Powell Industries — Blitz Index 47 (C)

Backlog 2× revenue but production is the bottleneck — $400M of it anonymous.

Last re-scored: October 18, 2026.

Score breakdown

Think we scored it wrong? Score it yourself →

Radar axes (0–100): Moat 46.7 · Speed 64 · Creative 53.3.

Watch item

Gates: G1 ✓ · G2 ✓ (borderline — 10× in 10 years needs secular DC+grid demand). Capacity (Jacintoport, FY2026-end) is the unlock.

Strengths

• Quarterly orders $934M (+158%, record) · book-to-bill 3.0 — demand visibility at its best
• Backlog $2.4B (+69%) ≈ 2× FY2026 revenue; DC $400M+ largest single order ever
• Gross margin 21.1% → 29.4% in 2 years — margin proof complete; debt-free + $634M cash
• Petrochemical −49% absorbed by DC/Utility mix shift — 'de-cycling' evidenced in the mix

Concerns

• Revenue can't catch orders — capacity bottleneck; anonymous $400M ≈ 17% of backlog
• 3 straight quarters of EPS consensus misses — narrative vs. numbers gap
• ETO project business remains cyclical — 'mitigated', not solved
• Raw materials 45% of revenue — steel/copper/aluminum exposure

Research notes

Thesis: a demand blitz with an execution bottleneck

Powell Industries, founded in 1947, makes custom-engineered electrical distribution and control equipment, centred on medium-voltage switchgear. It is not a platform company, and its score does not pretend otherwise. The case for Powell in this scoring period is narrower and, in its own way, more concrete: a long-established maker of engineered-to-order equipment is in the middle of a re-blitz, shifting the market it serves from petrochemical work toward electric utilities and data centres, and the order book is moving far faster than recognised revenue.

That tension is the whole story. Revenue rose from $699.3 million in fiscal 2023 to $1,012.4 million in fiscal 2024, an increase of 44.8%, and then to $1,104.3 million in fiscal 2025, an increase of 9.1%. Realised growth, in other words, slowed to single digits even as demand, measured by orders, accelerated sharply. Quarterly orders moved from $271 million to $439 million to $490 million and then to $934 million, the last figure described in the score file as a record and as up 158% year on year. The book-to-bill ratio reached 3.0, and backlog rose from $1.4 billion to $2.4 billion. The score file characterises this gap, on a press-reported basis, as an execution problem rather than a demand problem. That characterisation matters because it frames everything that follows: the question is not whether customers are placing orders, but whether Powell's manufacturing capacity, project execution, and delivery timetable can convert a much larger backlog into recognised revenue at the margins it has recently achieved.

Powell is classified in the score file as a mature company undergoing a re-blitz, with revenue of about $1.1 billion. The old axis is shrinking. Petrochemical revenue fell 19% in fiscal 2025, and the quarterly declines in that segment steepened through the scoring period, from 31% to 37% and then to 49% in the third quarter of fiscal 2026. The new axes are growing. Electric utility revenue rose 50% in fiscal 2025, to $279.0 million, and commercial and other industrial work, which includes data-centre orders, grew alongside it. In fiscal 2025, the score file records, the petrochemical decline of 19% was offset by utility growth of 50%, commercial and other industrial growth of 19%, and growth of 87% in the segment abbreviated in the file as LRT. In the third quarter of fiscal 2026, the pattern repeated in a sharper form: petrochemical revenue down 49%, set against commercial and other industrial revenue up 54% and utility revenue up 18%. This is a mix transition demonstrated in reported segment numbers, not merely asserted in management commentary. Whether that transition is durable is a separate question, and it is one the score treats with caution, because engineered-to-order project work is cyclical by nature. The score file records that cyclicality as mitigated by diversification rather than resolved.

The as-of date for this assessment, denoted T, is 18 October 2026. The latest reported results at T were the third quarter of fiscal 2026, published on 3 August 2026. Powell has a September year end, so fourth-quarter and full-year fiscal 2026 results had not yet been reported at T and are excluded from this assessment. Per-share figures in the score file are shown on a retroactively adjusted basis following a 3-for-1 stock split on 2 April 2026. Readers should keep that dating in mind throughout: everything below describes the company as it stood at T, and later results appear only as future markers, without any anticipated outcome attached to them.

The score at a glance

Powell Industries scores 47 out of 100, grade C, under the BlitzLab v1.0 model. The section totals are Speed 16 out of 25, Moat 28 out of 60, Creative Destruction 8 out of 15, and Demerits minus 5. The arithmetic is 16 plus 28 plus 8 minus 5, which equals 47. The score falls 18 points below the 65-point threshold used in the model, and no FLAG is triggered, because total demerits of 5 are below the 15-point FLAG threshold.

The item scores are as follows. In Speed: A1 Growth 7 out of 10, A2 Capital velocity 3 out of 5, A3 Hiring and productivity 3 out of 5, and A4 Share-gain velocity 3 out of 5. In Moat: B1 Network effects 3 out of 18, B2 Distribution scalability 9 out of 18, B3 Market size 8 out of 12, and B4 Gross margin 8 out of 12. In Creative Destruction: C1 Order being replaced 3 out of 5, C2 Irreversible behaviour change 3 out of 5, and C3 Cost-structure change 2 out of 5. In Demerits: D1 Debt 0 out of minus 10, D2 Revenue quality minus 3 out of minus 10, D3 Governance minus 1 out of minus 5, and D4 Regulatory and social cost minus 1 out of minus 5.

Two gate verdicts precede the score. Gate G1, which asks whether growth continues or a second growth curve is present for a mature company, is recorded as a confirmed pass. Gate G2, which asks whether the ten-year ceiling of the new market is at least ten times current revenue, is recorded as a borderline pass on an estimated basis. The G2 bar is calculated in the score file as about $11.9 billion, derived from estimated fiscal 2026 revenue of about $1.19 billion multiplied by ten. That bar is demanding: it requires roughly tenfold growth over ten years, and the score file is explicit that the pass is borderline and estimated, not confirmed.

Speed: 16 out of 25 — fast orders, slower revenue

A1 Growth, scored 7 out of 10, captures the central contradiction of the Powell story. On the realised side, revenue growth decelerated from 44.8% in fiscal 2024 to 9.1% in fiscal 2025. Net income over the two-year span from fiscal 2023 to fiscal 2025 rose from $54.5 million to $180.7 million, a 3.3-fold increase, which shows that profitability expanded even as the top-line growth rate cooled. On the demand side, orders, book-to-bill, and backlog all point the other way. Quarterly orders reached $934 million, book-to-bill reached 3.0, and backlog rose from $1.4 billion to $2.4 billion. The score deducts two points for the slowdown in realised growth and one point for the size of the revenue base, at about $1.1 billion, arriving at 7. The logic is deliberately conservative: orders are not revenue, and a scoring system that rewarded orders as if they were already delivered would overstate what has actually been achieved. At the same time, the score does not ignore the order book. A book-to-bill ratio of 3.0 means that, in the quarter in question, new orders were three times the revenue recognised in that period, which is why backlog grew so quickly and why capacity, rather than customer interest, is identified as the constraint.

A2 Capital velocity, scored 3 out of 5, reflects a balance sheet that is strong but deliberately unhurried. Powell is debt-free in the sense recorded in the score file: there is no borrowings account, and the balance on its United States revolving credit facility was $0 at the September year ends in both 2024 and 2025. Cash and short-term investments stood at $633.6 million at 30 June 2026. Access to capital is therefore not the problem. The score is nevertheless only 3, because that capital is not being deployed at blitz speed. There is no share repurchase programme, treasury shares were unchanged over three years, the dividend payout ratio is about 7%, and a greenfield plant costing an estimated $70 million to $100 million is described as under review, not confirmed. Cash is accumulating rather than being converted into capacity, acquisitions, or returns to shareholders at a pace that would match the order book. The score file treats this as an observation about capital allocation, not as a leverage demerit, and that distinction is preserved in the demerits section below.

A3 Hiring and productivity, scored 3 out of 5, is consistent with the same picture of gradual rather than explosive scaling. Powell employed 3,143 regular employees plus 315 contract workers at the date recorded in the score file, with no unionised workforce. Revenue per employee in fiscal 2025 was about $351,000. Investment in production capacity is centred on the expansion of the Jacintoport fabrication yard, expected to be completed by the end of fiscal 2026, together with the greenfield plant under evaluation at $70 million to $100 million. Hiring is therefore described as focused on manufacturing capacity and as a progressive ramp-up rather than a surge. That is a coherent response to a backlog problem, but it is not the hiring blitz that the model's speed section is designed to reward, and the score of 3 reflects that.

A4 Share-gain velocity, scored 3 out of 5, is the most judgement-heavy of the speed items, and the score file is careful about its limits. No official market-share figure is published by the company, and any market-share percentage for Powell should therefore be treated as unconfirmed. What can be documented is segment momentum: utility revenue rose 50% in fiscal 2025 to $279.0 million, described in the score file as the result of strategic expansion efforts, and data-centre work moved from a first mega order in the first quarter of the fiscal year to a single order of more than $400 million in the third quarter. Against that momentum, the score file sets a counterweight: three consecutive quarters of earnings per share coming in below consensus, with third-quarter earnings per share of $1.42 against a consensus range of $1.47 to $1.49, on a press-reported basis. The score file explicitly states that it judges by numbers rather than by narrative, and that the gap between management's expressed confidence in maintaining trailing-twelve-month margins and the reported earnings figures is a reason for restraint. With no verified share figure to anchor a higher score, A4 is held at 3.

Moat: 28 out of 60 — an execution moat, not a network moat

B1 Network effects, scored 3 out of 18, is the lowest moat item relative to its maximum, and the reason is structural. Powell's business is engineered-to-order project work. Each system is designed for a particular customer and site, and the value of the product does not increase for one customer simply because another customer has bought a similar system. There is no platform network effect of the kind the model rewards most highly. The score file locates Powell in the lowest anchor range for this item. What Powell does have, in the file's account, is an execution moat: engineering capability, project management, systems integration, and customer relationships, with the company's own filing language describing Powell as a preferred service provider. The single data-centre order of more than $400 million is presented as evidence of that capability, because it required assembling a multi-division, multi-country execution plan. That is a real strength, but it is not a network effect, and the score of 3 keeps the distinction clear.

B2 Distribution scalability, scored 9 out of 18, is the strongest moat item in relative terms, and it is where the channel-shift thesis is tested. Utility revenue rose from 23% of total revenue in fiscal 2023 to 19% and then to 25% in fiscal 2025, while growing 50% in absolute terms in fiscal 2025. Data-centre distribution was opened from a first mega order to an order of more than $400 million. Liquefied natural gas work reignited, with about $60 million of orders in the third quarter, overseas revenue grew 35% to $224.1 million, and Powell acquired Remsdaq, a United Kingdom-based supervisory control and data acquisition company. Taken together, these are the marks of a company that has found new channels for an existing capability. Two discounts are applied, and both matter. First, distribution is not the bottleneck; production capacity is. Backlog grew far faster than revenue, and the Jacintoport expansion and the greenfield plant under review are the proposed remedies, not yet completed cures. Second, the single order of more than $400 million represents about 17% of backlog, so the quality of distribution growth carries a concentration risk that is examined further under revenue quality. The score of 9, at the upper end of the middle anchor range, reflects a distribution story that is genuinely under way but not yet proven at scale.

B3 Market size, scored 8 out of 12, follows from the borderline G2 pass. The relevant markets are electrical distribution equipment, data-centre power infrastructure centred on medium-voltage switchgear, liquefied natural gas, and grid reinforcement. The structural demand hypothesis, attributed in the score file to the chief executive, links artificial intelligence and data-centre power demand, expanding United States liquefied natural gas exports, and grid strengthening, and it is supported by the order figures already described, including $934 million of quarterly orders and a book-to-bill ratio of 3.0. The score is nevertheless held below the maximum because Powell does not disclose a quantified total addressable market, so that figure remains unconfirmed, and because reaching the $11.9 billion bar requires roughly tenfold growth over ten years, which presupposes that the data-centre and utility axes remain secular rather than cyclical. The score file notes that this bar is tight but not arithmetically impossible, distinguishing Powell's case from a case in which the bar exceeds the entire addressable market. The score of 8 records a large, structurally supported market with an unquantified ceiling.

B4 Gross margin, scored 8 out of 12, records a margin expansion that has been proven and has then stalled. Gross margin rose from 21.1% in fiscal 2023 to 27.0% and then to 29.4% in fiscal 2025, and stood at 29.6% for the first nine months of fiscal 2026. The quarterly peak was 31.4% in the fourth quarter of fiscal 2025, easing to 30.6% in the third quarter of fiscal 2026. The score file describes the result as expansion that has stopped inside a 29% to 31% box. The company's explanation, as recorded in the file, is volume leverage, strong project execution, and stable pricing, with the chief financial officer expressing an outlook of maintaining trailing-twelve-month levels. Two structural facts sit alongside that explanation. Raw material costs represent 45% of revenue, covering steel, copper, and aluminium, so input-price exposure remains material. And interest income represents about 9% of net income, a by-product of the debt-free balance sheet and large cash holdings rather than of operations. Margin proof is therefore established, but the next chapter of margin expansion is unconfirmed, and the score of 8 reflects that balance.

Creative destruction: 8 out of 15 — a beneficiary of transition, not its author

C1 Order being replaced, scored 3 out of 5, asks whether Powell is displacing an existing order. The honest answer in the score file is partial. A structural shift in power demand, driven by artificial intelligence, data centres, liquefied natural gas, and grid reinforcement, is changing the priorities of the markets Powell serves, and the mix transition from petrochemical work down 49% in a quarter to data-centre and utility work filling the gap is evidence of that shift. But Powell is not the agent replacing the old order; it is a supplier benefiting from the transition, competing in engineered-to-order work against companies including ABB, Siemens, Eaton, and Schneider. Displacement is under way, but Powell's role in it is bounded, and the score of 3 reflects that limit.

C2 Irreversible behaviour change, scored 3 out of 5, looks at whether customer behaviour is changing in a way that is hard to reverse. The score file points to behind-the-meter self-generation designs for data centres and to greenfield data-centre work in its first phase, as structural changes in how power is procured. The single order of more than $400 million is also relevant here: on a press-reported basis from the second-quarter call, its burn period is two to two and a half years, extending into fiscal 2028, which makes it a long-duration contractual commitment rather than a short, easily cancelled purchase. The discount is that orders still arrive project by project. A portfolio of long contracts is more durable than a single spot sale, but it is not the same as a permanent change in how an entire market behaves, and the score of 3 holds that distinction.

C3 Cost-structure change, scored 2 out of 5, is the weakest creative-destruction item. The Remsdaq acquisition is described as margin-accretive, and the Jacintoport expansion is intended to improve unit productivity. Against those efforts, the score file sets two constraints: raw material costs at 45% of revenue, and the engineered-to-order character of the business, which limits economies of scale. There is, in the file's words, no demonstrated collapse in cost structure. The score of 2 therefore records genuine productivity attempts without crediting a structural cost transformation that has not been shown.

Deductions and risks: minus 5 — clean leverage, concentrated backlog

D1 Debt is scored 0, meaning no deduction. This is the cleanest part of the score. Powell has no borrowings account, its revolving facility balance was $0, and it held $633.6 million in cash and short-term investments at 30 June 2026, alongside working capital of $606.5 million. There is no excess leverage to deduct for. The score file is careful to separate this from a different observation: the combination of no borrowings, large cash accumulation, no share repurchase programme, and a dividend payout ratio of about 7% raises a question about capital allocation efficiency. That question is recorded as an observation, not as a demerit, and this draft preserves that treatment rather than converting caution about cash into a leverage penalty.

D2 Revenue quality is scored minus 3 out of minus 10. On the positive side, no single customer exceeded 10% of revenue in either fiscal 2024 or fiscal 2025, as stated in the annual filing, and no accounting issue is identified in the score file. Two discounts drive the deduction. First, the customer behind the single order of more than $400 million is anonymous, so the counterparty risk on a project representing about 17% of backlog cannot be disclosed or assessed from public filings. Second, backlog itself is not necessarily a reliable indicator, because customers retain cancellation rights for their convenience, as stated in the annual filing, and the filing's risk factors state dependence on the cyclical nature of end markets. Backlog is therefore best read as visibility with conditions attached, not as guaranteed future revenue. The deduction of 3, rather than a larger figure, reflects that the underlying revenue recognition and customer diversification in reported fiscal years are otherwise sound.

D3 Governance is scored minus 1 out of minus 5. The capital structure is a single class of common stock, with no differential voting rights. The largest holders recorded in the score file are Thomas W. Powell, of the founding family, with 19.1% through TWP Holdings, BlackRock with 11.4%, Vanguard with 5.8%, and the chief executive with 1.4%. The file also records ten Form 4 filings in 2026 and numerous Form 144 filings, described as executive sale disclosures whose individual details are unconfirmed. There is no structural governance problem identified, but the combination of a founding-family block and executive sales is given a small deduction of 1. That is a proportionate treatment: concentrated ownership and insider sales are noted, but they are not presented as evidence of a governance failure.

D4 Regulatory and social cost is scored minus 1 out of minus 5. No material pending litigation is disclosed in the score file. The deduction reflects exposure to tariffs and raw material prices, with raw materials at 45% of revenue, an exposure explicitly listed among the risk factors in the annual filing. It is a small, specific deduction for a trade and input-cost risk, not a general regulatory judgement.

The largest single risk in the score file's own blitzscaling analysis deserves separate emphasis, because it cuts across several items at once. It is the anonymous $400 million order. A single project representing about 17% of backlog, with an undisclosed counterparty, means that a delay, disruption, or cancellation affecting that one project would affect backlog, revenue, and narrative together, rather than being absorbed across a dispersed customer base. The score file calls this the maximum risk from a blitzscaling perspective. It is not a reason to ignore the order, which is also the clearest evidence that Powell can win data-centre work at scale, but it is a reason to read backlog growth with the concentration qualifier attached.

Other risks follow from the structure of the business. Engineered-to-order project work is cyclical, and the score file records that cyclicality as mitigated by diversification rather than eliminated. The chief executive's description of that mitigation, on a press-reported basis from the second-quarter call, is reflected in the mix numbers, but a reader should not mistake mitigation for immunity. Input costs at 45% of revenue leave margins exposed to steel, copper, and aluminium prices and to tariffs. Three consecutive quarters of earnings per share below consensus, on a press-reported basis, show that reported results have trailed the expectations set around them, even in quarters that were records in absolute terms. Third-quarter revenue of $311.7 million against $316.7 million in the comparison figure cited in the score file points in the same direction: absolute records and relative misses can coexist, and the score judges by the reported numbers.

What to watch

The following items are future markers only. They are listed as matters to observe when new public information appears, without any anticipated outcome attached to them.

First, fourth-quarter and full-year fiscal 2026 results, which had not been reported at T. The score file notes an expected date of around 17 November 2026 for those results, explicitly labelled as company-unconfirmed, and those results are excluded from the present score. When they are published, the relevant observations will be the relationship between orders, backlog, and recognised revenue, and whether the margin range described above is maintained.

Second, the Jacintoport fabrication yard expansion, expected to be completed by the end of fiscal 2026. The relevant observation will be whether capacity changes are reflected in subsequent reported delivery and revenue figures.

Third, the greenfield plant evaluated at $70 million to $100 million. At T this project is under review and unconfirmed. The relevant observation will be whether a decision is announced and, if so, on what terms as disclosed in public filings.

Fourth, the single data-centre order of more than $400 million, with a burn period of two to two and a half years extending into fiscal 2028 on a press-reported basis from the second-quarter call. The relevant observation will be how that project progresses as disclosed in subsequent filings and reporting, given its weight in backlog and its anonymous counterparty.

Fifth, the segment mix. The relevant observations will be whether petrochemical revenue continues to decline at the pace recorded in the scoring period, and whether utility and commercial and other industrial revenue continue to offset that decline in reported segment figures.

Sixth, capital allocation. The relevant observations will be the cash and short-term investments balance in subsequent filings, and whether any share repurchase programme, dividend change, or capacity investment is disclosed.

A note on figures in the score file

This draft does not reconcile or adjust any figure from the score file. Two internal tensions in that file are preserved here rather than smoothed over, so that a reader adding up the file's own numbers can see both sides.

First, backlog growth is given as 71% in the gate and growth-item sections, where backlog rises from $1.4 billion to $2.4 billion, and as 69% in the list of positive points in the same file. Both figures are quoted in this draft's sources of evidence as they appear in the score file, and neither has been chosen over the other.

Second, the size of the first-quarter data-centre mega order is given in two forms in the score file. The gate section describes data-centre orders in commercial and other industrial work as more than $100 million in the first quarter, rising to a single order of more than $400 million in the third quarter. The share-gain item describes the first mega order in the first quarter as about $75 million, rising to more than $400 million. Both descriptions are preserved here, and no single first-quarter figure has been adopted as definitive.

Labels from the score file are likewise preserved. Market-share percentages are unconfirmed, because the company does not officially disclose a market-share figure. The three consecutive quarters of earnings per share below consensus are press-reported, not filing-confirmed, and the third-quarter comparison of $1.42 against $1.47 to $1.49 should be read on that basis. The expected date for fourth-quarter and full-year fiscal 2026 results is company-unconfirmed. The greenfield plant cost of $70 million to $100 million is under review and unconfirmed. Individual details of the 2026 Form 4 and Form 144 filings are unconfirmed. Where the score file marks a figure as confirmed, estimated, or press-reported, this draft keeps that marking.

Powell at T is therefore best understood as a company that has proven it can win the next market's orders and expand its margins, that has the balance sheet to fund the next stage without borrowing, and that has not yet proven, in recognised revenue and earnings, that it can deliver the backlog at the pace the order book implies. The score of 47 out of 100, grade C, records that position faithfully: strong demand evidence and a clean balance sheet, held back by a moat that rests on execution rather than networks, by creative destruction that is real but bounded, and by the simple fact that orders are not yet revenue.

More Power & Energy companies

Other Power & Energy names on the Blitz Index, highest score first: Solaris Energy Infrastructure (59/C), Bloom Energy (58/C), Forgent Power Solutions (50/C).

Next earnings

Nov 16, 2026 · After close · Confirmed

Coverage

Powell Industries, Inc. · Q4 FY2026 due ~Nov 17, 2026 (expected) · Score 47 · Grade C

Embed this score

Copy the HTML below to show this score on your site. The badge links back to this page and updates automatically when the score changes.

<a href="https://blitzlab.app/companies/powl/"><img src="https://blitzlab.app/badge/powl.svg" alt="Powell Industries Blitz Index 47/C — BlitzLab"></a> Copy HTML

Preview:

BlitzLab scores businesses against a scholar’s standard — not investment advice. No price targets, no buy/sell recommendations. Figures are from SEC filings (XBRL) as labeled in the source scorecard.