Phase Space AI

Valuation

Enphase Energy [ENPH]

Enphase Energy [ENPH] — Valuation

Spot $35.05 (2026-07-29 close) · EV $4,266.8m · TTM revenue $1,328.5m · EV/Sales 3.21x Two outputs, two horizons. Reporting only one is a defect.

---

1. Fixed inputs, named

---

2. Implied-path test (Valuation Criteria) — FAIL

2.1 Primary formulation: on EV/Sales, where the multiple anchor IS identified

The EV/EBIT route requires a terminal operating margin, and for this company that quantity is the thing in dispute — reported operating margin is 152–221% financed by a statute that is zero from 2033. The EV/Sales formulation removes that unanswerable parameter and puts all the weight on the multiple, which is observable in the company's own trading record.

Own multiple history. Point-in-time daily EV/Sales, using TTM revenue as known at each date (lagged to filing) and the share count and net cash from the most recently filed balance sheet at each date. Window 2025-01-01 → 2026-07-29, n = 393 — chosen because the full 2020-2026 history spans a regime change of 1.90x to 84.5x and cannot be pooled.

The multiple is not the problem. It sits within 1½ points of its own median.

Required five-year revenue CAGR to justify $35.05. Solved with assets/reverse_dcf.py, --solve cagr, multiple_basis sales, all parameters in §1 held fixed:

Result: FAIL. At its own median multiple the price requires a +11.7% five-year revenue CAGR. The demonstrated three-year CAGR is −14.2%; the latest quarter is −19.6% year on year, and −42.8% year on year excluding safe-harbour revenue (Q2-2026 ex-SH $207.6m against a Q2-2025 base with no disclosed safe-harbour component). Margin −25.9 percentage points.

Implied compression, stated as a number: today 3.21x EV/Sales → exit 2.98x = 0.23 turns of compression (−7.2%). The test does not fail because of an aggressive multiple haircut. It fails because a business shrinking 14% a year is priced for growing 12% a year.

Is there a "PASS WITH ARGUMENT"? The rule requires a specific, evidenced reason — a named product cycle, mix shift or pricing action. The candidate is the IQ Solid-State Transformer data-centre programme. Its current commercial evidence is "a few opportunities advancing to the RFI and RFP stages" and a "potential multi-gigawatt pipeline." An RFI is not a mix shift. That is narrative, and narrative does not qualify. The GaN cost-down (IQ9N/IQ9S-3P, IQ Battery G5 at ~50% higher energy density) is genuine and evidenced but it is a margin lever, not a 26-point revenue-growth lever. FAIL stands.

2.2 Cross-check on EV/EBIT, with the terminal margin taken from the best product year ever

Terminal margin = 19.23% — FY2022 operating margin, the last clean pre-45X year (revenue $2,330.9m, operating income $448.3m, zero AMPTC recognised). This is deliberately the most generous defensible number: it is the best product operating margin the company has ever recorded, achieved in the peak of the US residential boom with a 30% federal consumer credit in force and before NEM 3.0. The exit EV/EBIT multiple is declared UNIDENTIFIED — the company's own EV/EBIT history splits into a pre-45X regime and a subsidy-inflated regime, and on a product basis EBIT is negative in every period since 2023, so no growth-matched anchor exists. The sensitivity is therefore the output, not a point estimate.

The EV/Sales and EV/EBIT routes agree to within 0.7pp at their reference points (−25.9pp vs −26.6pp), which is the useful thing: the conclusion does not depend on which unobservable you choose to fix.

2.3 Inverting it — what terminal margin the price requires

--solve terminal_margin, exit 15x EBIT, all else fixed:

The best operating margin Enphase has ever reported, on any basis, is 19.5% (FY2023, subsidy-inclusive) and 19.2% (FY2022, clean. So even granting five years of +10% revenue growth — against a −14.2% demonstrated CAGR — the price requires a terminal operating margin above anything the company has achieved, in a perpetuity where §45X is zero. At flat revenue it requires 34.5%, which is nearly double the record.

2.4 What the screen's version did instead

The screen reported valuation: INDETERMINATE with the note "reverse DCF unsolvable in range", using terminal_margin: 0.107 (FY2025 reported operating margin — 152% §45X) and exit_multiple: 18.0 with exit_multiple_basis: "UNIDENTIFIED" and exit_multiple_peer_n: 0. Two independent errors partially cancelled: a terminal margin too high by roughly the whole subsidy, against a multiple that was declared unidentified and used anyway. The correct handling of a declared-unidentified multiple is to propagate INDETERMINATE or to run the sensitivity — not to substitute a default.

---

3. Twelve-month target — $29, −17.3% to spot

Built per valuation.md: near-term estimates, named product-cycle events inside twelve months, multiple anchored on the name's own trading range with the percentile stated. Not a DCF; not a peer median.

3.1 Near-term revenue base

No consensus pull was made — the Alpha Vantage 25/day quota is shared across the cluster. Two bases are reported rather than one, because the choice matters and it is a judgement:

Both are below TTM revenue of $1,328.5m. That is the point: the near-term estimate line is falling, and it is falling for a dated statutory reason, not a cyclical one.

3.2 Multiple, on its own history

Anchor at the 50th percentile of the 2025-01-01 → 2026-07-29 window, 2.98x. The stock trades at 3.02x today, the 54th percentile. No haircut is applied and none is needed — the base case holds the multiple essentially where it is.

3.3 The target

Equity value = NTM revenue × exit EV/Sales + net cash $364.9m, ÷ 132.145m shares.

Base 12-month target: $29 (the p50 across both revenue bases), −17.3% to spot. Range: $25 – $40, set by the p25 and p90 of the name's own multiple distribution.

3.4 Why a below-spot target is the correct output here, not a house view

Calibration item B16 recorded 16 of 16 house targets below spot, a median 46.1% below Street, and the standing instruction is that targets above spot should be common. This target is below spot and the reason is mechanical and checkable:

• The multiple is set at the 50th percentile of the name's own recent history and the stock trades at the 54th — the multiple assumption contributes essentially nothing to the gap.
• The whole gap comes from NTM revenue being 8–14% below TTM revenue, which is the company's own guided trajectory plus the expiry of §25D on 2025-12-31.
• No distant-year haircut is stacked on top. The base multiple sits above the p25 anchor, so the NTRA double-counting failure is not repeated.

A target below spot driven by falling near-term estimates is a valuation output. A target below spot driven by a multiple haircut would be a house view about the market. This is the first.

3.5 Named events inside twelve months that would move the target

Each appears in ENPH_Catalyst_Calendar.md.

1. IQ SST full-system demonstration — company says "on track for a full-system demonstration later this year" (2026). A first named data-centre customer or order would justify the p90 multiple and a $38–41 target. This is the only path to a target above spot.
2. First quarter with no safe-harbour revenue. Q3-2026 still carries ~$75m. The first clean quarter reveals the underlying US run-rate. If it prints near the Q2-2026 ex-SH level of ~$207.6m, the p25 multiple and the $25–26 target become the base.
3. Further Treasury/IRS FEOC guidance on §45X. The 10-K discloses that LFP battery cells are "supplied exclusively by two vendors located in China." An adverse material-assistance determination removes the credit that is currently 221% of operating income.
4. §48E placed-in-service cliff at 2027-12-31 begins to bind on the $878.6m Physical Work Test pipeline during the twelve-month window.

3.6 External target check

No Street price target is available without a paid feed and none is asserted. The gap cannot be reported; this is a stated gap, not an estimate.

---

4. Peer spread — INDETERMINATE, declared

The only comparator with the identical accounting structure (§45X recognised as a reduction of cost of sales, credit larger than operating income) is First Solar [FSLR], and its end-market is utility-scale modules sold to developers, not residential microinverters sold through distributors. FSLR's TTM 45X was 100.2% of operating income against Enphase's 220.5% — different magnitudes of the same disease, but not a spread that can be traded. No comparator with an equivalent margin structure and end-market exists. Declared UNIDENTIFIED rather than substituted with a sector median.

---

5. Downside Criteria — the permanent-loss case, with a named cause

Type: MEASURED. Logged; blocks nothing.

Named cause: safe-harbour exhaustion revealing the underlying US run-rate, into a §45X step-down.

Not a volatility figure. The mechanism is specific: Q2-2026 revenue excluding safe-harbour shipments was $207.6m, an $830m annual run-rate. §25D is gone. The §48E begin-construction window closed 2026-07-04, so no new safe-harbour agreements of that type can be written. The existing $1.08bn of executed agreements converts into shipments over roughly 2026–2027 and then stops.

Going concern: NO. This must be stated explicitly and it is not a going-concern case. Cash and marketable securities are $937.7m against a single $575.0m maturity on 2028-03-01, and the business generated $143.2m of operating cash in the first half of 2026. Enphase can pay its debt from the balance sheet — it has already demonstrated exactly that, settling $632.5m of 2026 converts in cash on 2026-03-02 without issuing a share. The risk here is permanent impairment of equity value through a re-rating onto product economics, not insolvency. That distinction matters for sizing: a name that cannot go to zero can be sized on volatility.

Dilution: not a factor. Unlike the other two names in this cluster, Enphase has not funded losses with equity. Shares outstanding: 132.47m (2025-02-03) → 131.10m (2026-02-05) → 132.14m (2026-07-20) — flat to slightly down over eighteen months, with $130.0m of buybacks in the first half of 2025. The 2028 converts are struck at $284.87 and are 8.1x away from being equity. The 2028 Warrants sold in 2021 strike at $397.91. Per-share revenue and per-share economics track the company's, one for one. This is the sharpest distinction within the cluster and it is in Enphase's favour.

---

6. Model integrity notes

• net income ÷ shares ≈ filed EPS: Q2-2026 $36,079k ÷ 132.145m = $0.273 vs filed diluted $0.27. ✓
• TTM revenue reconciles: FY2025 $1,472.985m − 1H-2025 $719.237m + 1H-2026 $574.754m + Q3-2025 $410.427m … computed as Q3-25 410.427 + Q4-25 343.321 + Q1-26 282.900 + Q2-26 291.854 = 1,328.502, where Q4-2025 is FY2025 less the three filed quarters. ✓
• Balance sheet ties: total liabilities and equity $2,919,259k at 2026-06-30 as filed; equity $1,181,836k; implied total liabilities $1,737,423k, which reconciles once the $572,836k non-current debt caption (shown separately in the debt note, not in the summarised long-term liabilities block) is included. ✓
• AMPTC figures are the company's disclosed amounts, not derived: $238.7m FY2025, $157.5m FY2024, 16.2pp and 11.8pp of gross margin respectively. Q2-2026 $64,922k net / $71,133k generated. ✓
• FY2023 AMPTC is not disclosed in the FY2025 10-K and is left blank. No estimate substituted.