AGX · Argan, Inc · Stock research

Current Thesis

Argan builds U.S. gas-fired generation through Gemma Power Systems — the asset class hyperscalers now contract for firm, round-the-clock AI load. The demand side is unchanged since the June print: ~$2.8B backlog at 2026-04-30, anchored by 4.1+ GW of combined-cycle work, funded off $973.6M of cash and investments with zero debt. The equity is the moving part. After round-tripping the June gap, the decline stopped: $563.06 close on 2026-07-25, $578.22 on 2026-08-14, RSI(14) 56.3, three-month return -19.9%, 27.5% under the $797.9 52-week high. That is a first higher low with the $600 shelf and the ~$675 June breakout shelf both overhead and unreclaimed. The narrative is maturing — the upstream theme still gets funded (Alphabet raised its 2026 capex outlook at its late-July print; hyperscaler capex scrutiny was the market story on 2026-07-28), while AGX-specific news flow has thinned to an $8.3M teledata bolt-on (2026-08-04) and a 15-year-return retrospective (2026-08-05). The Q2 FY27 print — unscheduled, due around 2026-09-03 on last year's 2025-09-04 cadence — decides whether the first QoQ backlog dip was an air pocket.

Bullish and bearish views on Argan, Inc

The model's bull view on Argan, Inc (AGX), in brief: Q1 FY27, reported 2026-06-04: revenue $291.0M, +50.2% YoY; diluted EPS $3.24 vs $2.33 consensus; adjusted EBITDA $56.4M, +79%; gross margin 19.0% → 21.0%. The bear view: The June leg is fully retraced. From the $797.9 52-week high to $578.22 on 2026-08-14 is -27.5%; the path down went through $630.32 (2026-07-10), $599.74 (2026-07-14) and $563.06 (2026-07-25). Both the ~$675 and $600 shelves are now resistance. Backlog fell QoQ for the first… Both cases follow in full.

Bull Case

• Q1 FY27, reported 2026-06-04: revenue $291.0M, +50.2% YoY; diluted EPS $3.24 vs $2.33 consensus; adjusted EBITDA $56.4M, +79%; gross margin 19.0% → 21.0%. Margin expanded on a record top line.
• Backlog ~$2.8B at 2026-04-30, ~79% gas, anchored by CPV Basin Ranch — 1,350 MW combined-cycle in Ward County, TX, full notice to proceed 2025-10-30, GE 7HA.03 turbines, completion scheduled 2028 — which recognizes revenue across FY27–FY28.
• $973.6M cash and investments, zero debt at 2026-04-30: growth and capital return are self-funded, with no issuance overhang to absorb.
• Capital return intact through the drawdown: buyback raised from $150M to $200M and extended through 2030-01-31 (2026-04-08); dividend at $0.50/qtr, $2.00 annualized.
• Corporate action restarted on 2026-08-04: Southern Maryland Cable acquired ValCor Communications for ~$8.3M, extending the teledata footprint into New England — small in dollars, and the first deal announced since the price leg broke.
• Sell-side targets still sit above the market: 5-analyst consensus Buy, average target $679.80, implying 17.6% upside from the 2026-08-14 close (stockanalysis.com, 2026-08-15).

Bear Case

• The June leg is fully retraced. From the $797.9 52-week high to $578.22 on 2026-08-14 is -27.5%; the path down went through $630.32 (2026-07-10), $599.74 (2026-07-14) and $563.06 (2026-07-25). Both the ~$675 and $600 shelves are now resistance.
• Backlog fell QoQ for the first time: $2.929B at 2026-01-31 → ~$2.8B at 2026-04-30 — book-to-bill under 1, revenue burning the book faster than awards refill it.
• Still ~50x trailing: market cap $8.11B, P/E 50.82 (stockanalysis.com, 2026-08-15). A second soft backlog quarter compresses that multiple even if the demand story holds.
• Target risk is asymmetric downward: the $679.80 average and Lake Street's $600 Hold (raised from $375 on 2026-06-05) both sit above the market, so the next revisions are more likely cuts toward price than raises.
• No new EPC award release since the 2025-10-30 full NTP. The award cadence that produced the re-rate has not printed a headline through the entire drawdown; the 30-day news set is a bolt-on acquisition, an options-flow screen (2026-08-04) and a retrospective performance piece (2026-08-05).
• Top-customer concentration above 60% per the 10-K risk factors — a single project schedule slip produces outsized single-day gaps in a name with a thin float and a ~50x multiple.

Setup & Price Structure

• Reference levels: 2026-08-14 close $578.22. The 2026-07-25 close of $563.06 is the swing low the current higher low depends on. $600 (round number and Lake Street's target) is the first proof point; ~$675 is the June breakout shelf; the pre-run $500s are the base the July note flagged.
• Momentum: RSI(14) 56.3 — restored to neutral by the bounce, with no overbought condition to work off. The 2026-08-14 session closed +3.46%.
• Crowding and positioning observables, stated as observed: price trades below every published target in the 5-analyst set (average $679.80); trailing P/E 50.82; an earnings print falls inside the next ~3 weeks and has not yet been scheduled; AGX appeared on an industrials options-flow screen on 2026-08-04; retail-facing coverage in the window is a 15-year total-return retrospective (2026-08-05), the second of that shape this year (2026-03-31); the recent filing feed shows no insider Form 4 activity and there is no equity issuance to absorb.
• What the structure is: a base attempt off the July low, unconfirmed until $600 is taken back on a weekly close. Buying here is buying the reclaim before the backlog number that would justify it.

Catalyst Calendar (next 30 days)

• ~2026-09-03 (est.) — Q2 FY27 results, quarter ended 2026-07-31, with the conference call the same afternoon. The backlog figure is the number that matters: a rebuild above $2.9B ends the book-to-bill question; a second sequential decline confirms it.
• ~2026-09-03 (est.) — quarterly dividend declaration, which has historically accompanied the release; the rate has been raised three consecutive years, most recently to $0.50/qtr.
• Ongoing through the window — GE Vernova turbine order and lead-time commentary. AGX awards have lagged OEM bookings by roughly 2–4 quarters, so GEV headlines read forward on Argan's funnel; GEV's next scheduled report falls outside this window.

Elapsed catalysts

• ~2026-08-24 (est.) — Q2 FY27 earnings-date announcement. Argan issues a scheduling release ahead of the print; last year's Q2 release landed 2025-09-04. The notice fixes the binary date. (passed 2d ago)

What Would Change Our Mind

The base attempt is the whole constructive case, so it fails when the July low stops holding: a weekly close below $555 undercuts the 2026-07-25 close of $563.06, turns the higher low into a lower low, and puts the pre-run $500s and then the $400s back in play as the mean-reversion target. Fundamentally, the read breaks if the ~2026-09-03 print shows backlog below ~$2.8B — a second straight sequential decline and book-to-bill under 1 for two quarters — or if gross margin retraces below the 19.0% of Q1 FY26 on fixed-price EPC execution. A covering analyst cutting a target beneath the market price would confirm that sell-side support is repricing to the tape rather than the tape catching up. In the other direction, a weekly close above $600 alongside a backlog rebuild above $2.9B would re-establish the leg that broke in June, with ~$675 the next structural level. If the print date comes and goes with backlog flat and no new EPC award announced, the theme moves toward saturated for this name regardless of what hyperscaler capex does.

Correlation Notes

• GE Vernova (GEV) is the upstream indicator: turbine slot availability and bookings set Argan's award pipeline 2–4 quarters later. AGX trades as a higher-beta derivative of GEV's tape.
• Hyperscaler capex prints drive the funnel. Alphabet raising its 2026 capex outlook at its Q2 print, and the sell-off in peers that followed on 2026-07-28, showed capex direction and equity reaction can diverge — the spend supports Argan's order book even when the AI complex de-rates.
• E&C comparables (PWR, EME, PRIM, MYRG) move on the same power-buildout flow; relative weakness in AGX against that group while backlog is under question separates a company issue from a sector rotation.
• Rates and small-cap multiple beta: at ~50x trailing on an $8.11B cap, AGX carries duration risk that the backlog does not — repricing of long-duration growth hits it before any project news does.
• Natural gas and ERCOT policy set the sponsor economics behind the ~79% gas backlog; the ~13% renewable sleeve carries separate policy beta.