AMSC · American Superconductor Corporation

Refresh of coverage first published 2026-04-22; advances the 2026-09-02 note.

AMSC — American Superconductor Corporation

Current Thesis

The 2026-09-02 close of $28.49 is the same number as the 2026-08-28 close of $28.49 — the close that had taken out the 2026-08-24 closing low of $28.92. Six sessions apart the tape printed an identical price, the first thing it has done since 2026-08-05 that is not a lower close. RSI(14) read 28.9 against 26.3 on 2026-09-01, so the oscillator lifted 2.6 points while price slipped seven cents from $28.56. A decline that has stalled is not the same object as a base, and that distinction is the entire question here.

The leg an investor would be buying is mean reversion against a compounding order book. The order book is measured: Q1 FY2026 revenue of $94.073M (+30% YoY) beat the $87.800M consensus by $6.3M when reported 2026-08-05; quarterly orders exceeded $130M; twelve-month backlog stood above $300M and total backlog above $400M at 2026-06-30; cash, equivalents and restricted cash were $153.1M against $4.00M of total debt. What stays inferred, and what four months of tape has not supported, is that any of it produces a bid.

The narrative is dead rather than merely tired, and four dates carry that reading: the 2026-05-28 beat-and-soft-guide that opened the de-rating, the late-June 2026 Russell 3000E and Russell Microcap deletions that removed the passive holder base, 2026-08-05 when record revenue, record orders and record backlog were met with lower closes, and 2026-08-28 when the August closing low went. No new company flow has arrived behind the drawdown — the last dated items are the 2026-08-05 results, the 10-Q filed 2026-08-06, Roth Capital's 2026-08-10 target cut to $55 from $58 with the Buy kept, and an investor-presentation posting on 2026-08-11.

Bullish and bearish views on American Superconductor Corporation

The model's bull view on American Superconductor Corporation (AMSC), in brief: Q1 FY2026 (quarter ended 2026-06-30, reported 2026-08-05): revenue $94.073M versus $72.4M a year earlier and $87.800M consensus. The bear view: Margin, not demand, is where Q1 broke: adjusted EPS of $0.16 missed the $0.21 estimate on a $6.3M revenue beat (2026-08-05). Both cases follow in full.

Bull Case

• Q1 FY2026 (quarter ended 2026-06-30, reported 2026-08-05): revenue $94.073M versus $72.4M a year earlier and $87.800M consensus. Grid $76.3M, Wind $17.8M.
• Orders exceeded $130M in the June quarter; excluding the single largest recently announced award, bookings still exceeded $100M against roughly a $70M quarterly average in the prior fiscal year (CEO Daniel McGahn, 2026-08-06 call).
• Twelve-month backlog above $300M and total backlog above $400M at 2026-06-30, versus roughly $280M twelve-month at 2026-03-31.
• The 2026-07-28 award is $25M of turnkey integrated power quality for a North American utility serving a large-scale mining development — STATCOM, capacitor banks, shunt reactors, a 138kV transformer — for delivery in fiscal 2027, the company's largest single mining-development order to date.
• Operating cash flow of $16.0M in the June quarter; cash, equivalents and restricted cash of $153.1M at 2026-06-30 against $147.6M at 2026-03-31; total debt $4.00M. No financing event sits behind the drawdown and no dilution has been announced into it.
• Revenue guidance has behaved as a floor: the 2026-05-28 guide of "above $85.0M" produced $94.073M, and the Q2 guide issued 2026-08-05 repeats the identical wording.
• Sell-side has not capitulated. Four analysts polled by S&P Global carry a Strong Buy consensus averaging $61.75 (low $54, high $71) as of 2026-09-02, and Roth's 2026-08-10 cut kept its Buy.
• The 2026-09-02 close matched the 2026-08-28 close to the cent while RSI(14) rose from 26.3 to 28.9, the first non-confirming low in the sequence that began 2026-08-05.

Bear Case

• Margin, not demand, is where Q1 broke: adjusted EPS of $0.16 missed the $0.21 estimate on a $6.3M revenue beat (2026-08-05).
• The Q2 guide issued 2026-08-05 sits below consensus on every profit line — GAAP net income above $1.0M ($0.02) against $0.12, non-GAAP net income above $8.0M (about $0.17) against $0.23, revenue above $85.0M against $87.55M.
• The 50-day moving average at $33.90 and the 200-day at $35.96 (2026-08-30) are both above the last close, against $39.12 and $38.17 on 2026-08-05; the 50-day crossed under the 200-day inside that window.
• Passive demand is structurally gone. The late-June 2026 Russell 3000E and Russell Microcap deletions removed the index holder base, and the 2026-09-18 rebalance is the first quarterly liquidity event since.
• Short interest is building into the decline: 3.48M shares, 7.44% of float, 4.55 days to cover on 2026-08-30, against 3.40M, 7.27% and 3.95 days on 2026-08-05, on 48.44M shares outstanding.
• Sell-side targets averaging $61.75 against a $28.49 close leave roughly 2.2x of gap that can close by revision rather than by rally; each cut arrives as a negative headline, as Roth's did on 2026-08-10.
• A five-year beta near 3.28 means any de-rating of the AI-power and electrical-equipment complex arrives amplified, regardless of what the backlog does.
• Trailing EPS of $3.05 rests on FY2025 GAAP net income of $133.81M, which included a deferred-tax valuation-allowance release — the headline multiple is not comparable to a peer operating multiple.

Setup & Price Structure

The last completed close is $28.49 (2026-09-02), 57.3% below the $66.68 52-week high on the split- and dividend-adjusted series, with a three-month price change of -38.6%. stockanalysis.com lists the 52-week range as $24.87 to $70.49 intraday, so the August–September shelf sits roughly 15% above the year's intraday low.

The structure is a stalled downtrend under declining averages. Both the 50-day at $33.90 and the 200-day at $35.96 (2026-08-30) sit above price, and the death cross formed between the 2026-08-05 and 2026-08-30 readings. What changed in one session is the shape of the low: two identical closes at $28.49 on 2026-08-28 and 2026-09-02, with RSI(14) 2.6 points higher on the second. That is a condition for a base, not a base — the level has been touched twice and never reclaimed upward, and the first genuine confirmation would be a weekly close back above the declining 50-day at $33.90.

On crowding, the observables point the opposite way from a late-stage momentum name. There is no retail-sentiment cluster and no distance above a rising moving average to unwind; price sits below both major averages and the passive holder base was deleted in late June. The crowded side is short — 7.44% of float and 4.55 days to cover on 2026-08-30, up from 7.27% and 3.95 days on 2026-08-05 — and long in the sell-side book, where a $61.75 average target faces a $28.49 close. Institutional ownership stands at 69.61%. Insider selling exists but is mechanical: CEO and CFO sales run under Rule 10b5-1 plans (CEO plan adopted 2025-08-15, modified 2026-02-26) covering tax withholding on vesting restricted stock, not discretionary distribution into strength. No equity issuance has been announced into the decline.

Catalyst Calendar (next 30 days)

• 2026-09-18 — September quarterly index rebalance and quadruple witching. First quarterly liquidity event since the late-June Russell 3000E and Microcap deletions; a flow date carrying no fundamental information, and the only scheduled item inside 30 days that moves size.
• ~2026-09-25 (est.) — FINRA short-interest publication covering the mid-September settlement date. Second data point on whether the build between the 2026-08-05 and 2026-08-30 readings continues at the lower price.
• 2026-09-30 — Close of the September quarter covered by the guided floors of revenue above $85.0M and non-GAAP EPS around $0.17.
• Nothing company-scheduled resolves the guide sequence inside the window. The next fundamental trigger is the Q2 FY2026 print, ~2026-11-04 (est.), outside 30 days.

Elapsed catalysts

• ~2026-09-10 (est.) — FINRA short-interest publication covering the end-August settlement date. First reading fully spanning the 2026-08-28 break of the August low; separates fresh short demand at the lower level from covering into it. (passed 1d ago)

What Would Change Our Mind

The whole structure now rests on one number: the $28.49 shelf set by the identical closes of 2026-08-28 and 2026-09-02. Lose it and the RSI divergence that formed across those six sessions is void, with nothing between price and the stockanalysis-listed $24.87 52-week intraday low. A weekly close below $27.00 grades that break. A second condition would compound it: the 2026-09-18 rebalance passing without a volume expansion, confirming that the post-deletion holder base has not been replaced by a discretionary one.

On the fundamental side, the read turns worse if the Q2 FY2026 print (~2026-11-04 est.) shows revenue clearing the $85.0M floor while gross margin declines sequentially again — that would make the 2026-08-05 EPS miss a trend rather than a Comtrafo purchase-accounting artifact — or if twelve-month backlog contracts from the above-$300M level reported at 2026-06-30.

The read turns constructive on a weekly close above $33.90, reclaiming the declining 50-day, ideally alongside a FINRA reading showing short interest below the 3.48M shares of 2026-08-30. Absent that, the base has not formed and there is nothing here to chase.

Correlation Notes

• A five-year beta near 3.28 means the name trades as roughly 3x the broad tape, so sector risk-off in the AI-power and electrical-equipment complex arrives amplified in both directions.
• The cleanest same-session comparables are GE Vernova and Powell Industries; AMSC making lower lows alongside both indicates complex de-rating rather than company-specific news, which matters because company-specific news has been absent since 2026-08-11.
• After the late-June 2026 deletions from the Russell 3000E and Microcap indices, the name no longer moves with microcap index flow, which removes both a source of forced selling and a source of mechanical rebalancing demand.
• Wind revenue of $17.8M against Grid's $76.3M in Q1 ties roughly one fifth of the top line to a lumpy, concentrated order cycle unrelated to data-center demand.