline loss "worsens."

Bear case. The equity is structurally unownable because the cap table is a wealth-transfer machine. Three impairment risks: (1) Dilution is the business model — ~47% share growth in a year, $1.9B of pre-authorized issuance capacity, $7.75 warrants for 185M more shares; even if the enterprise recovers, the per-share claim keeps shrinking, so equity holders may never capture the turnaround. (2) The core economics still don't work — fuel-delivery at (85.9)% and PPAs at (66.2)% gross margin mean Plug loses money on its recurring, contracted revenue; that's a contract-book problem, not a scale problem, and "$200M/quarter to breakeven" is a moving target it has missed before. (3) Policy dependency — the whole hydrogen economics lean on 45V/ITC clean-hydrogen credits and DOE loans; the 10-K explicitly flags subsidy reduction/elimination as a demand risk, and a hostile federal posture on clean energy could gut the model. Pre-mortem (18 months out, thesis broke): the FY2026 revenue guide slipped to single digits, gross-margin breakeven pushed to 2027, the company tapped the ATM at ~$2 to fund burn (another ~300M shares), a 45V rollback hit demand, and EBITDAS-positive never arrived — the stock is back under $1.50 and a reverse split is back on the table. Are multiples too high? At ~6–7x sales for a company with negative gross margin and structural dilution, yes on fundamentals — the multiple is a survival-cum-squeeze multiple, not an earnings multiple.

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

Dismantling the bull case.

• Revenue concentration: Walmart 24.2% + #2 at 14.3% = ~38.5% of revenue in two customers Plug must pay equity to retain. If Walmart in-sources, electrifies with batteries, or simply renegotiates, ~a quarter of revenue is at risk overnight — and the warrant bribes prove Plug has no leverage to stop it.
• The moat is a niche, and batteries are eating the rest. Plug's own 10-K names advanced batteries as the competitive threat. Outside multi-shift cold-chain forklifts, the addressable market keeps narrowing as battery energy density and fast-charging improve.
• Most dangerous competitor bulls underestimate: Bloom Energy. Bloom has won the stationary power / data-center trade decisively (Oracle 2.8GW). If on-site clean power for AI is the prize, Bloom — profitable-ish, $79B, cash-generative — will out-finance and out-execute a cash-burning Plug for every datacenter RFP. Plug's "data-center optionality" may already be lost to Bloom.
• Capital allocation is the worst on the board: ~$8.5B accumulated deficit, >$2.0B of impairments in three years, serial dilution, customer-warrant giveaways. Management has never demonstrated it can turn invested capital into returns.
• Assumptions that must hold for today's price: continued capital-market access at non-catastrophic prices; FY2026 guide actually hits; fuel/PPA margins inflect positive; 45V/DOE support survives the political cycle. Break any one and the model wobbles.
• −20–30% revenue disappointment: at ~$570–650M FY2026 instead of ~$810M, gross-margin breakeven slips to 2027+, burn stays ~$450M+, the ATM gets tapped at distressed prices → EPS losses persist and the equity re-rates down toward the $0.75–1.20 analyst floor.
• Single scenario that permanently impairs: a 45V/IRA clean-hydrogen credit repeal combined with loss of DOE loan disbursement — the production-plant economics collapse, the molecule becomes uneconomic, and Plug is forced into asset sales / restructuring at fire-sale prices. Plausibility: moderate (policy risk is live under the current administration; not base case, but far from tail).

Lens 14 · Management Questions (ordered by information value)

1. The FY2026 guide implies gross-margin breakeven by year-end and EBITDAS-positive in 4Q26 — what specifically in the cost base or mix bridges from Q1's (13.2)% gross margin to breakeven in nine months, and how much of it is already contracted vs. still to be won?
2. Fuel-delivery ran (85.9)% and PPAs (66.2)% gross margin in 2025 — what is the dollar/kg gap between your own produced hydrogen and the merchant hydrogen you buy, and at what plant-utilization level do these two lines reach zero gross margin?
3. With ~$1.9B of pre-authorized ATM+SEPA capacity and the stock near $3, what is the maximum dilution shareholders should expect over the next 24 months, and what milestones would let you stop issuing equity entirely?
4. Bloom and FuelCell have been re-rated on data-center power while Plug hasn't — what is your concrete pipeline of stationary/data-center contracts, and why should the market believe GenSure can win against Bloom's installed base?
5. Walmart is 24.2% of revenue and your #2 is 14.3% — what is the renewal/expansion status of each, and how do you reduce this concentration over the next two years?
6. You pay anchor customers in warrants that reduce revenue — at what point does your bargaining power improve enough to stop, and what's the all-in equity cost of the Amazon and Walmart relationships?
7. The $7.75 warrants (185M shares, to Mar 2028) and 431M other dilutive securities — walk through the fully-diluted share count under your base case at FY2028.
8. The going-concern qualifier was removed in the FY2025 10-K — what changed in the auditors' analysis, and what would have to deteriorate for it to return?
9. The Stream Data Centers asset sale (~$132.5M, expected by 30 Jun 2026) — is it closed, and what's the remaining pipeline of "infrastructure optimization" monetizations and their expected proceeds?
10. How exposed is the production-plant economics to 45V/ITC clean-hydrogen credits and the DOE loan, and what's the plan if either is curtailed by federal policy?
11. You took >$2.0B of impairments in three years — what governance/discipline change ensures the Texas/Georgia plants and electrolyzer capacity won't be the next write-downs?
12. New CEO Jose Luis Crespo comes from sales — what are the top three operational changes he's making that Andy Marsh wasn't, and how is the Exec-Chairman/CEO division of responsibility actually working?
13. Restricted cash of $328M is pledged to sale/leaseback financiers — how much of your reported liquidity is genuinely usable, and what's the runway in months on usable cash at the current burn?
14. Backlog/orders were ~$724M at year-end with 90-day-to-10-year delivery — what's the near-term (≤12-month) deliverable portion, and what's the cancellation/renegotiation risk in it?
15. What is the single internal metric you watch most closely as the leading indicator that the turnaround is on or off track?

---