GNRC — Generac Holdings Inc.

Two forces are pulling in opposite directions. Fundamentals look constructive: Q2 2026 (reported 2026-07-29) net sales rose ~11% YoY, adjusted EPS of $2.91 crushed the ~$1.95 consensus, C&I sales grew ~29% to $556M on large-megawatt demand, and the Street sits at a Buy consensus with price targets averaging ~$298 (low $214, high $340). The tape disagrees: as of 2026-09-13 the stock is -15.7% over one month and -28.7% over three months, at $186.99 versus a 52-week high of $296.44. Investors appear to be discounting the fact that ~$71M pre-tax (~$0.89/share) of the Q2 beat came from one-off tariff refunds, and waiting to see whether the June 2026 hyperscale supply agreement actually converts into shipped megawatts.

The three-year record shows a trough-and-inflection, not a structural decline. Revenue grew from $1,070.7M in Q3 2023 to a peak $1,234.8M in Q4 2024, then went negative YoY (-5.0% in Q3 2025, -11.6% in Q4 2025) before snapping back to +12.4% in Q1 2026 ($1,059.4M) and +10.6% in Q2 2026 ($1,173.5M). Margins follow the same arc with one important caveat: gross margin bottomed at 36.3% in Q4 2025 — the quarter that produced a $9.3M operating loss and -$0.42 diluted EPS — then recovered to 38.7% in Q1 2026 and 44.5% in Q2 2026, but that Q2 figure includes roughly 6 points of gross margin from the $71M tariff refund, so underlying gross margin is closer to ~38.5%, roughly flat YoY. Q2 2026 operating income of $210.4M is the highest in the trailing three-year window, versus -$9.3M in Q4 2025 and $198.0M in Q4 2024. Free cash flow is recovering off a 2025 trough — trailing-4Q FCF went from $268.1M (through Q4 2025) to $379.2M (through Q2 2026) — but is still below the ~$486M generated in the year through Q2 2025, even though the most recent quarters are sharply better YoY ($89.9M in Q1 2026 vs $27.2M; $62.9M in Q2 2026 vs $14.5M). Net: accelerating off a trough, but the reported magnitude is flattered by a one-off.

As of 2026-09-13, at $186.99 GNRC screens expensive on trailing numbers and reasonable on forward ones: 42.4x TTM P/E on $4.405 of trough-distorted EPS, 3.81x book, 2.48x sales (just under its own 5-year average of 2.66x), 22.0x trailing EV/EBITDA, and a PEG of -1.53 that is meaningless because TTM EPS sits near a cyclical low. The forward math is the real story — against next-FY consensus EPS of $9.723 the stock trades at roughly 19.2x, a steep discount to its 5-year average TTM P/E of 35.4x. That discount is not a free lunch; it is the market pricing a near-doubling of EPS from $4.41 to $9.72, driven by C&I data center demand rather than by anything already delivered. My read: fair-to-slightly-cheap. ~19x forward is defensible for a mid-teens grower with a 91.2/100 financial-health score, but the 5.3% TTM ROIC and 5.8% net margin cap how much multiple the business deserves. (Note: the 13.7% "dividend yield" in the data feed is almost certainly a feed error and was excluded from this analysis.)

The gap between 42x trailing and ~19x forward IS the trade. The market has already priced a full earnings recovery — roughly +120% EPS growth from TTM $4.41 to consensus $9.72 — so there is no valuation cushion on trailing numbers if estimates slip. That is why a quarter with a big headline beat still left the stock down 28.7% over three months: strip the tariff refund and Q2 2026 adjusted EPS was roughly $2.02 against the $1.95 consensus, a ~3% beat rather than a blowout. The playbook is to treat the forward estimate as the thing on trial, not the reported beat. Two confirmations matter: (1) Q3 2026 gross margin holding in the high-30s with no tariff refund; (2) C&I revenue sustaining its ~29% YoY pace as the hyperscale agreement converts. If both hold, ~19x forward is too cheap; if either breaks, the stock can retest the $150s given how far TTM earnings sit below the estimate.

As of 2026-09-13, the technicals are decisively negative rather than merely neutral. RSI(14) is 40.4 — weak, but not oversold, leaving room to fall further before a washout signal. At $186.99, the stock sits 4.9% below its 20-day SMA (~$196.6), 10.5% below its 50-day (~$208.9), and 10.7% below its 200-day (~$209.4); the 50- and 200-day averages have converged near $209, a setup that typically resolves into a death cross and persistent overhead supply. Momentum: -15.7% over one month and -28.7% over three months, but only -0.19% over the past week, so the selling is at least decelerating. Against a 52-week range of $134.80-$296.44 the stock sits in the lower third. Entry implication: there is no technical reason to rush — better odds come from an RSI washout below 30, a reclaim of the 20-day near $197, or the November 4 Q3 print confirming the fundamental trend.

• AI/data center power demand — grid congestion and hyperscale buildouts are the single biggest swing factor for large-megawatt C&I generators and microgrids; the Belvidere, IL plant expansion and the June 2026 global hyperscale supply agreement are direct plays on it.
• Grid reliability and severe-weather frequency — drives baseline residential standby demand, but it is weather-lumpy and can pull demand forward rather than create it, making segment results noisy quarter to quarter.
• Tariffs and trade policy — Q2 2026 included a $71M pre-tax tariff refund worth roughly 6 points of gross margin; the same policy channel can reverse and raise steel and engine input costs with little warning.
• Interest rates and the housing/renovation cycle — home standby systems are big-ticket and frequently financed, and with home turnover low the residential segment was still -2% YoY in Q2 2026; rate cuts would be a meaningful tailwind.
• Electrical-equipment peer multiple rotation — GNRC trades alongside higher-multiple electrification names (Eaton, Vertiv, GE Vernova), so sector rotation in or out of that theme can move the stock more than its own results do.

• Q3 2026 earnings, expected around 2026-11-04 — the key test of gross margin ex-tariff-refunds, C&I backlog conversion, and whether management resets full-year guidance.
• Conversion of the June 2026 global hyperscale supply agreement into formal Master Supply Agreements with announced megawatt volumes; further MSA news is the most binary catalyst on the board.
• A return to YoY growth in the residential segment as channel inventory normalizes, aided by a major outage event or lower financing rates.
• Ramp of the Belvidere, IL large-megawatt packaging facility — expands C&I capacity, but adds fixed cost if data center volumes disappoint.

• Next-FY EPS of $9.723 is unproven, and the Q2 2026 beat leaned on ~$0.89/share of one-off tariff refunds; ex-refund, adjusted EPS was roughly $2.02 vs the $1.95 consensus.
• Data center concentration and lumpiness — a hyperscaler capex pause, or a shift toward on-site turbines/fuel cells instead of packaged gensets, would undercut the C&I story the forward multiple now rests on.
• Residential remains soft (-2% YoY in Q2 2026) with competition from battery storage and low home turnover; a mild storm season would leave the segment flat and force the whole recovery onto C&I.
• Weak underlying returns and earnings quality — TTM ROIC of 5.3%, ROE of 6.1%, net margin of 5.8%, and Q4 2025 showed the downside with a $9.3M operating loss on $1.09B of revenue.
• High beta (1.91) and a 52-week range of $134.80-$296.44 — a spread of roughly 120% means position sizing and entry timing matter more than usual here.

Neutral, with a bias toward patient accumulation on weakness rather than buying here. The data center power thesis is genuine, and ~19.2x forward EPS is not a demanding multiple for a business with a 91.2/100 financial-health score ($1.42B long-term debt against $265M cash, current ratio 2.05). But the recovery is less proven than the Q2 headline suggests: strip the $71M tariff refund and the margin expansion and the EPS beat mostly evaporate, TTM ROIC is only 5.3%, and the stock is below all three major moving averages after a 28.7% three-month decline. My fair value estimate of ~$214 implies about 14% upside — real, but not enough to override a weak tape and a forward estimate that still has to be earned. Wait for the November 4 print.

I applied a 22x forward P/E — a mid-cycle multiple for a high-beta (1.91) but financially strong (91.2/100 health score) industrial growing EPS off a trough, and well below GNRC's own 35.4x five-year average TTM P/E — to next-FY consensus EPS of $9.723, giving 9.723 x 22 = $213.91. Cross-checks land in the same place: 2.55x forward sales (a modest discount to the 5-year average P/S of 2.66x) on $4,939M of next-FY revenue across ~59M diluted shares implies ~$214, and 18x forward EBITDA of roughly $770M less ~$1.16B net debt on ~59M shares implies ~$215.