A real business finally inflected to GAAP profit on the AI-power boom — but the stock is priced as if the 2.8 GW Oracle deal is the floor, not the ceiling; the asymmetry now runs short.
| Date | Type | What happened | Source |
|---|
| 2026-08-10 | editorial note | Valuation figure revised: $2.3B → 0.7%Valuation moved from $2.3B (deep-dive-2026-06-21.md) to 0.7% (deep-dive-2026-08-10.md). | dossier |
| 2026-08-10 | editorial note | Capex figure revised: $56.8M → $77.8MCapex moved from $56.8M (deep-dive-2026-06-21.md) to $77.8M (deep-dive-2026-08-10.md). | dossier |
| 2026-08-10 | editorial note | Revenue figure revised: $21.9M → $1,065.4MRevenue moved from $21.9M (deep-dive-2026-06-21.md) to $1,065.4M (deep-dive-2026-08-10.md). | dossier |
| 2026-08-10 | editorial note | Verdict changed: A real business finally inflected to GAAP profit on the AI-power boom — but the stock is priced as if the 2.8 GW Oracle deal is the floor, not the ceiling; the asymmetry now runs sho…Before (deep-dive-2026-06-21.md): A real business finally inflected to GAAP profit on the AI-power boom — but the stock is priced as if the 2.8 GW Oracle deal is the floor, not the ceiling; the asymmetry now runs short. After (deep-dive-2026-08-10.md): The valuation short mostly worked and is now two-thirds spent — but what replaced it is worse for the bulls than a high multiple, because the new bear case is about whether the disclosed numbers describe the business. | dossier |
The verdict
The valuation short mostly worked and is now two-thirds spent — but what replaced it is worse for the bulls than a high multiple, because the new bear case is about whether the disclosed numbers describe the business.
Primary sources
SEC filings
Earnings calls
Source documents — open to read in full
Q2 was the biggest quarter in company history and it beat by a mile. Revenue $1,065.4M (+165.5% YoY) — first billion-dollar quarter — vs consensus in the $815.6M–$851.4M range depending on outlet; non-GAAP diluted EPS $0.78 vs ~$0.40–0.42 expected; GAAP diluted EPS $0.62.
Guidance raised a second time in one quarter. FY2026 revenue $3.4–3.8B → $3.9–4.2B (~100% YoY at midpoint); non-GAAP operating income $600–750M → $800–900M; non-GAAP EPS $1.85–2.25 → $2.55–2.85.
Brookfield went from $5B to $25B — a 5× expansion in nine months (announced 2026-06-30). Separately, IDF partnered with Oaktree, MUFG and Morgan Stanley for $2.6B. The capital-partner chokepoint identified in the prior dive materially loosened.
A short seller hit the supply chain, and a securities class action followed. Hunterbrook Media published "Bloom's Big Lie" on 2026-07-08, alleging China-routed scandium sourcing (the dopant stabilising the zirconia electrolyte) and challenging the unaudited ~$20B backlog claim against audited performance obligations of ~$492M. BE closed −5.7% at $254.29 that day. Bloom rejected the report by Form 8-KA filing for something that happened between reports and matters enough to tell shareholders now. on 2026-07-09. A class action now covers purchasers 2025-02-27 → 2026-07-08; lead-plaintiff deadline 2026-09-28.
Related-party revenue collapsed to nothing — and single-customer concentration replaced it. Q2 related-party revenue was $2.818M, i.e. 0.26% of the quarter, versus $373.3M (~50%) in Q1. But the concentration note discloses that for the six months, one customer that is not a related party accounted for ~73% of total revenue. The risk did not go away; it changed counterparty and stopped being labelled "related party."
The stock round-tripped violently and the multiple roughly halved. $217.94 as of 2026-08-10 11:58 EDT, Market capitalisationThe share price multiplied by the number of shares. What the market says the equity is worth. $64.19B, trailing P/E 245.69, forward P/E 62.58, 52-week range $36.80–$351.28, consensus PT $273.51 — cross-checked at $217.56 the same lunchtime. Against the prior dive's ~$329 (2026-06-18), that is −34% in seven weeks while the guide went up 32%.
The structural thesis of the prior dive holds, and the operating half of it strengthened. Bloom is still an SOFC equipment OEM selling time-to-power into an AI-power shortage; the moat is still "best non-combustion platform + installed base + financing rails"; the pricing window is still demand-shock-conferred rather than permanent. Every operating claim in the prior dossier got better: gross margin 30.0% → 33.4% GAAP / 34.3% non-GAAP; operating margin 9.6% → 17.1% GAAP / 22.5% non-GAAP; the company generated $226.4M operating cash flow and ~$174.8M free cash flow in Q2 alone and now sits in a net cash position (−$188.9M net debt).
What broke is the prior dive's central bear argument, and what replaced it is different in kind. The old bear case was "the multiple is the trade" plus "~50% related-party concentration." At ~15.8× EV/FY26-guided sales and ~81× FY26 non-GAAP EPS, the multiple argument is roughly half-spent; the related-party argument is factually gone. In their place: an unresolved allegation that a key input is China-sourced, a live securities class action, a $306.5M "Customer Consideration Asset" that is contractually committed to reducing future revenue, and disclosed performance obligations of $494.1M against a marketed backlog reported at ~$20B. The prior dive worried the business was priced too richly. The open question now is whether the reported business is described accurately.
The platform description is unchanged from the prior dive (solid-oxide fuel cells, non-combustion, ~90-day time-to-power, native DC output). Three structural changes landed in Q2:
One reportable segment confirmed under ASC 280 — the CODM (the CEO) reviews consolidated results only.
The prior dive flagged that "specific upstream suppliers are not named in the filing — chokepoint risk on specialty ceramics... is real but undisclosed." Seven weeks later that undisclosed chokepoint became the story.
The allegation. Hunterbrook Media traced four China-linked routes into Bloom's scandium supply using trade data, Chinese corporate filings and satellite imagery, and argued that Bloom's stated five-gigawatt production ambition would require roughly 220 tonnes of scandium oxide annually against a projected global supply near 240 tonnes. Scandium is used as a dopant to stabilise the zirconia-based ceramic electrolyte.
Bloom's answer, in its own filing. The Form 10-QThe quarterly version of the annual report. Lighter, and not audited. addresses the report directly: "we rejected the report's conclusions regarding our supply chain, and we believe we have sufficient supply of the relevant raw materials to meet our current fuel cell demand and backlog". On the call, Sridhar gave three points and refused the rest as proprietary: enough scandium is economically recoverable, visibility for 25 GW of deployments, and "we are not dependent on China".
What is verifiable and what is not. Verifiable from the filing: non-cancelable purchase commitments beyond 12 months are $16.4M as of 2026-06-30, up from zero at YE2025. For a company guiding to $4B of revenue and claiming multi-gigawatt scale, $16.4M of long-dated committed supply is a strikingly thin book — consistent either with genuinely broad, spot-available materials (management's story) or with the absence of secured long-term access (the short's story). The filing does not resolve it, and neither can this dive. Bloom also flagged elevated freight and selective air-freighting of materials to hit customer timelines — a real, disclosed margin drag and a tell that the chain is running tight.
Capital partners (the other half of the chain) got dramatically stronger: Brookfield $5B → $25B, part of Brookfield's $100B AI Infrastructure Fund; IDF/Oaktree/MUFG/Morgan Stanley $2.6B. Manufacturing capacity is no longer claimed as the constraint: Sridhar now says the current footprint "will allow us to deliver five gigawatts of product annually" and that Bloom is "not order constrained and not capacity constrained" — a large upgrade from the 1 GW → 2 GW plan in the prior dive, and one that carries no filing corroboration.
Two claims strengthened, one weakened.
Revenue by line, Q2 FY2026:
| Line ($M) | Q2-25 | Q1-26 | Q2-26 | YoY | Q2-26 GAAP GM | Share of Q2 |
|---|---|---|---|---|---|---|
| Product | 296.6 | 653.3 | 935.4 | +215.4% | 36.5% | 87.8% |
| Installation | 37.4 | 25.9 | 51.0 | +36.4% | (3.6)% | 4.8% |
| Service | 54.4 | 61.9 | 69.0 | +26.8% | 18.7% | 6.5% |
| Electricity | 12.8 | 9.9 | 10.0 | (22.3)% | 31.1% | 0.9% |
| Total | 401.2 | 751.1 | 1,065.4 | +165.5% | 33.4% | 100% |
Product gross margin of 37.2% non-GAAP (+291bps YoY) is the load-bearing number. Sequential revenue growth was +41.8%.
The concentration cut, which matters more than the mix — and contains a contradiction worth naming. The filing's Concentration of Risk note says:
The Q2 line does not reconcile with Note 11. Note 11 puts total Q2 related-party revenue at $2.818M — 0.26% of the quarter, not 21% ($224M). The 21% figure does reconcile against the six-month related-party total ($376.1M ÷ $1,816.4M = 20.7%). Either the concentration paragraph mis-scopes its periods, or "related party" means something different in that paragraph than in Note 11. I am surfacing this rather than resolving it. Bloom itself adds an unusual footnote defining "customer" as "the contractual counterparty... which in certain transactions may be a project-finance affiliate rather than the ultimate end user" — an admission that the disclosed customer is often not the buyer of the power. Read at its plainest, two counterparties accounted for ~94% of first-half revenue.
The print:
| Q2'25 | Q1'26 | Q2'26 | |
|---|---|---|---|
| Revenue | $401.2M | $751.1M | $1,065.4M |
| GAAP gross margin | 26.7% | 30.0% | 33.4% |
| Non-GAAP gross margin | 28.2% | 31.5% | 34.3% |
| GAAP operating income | $(3.5)M | $72.2M | $182.2M (17.1%) |
| Non-GAAP operating income | $28.6M | $129.7M | $239.6M (22.5%) |
| Adjusted EBITDA | $41.2M | $143.0M | $253.4M |
| Net income to common | $(42.6)M | $70.7M | $196.3M |
| GAAP EPS basic / diluted | $(0.18) | $0.25 / $0.23 | $0.68 / $0.62 |
| Non-GAAP EPS diluted | $0.10 | $0.44 | $0.78 |
Versus consensus. Revenue beat by roughly 25–30% — outlets quote the bar at $815.6M, $827M and $851.4M; the spread itself is worth noting, since a "30% beat" and a "25% beat" are different stories. Non-GAAP EPS of $0.78 against ~$0.40–0.42 is roughly a 90% beat.
Guidance. Raised across the board: revenue $3.9–4.2B, non-GAAP gross margin ~34%, non-GAAP operating income $800–900M, non-GAAP EPS $2.55–2.85. Management says the guide is built bottom-up from backlog conversion plus in-year bookings, against customer site-readiness dates.
Balance-sheet flags — five, ranked by how much they should change your mind:
Also new: effective tax rate 0.7% — the full domestic DTA valuation allowance is still not released after two large profitable quarters. Management still does not underwrite durable US taxable income. Shares outstanding 280.0M → 293.4M in six months (+4.8%); diluted weighted-average 323.3M now that convertibles are dilutive. Disclosure controls concluded effective, no material change in ICFR.
Market reaction — messy, and I could not verify it tick-by-tick. The sequence across sources: ~$329 at the prior dive → 52-week high $351.28 → −5.7% to $254.29 on the Hunterbrook report → −11.34% to $166.84 in the regular session on print day, then +12.08% after hours to ~$187 → ~$185.85 on 07-29 and a cumulative ~+31.9% off the low by 07-30 → −6.35% on 08-07 → $217.94 on 08-10 with an intraday $212.20–$243.88. Only the 2026-08-10 quote is precisely timestamped and cross-checked; the intermediate path is reconstructed from secondary coverage and should be treated as directional. What is not in doubt: a violent round trip, and a stock still ~38% below its 52-week high despite a doubled guide.
The shift versus the prior 3–4 calls is from "we are becoming the standard" to "we are the standard, and scale is now the only question."
Net sentiment: the most confident call in company history, one quarter after the last most-confident call. In the prior dive I flagged peak confidence as a contrarian tell at 27× sales. At ~15.8× sales the tell is weaker, but the pattern — record confidence, refused composition, non-guided cash — is unchanged.
BE's own multiples, refreshed:
| Metric | Prior dive (2026-06) | Now (2026-08-10) |
|---|---|---|
| Price | ~$329 | $217.94 |
| Market cap | ~$93.6B | $64.19B |
| Shares outstanding | ~280M | 294.53M |
| Trailing P/E | n/a (just turned profitable) | 245.69 |
| Forward P/E (per source) | ~135× | 62.58 |
| P/E on FY26 non-GAAP EPS guide midpoint | ~160× (on $2.05) | ~80.7× |
| Enterprise valueWhat it would cost to buy the whole company: its market value plus its debt, minus the cash you would get with it. Often very different from market cap. / FY26 guided sales | ~27× | ~15.8× |
| Consensus price targetThe average of what published analysts think the share price should be. An opinion poll, not a forecast. | $203–283, below spot | $273.51 avg — now above spot (also quoted $286.20 avg / $390 high / $70 low across 26 analysts) |
| 52-week range | — | $36.80 – $351.28 |
Peer multiples were NOT re-sourced this refresh — the session's web-search budget was exhausted. The prior dive's peer table (GE Vernova ~$266.5B cap / ~31× forward / ~6× sales; CAT ~44× forward; CMI, PLUG, FCEL all n/a) is now ~7 weeks stale and must not be quoted as current. Chase it before any position sizing.
What can be said honestly about the comp read: the prior dive's headline was that BE traded at ~4× GE Vernova's EV/Sales and forward P/E for the same tailwind. On BE's side of the ratio, the premium has roughly halved. Whether the gap is now ~2× or something else depends on where GEV has traded since 18 June, which this dive did not establish. The single cleanest change: consensus price targets crossed from below spot to above spot — the sell-side was more bearish than the tape in June and is more bullish than the tape in August, with the same analysts and a doubled guide. That inversion is the most information-dense line in this lens.
New ±5% events since 2026-06-21, in order:
| Date | Event | Move |
|---|---|---|
| 2026-06-30 | Brookfield framework $5B → $25B | Up leg into the ~$351 high |
| 2026-07-08 | Hunterbrook "Bloom's Big Lie" — China scandium + backlog challenge | −5.7% to $254.29 |
| 2026-07-09 | Bloom 8-K rebuttal | Shares climbed |
| 2026-07-28 (regular) | Pre-print de-risking / broad AI-power weakness | −11.34% to $166.84 |
| 2026-07-28 (after hrs) → 07-30 | Q2 beat + double guidance raise | +12.08% AH, ~+31.9% cumulative off the low by 07-30 |
| 2026-07-30 → 08-04 | Securities class actions filed and publicised | Overhang |
| 2026-08-07 | Sector/rate/margin-doubt selloff | −6.35% |
| 2026-08-10 | Continued fade | −4.2% intraday, $212.20–$243.88 range |
What the market now reacts to has changed. In the prior dive the answer was (1) named GW deal scope and (2) beat-and-raise cadence. Both still work — but 2026-07-28 broke the pattern in the most informative way available: the biggest beat and the biggest raise in company history produced a stock that is still 38% below its 52-week high two weeks later. When a doubled guide cannot hold the price, the marginal buyer has stopped paying for growth and started pricing something else. On this tape that something else is the disclosure question, the daily-range volatility ($212–$244 in one session), and rate-sensitivity of a growth-capex model.
Forward catalysts, ranked by information value:
n/a — not verified against a primary FERC document in this refresh.A new CFO arrived and the prior dive did not have him. Simon Edwards is now Chief Financial Officer, signing the Q2 10-Q and 8-K. His offer letter was filed as an exhibit to the Q1 10-Q, dated 2026-03-18, and he received PSUs and RSUs on 2026-05-20 as "newly appointed Chief Financial Officer". His RSUs vest one-third on 2027-04-15 then quarterly over two years. A CFO transition in the middle of a 165%-growth quarter, a short-seller attack and a class action is a genuine execution risk the prior dive could not have priced.
The CEO received a very large new performance grant, five days before the prior dive's cutoff and disclosed in an 8-K it did not cover. On 2026-06-15 the Board granted Sridhar 271,076 PSUs at target, earnable up to 300%. The mechanics deserve reading closely:
The honest read is split. The long holding period to 2031 and the multi-year revenue-plus-margin structure are genuinely well-designed — far better than an annual-EPS award. But the primary metric is a best-four-quarters revenue peak, which pays for a single blowout window rather than sustained performance, and it was granted days before the stock's 52-week high on the explicit rationale that the previous award's targets had already been exceeded. Resetting the bar upward after outperformance is defensible; resetting it at a cyclical peak, on a peak-revenue metric, is the kind of thing that reads differently in 2029. Unamortised expense for the 2026 Executive Awards plus the CEO Award is $106.0M, on top of $401.1M total unrecognised stock-award compensation.
Capital allocation — one new decision, genuinely interesting. On the Oracle warrant, Bloom paid 248,798 shares ($72.3M) to induce Oracle to elect net settlement, forgoing the ~$400M cash exercise proceeds in exchange for issuing ~1.63M fewer shares. That implies a willingness to "repurchase" shares at roughly $246 each, before the $72.3M inducement. Against a stock now at $217.94, that is a decision that has not yet aged well — and it is a live example of management transacting in its own equity at prices it clearly believed conservative.
Dilution continues: 280.0M → 293.4M shares in six months. SBC is improving as a ratio — $100.4M for 1H on $1,816.4M revenue (5.5%), and Q2 specifically was 4.7% versus 6.7% in Q1 and ~8% in the prior dive's framing. Operating leverage is reaching the comp line too.
Insider selling: NOT re-verified this refresh. The prior dive cited ~$83M of net insider selling over the trailing twelve months. The web-search budget was exhausted before Form 4The filing an insider must make within two business days of buying or selling their own company’s shares. activity since 2026-06-21 could be checked. Treat the insider-flow picture as stale.
The 2019 Hindenburg episode stopped being history on 2026-07-08. The prior dive's closing judgment was that the forensic risk was "structural (accounting complexity + related-party concentration + non-GAAP reliance), not an open enforcement action." That is still literally true — there is still no SEC action — but the risk has now taken a concrete legal form.
Live legal exposure (new):
regulatory/regulatory-findings.md refreshed 2026-08-10: still ZERO SEC Litigation Releases and zero AAERs naming Bloom Energy, 2021-08-10 → 2026-08-10. A private class action is not an enforcement action — do not let the headline volume imply one.Three accounting judgments that deserve more scrutiny than they are getting:
Improved since the prior dive (state it plainly):
Unimproved: the $872.6M domestic DTA valuation allowance still stands (0.7% effective tax rate) — management has now produced $274.5M of pre-tax income in six months and still will not assert that future US taxable income is more likely than not. That is either extreme conservatism or an internal view of durability that does not match the external one.
Sourcing note, per the integrity flags above: model.xlsx produced NO computed values (51 uncached formulas, blank share price, only 2 quarters of history, growth seeded at a default 8%). Nothing below is cited from it. Every figure is `` with the arithmetic shown, anchored on the company's own raised guide.
Anchors: FY2026 revenue $3.9–4.2B, non-GAAP gross margin ~34%, non-GAAP operating income $800–900M, non-GAAP EPS $2.55–2.85. 1H26 actuals: revenue $1,816.4M, non-GAAP diluted EPS $1.22 ($0.44 + $0.78). Implied 2H26: ~$2.08–2.38B revenue and $1.33–1.63 non-GAAP EPS — i.e. the guide requires further sequential acceleration, not just a hold. Diluted share base 323.3M in Q2.
| Scenario | FY26 rev | FY27 rev | FY28 rev | FY26 nGAAP EPS | FY27 | FY28 | Logic |
|---|---|---|---|---|---|---|---|
| Bull | $4.2B | $6.7B (+60%) | $9.4B (+40%) | $2.85 | $5.14 | $7.10 | 5 GW capacity claim is real, Brookfield $25B converts, non-GAAP op margin 24%→25%, Oracle 2.8 GW fully deploys, more hyperscalers convert |
| Base | $4.05B | $6.1B (+50%) | $8.0B (+31%) | $2.70 | $4.29 | $5.55 | Guide midpoint, op margin 22%→23%, Customer Consideration amortisation and installation-mix shed roughly offset, ~2%/yr dilution |
| Bear | $3.9B | $4.7B (+20%) | $5.2B (+11%) | $2.55 | $2.78 | $2.62 | AI-capex digestion in 2027, one anchor project slips, op margin compresses 18%→16% on pricing normalisation as turbine supply clears |
Arithmetic, base FY27 : $6.1B × 22% non-GAAP operating margin = $1.342B; + ~$0.060B net interest (interest income ~$85M less interest expense ~$25M) = $1.402B; × 0.98 (2% effective tax, allowance still standing) = $1.374B; ÷ 320M diluted = **$4.29**. **Base FY28** : $8.0B × 23% = $1.840B; + $0.060B = $1.900B; × 0.95 (5% tax as allowance partially releases) = $1.805B; ÷ 325M = $5.55.
Bear FY28 ``: $5.2B × 16% = $0.832B; + $0.050B = $0.882B; × 0.98 = $0.864B; ÷ 330M = $2.62.
GAAP runs materially below non-GAAP — 1H26 GAAP diluted was $0.85 against non-GAAP $1.22, a $0.37 gap on SBC alone; extrapolating, FY26 GAAP diluted lands near ~$1.95 ``, i.e. ~112× GAAP at $217.94.
The valuation conclusion has genuinely changed, and it should be stated without hedging. In the prior dive the punchline was that even the bull FY28 case left the stock at ~63× a non-GAAP number three years out. At $217.94:
A 31–39× multiple on 2028 earnings for a business compounding revenue 30–50% is no longer an obvious short. The prior dive's "the earnings can triple and the stock can still be expensive" was correct in June and is roughly half-wrong in August — a −34% price move plus a +32% guidance raise is a ~50% de-rate. What replaces the valuation objection is the quality objection: every one of these scenarios extrapolates a guide that rests on a backlog Bloom will not itemise, delivered to a customer base where one counterparty was ~73% of the half, using an input whose sourcing is under active legal challenge.
No our model base case was committed — unattended run, per standing rules. For the record only: BE FY2026 non-GAAP EPS >= $2.55 (guide low), p≈0.75, resolves 2027-02-28.
Bull case (stronger than it was in June). Bloom just printed a billion-dollar quarter at 33.4% GAAP gross margin and 17.1% GAAP operating margin, generated $175M of free cash flow, sits in net cash, and doubled its own full-year guide — twice in one quarter. Brookfield, having watched a full quarter of execution, 5×'d its financing framework to $25B, and the CEO's read on that is hard to dismiss: capital of that size follows firm orders, not press releases. Service margin swung 43 points from IPO to +22%, turning the installed base into a real annuity. Installation — the only gross-loss line — is being deliberately handed to third-party installers, structurally raising blended margin. Related-party revenue went to essentially zero. All major US hyperscalers plus a dozen-plus neoclouds have validated the platform, and time-to-power is measurable in the Oracle deployment at 55 days from first engagement. And after all of that, the stock is 38% below its 52-week high at ~15.8× guided sales — the cheapest this business has looked relative to its own trajectory in a year.
Bear case — three ways this permanently impairs (and note the composition has changed):
Pre-mortem (18 months out — December 2027, thesis broke). The class action survives a motion to dismiss and discovery produces supplier documents. Nothing criminal emerges — but the disclosure record shows Bloom knew its scandium came through intermediaries, and the "not dependent on China" line is quietly retired. Meanwhile the anchor customer's second and third phases slip two quarters on site readiness, and because there is no itemised backlog, nobody can tell whether the shortfall is timing or demand. Revenue still grows ~25% in 2027 — a fine year for almost any industrial — but the Customer Consideration Asset amortises into the reported line, gross margin gives back 200bps on pricing normalisation as turbine capacity clears, and the DTA valuation allowance still isn't released. The business is intact. The stock is down 55% from $218, because the premium was never really about the multiple — it was about believing an unverifiable number, and the market stopped.
Are multiples too high? No longer unambiguously. At ~31–39× a 2028 base/bull non-GAAP EPS with 30–50% revenue growth and real free cash flow, the pure valuation short has largely closed. The contrarian view now is the inverse of June's: the market spent seven weeks re-rating this name for price reasons, and the thing that actually deserves re-rating is disclosure. Consensus targets crossing from below spot to above spot, in the same seven weeks a short report and a class action landed, suggests the sell-side is modelling the raised guide and not the question underneath it.
Dismantling the refreshed bull:
Unchanged — see the previous dossier, Lens 14 (15 questions). Questions 1, 3, 4 and 5 (contracted-vs-financed mix, Oracle firm-PO vs intent, capacity gating constraint, ASC 810 consolidation trigger) are now more load-bearing, not less. Four to add:
Every dossier we have written on Bloom Energy, newest first.
The valuation short mostly worked and is now two-thirds spent
A real business finally inflected to GAAP profit on the AI-power boom
Covered in the Knowledge Base
Datacenters & Digital Infrastructure
| Industry | Cloud Computing |
| Size | Public Company |
Where Bloom Energy sits against the other names we cover on this beat. Each line is that company’s verdict, not a summary of it.
The backlog stopped being a press release and became an audited receivable
Cash $1.6B · Runway ~3.6yr at this rate
The largest guidance raise in company history rests on a quarter whose headline was bought
Cash $979M
The inflection the market bought in Q1 was a one-off hyperscaler IRU
Cash $609M · Runway ~2 qtr at this rate
The print that mattered came in split — margin, cash and the raised guide all beat, revenue missed, and the reason it missed was Vertiv's own executio…
Cash $2.8B
No directional call this refresh (Socratic gate, pre-print).
Cash $2.4B