Phase Space AI

Valuation

Evolv Technologies Holdings [EVLV]

EVLV — Valuation

Two horizons, two instruments. Both are required; reporting only one is a defect.

Spot $5.545 · 2026-07-29 · shares 179,458,233 · market cap $995.1m · net cash $32.4m · EV $962.7m TTM revenue $160.226m · EV/Sales 6.01x

---

1. Implied-path test — the Valuation Criteria (5-year horizon)

Instrument: assets/reverse_dcf.py. Terminal value is ~100% of EV because current EBIT is negative, so the reverse DCF is mandatory as the primary long-horizon output; no forward DCF is presented as a verdict.

1.1 The solve

The market requires a revenue CAGR of 24.5%.

1.2 The margin — demonstrated − required

This is the number the strategy ranks on. It depends entirely on which "demonstrated" is used, and that is the finding of this memo.

The screen's +15.5pp margin rests on the single measure that the recognition mix shift inflates. On every recurring measure the margin is negative.

1.3 Exit multiple — UNIDENTIFIED

Per criteria.md: "An exit multiple may only be drawn from a comparator set whose growth brackets the subject's growth at the exit year. If no such comparator exists, the multiple is UNIDENTIFIED and must be declared so."

No such comparator set exists for EVLV and none was fabricated. AI weapons-detection screening has no public pure-play peer of comparable size and growth. The 18.0x above is a stated assumption, and the required CAGR is therefore reported as a surface.

Sensitivity over the exit multiple — the highest-variance parameter, as required:

Cells are the required 5-year revenue CAGR (%), WACC 10%.

Sensitivity over WACC (18% terminal margin, 18x exit): 9% → 23.3% · 10% → 24.5% · 11% → 25.6% · 12% → 26.7%.

1.4 The break-even, stated explicitly

For today's price to require only the ~21% the recurring business demonstrates, the analyst must assume one of:

• 21x exit EV/EBIT at an 18% terminal margin (required 20.7%), or
• 22% terminal EBIT margin at 18x (required 19.6%), or
• WACC below 7%, which is not defensible for a name at 52% realised volatility with unremediated material weaknesses.

None is impossible; none is evidenced. The price is not absurd — it is unsupported.

1.5 Implied compression from today's multiple — stated as a number

The price does not require multiple expansion. It requires a 46% de-rating and 24.5% compounding simultaneously — an internally coherent maturation path. The problem is the growth rate, not the multiple.

1.6 Result

Valuation Criteria: FAIL.

The price requires materially more than the recurring business has demonstrated (24.5% vs ~21%), and the argument for bridging the gap — that upfront hardware conversion is permanent incremental revenue — is contradicted by the issuer's own guidance, which steps FY2026 reported growth down from +45% in Q1 to +20–23% for the year, i.e. down to the ARR rate.

PASS WITH ARGUMENT is reachable on a 21x+ terminal multiple plus a five-year 20%+ ARR runway. That is a legitimate view and it should be recorded as such. It is not the view the screen scored.

---

2. Twelve-month target

Built per valuation.md: near-term estimates → named product-cycle events → the name's own multiple history with the percentile stated. Not a DCF; not a peer median.

2.1 Near-term revenue base

No consensus estimate was obtained (Alpha Vantage quota). The base is the company's own guidance, which is the correct near-term anchor and is at least as informative.

2.2 Named product-cycle events inside 12 months

Each appears in EVLV_Catalyst_Calendar.md with a date.

• Evolv eXpedite X-ray attach into the installed base — the ramp is visible in disclosure frequency (0 → 25 mentions across nine filings) and is management's named Q1 2026 growth driver.
• Purchase-subscription mix at ~55% of new 2026 unit deployments (guided) — mechanically supports reported revenue and mechanically suppresses gross margin.
• Columbia Tech licence run-off — $8.1m of 89%-margin FY2025 revenue; Q1 2026 licence revenue already −69% YoY. A known, quantified drag through 2026.
• Plexus contract-manufacturing onboarding (agreement signed 2025-11-05) — management expects reduced product cost and improved gross margin "over time"; the product line is at −12% to +12% gross margin, so this is the single largest identified margin lever.
• Class-action final fairness hearing, 2026-09-24 — removes the largest quantified private-litigation overhang.

2.3 Multiple anchor — the name's own history, with the percentile

Point-in-time P/S (market cap ÷ TTM revenue), built with no look-ahead: on each trading day the series uses only the TTM revenue and diluted share count that had actually been filed by that date.

The post-restatement window is the only usable one, and it is 15 months long. Every multiple observation before 2025-04-28 was computed on revenue the company subsequently withdrew. valuation.md warns that a history "too short or spanning a regime change" should be declared UNIDENTIFIED — the pre-restatement history is exactly that, and is discarded. The 15-month post-restatement window is used with that limitation stated.

EVLV trades at the 3rd percentile of its own post-restatement range. The market has already de-rated the name substantially while revenue accelerated — consistent with the market pricing the recognition-quality issue this memo quantifies.

2.4 The target

NTM revenue $187m; shares ~185m in twelve months (Q1 2026 diluted 177.1m, growing ~1.5–2%/quarter on stock-based compensation).

12-month target: $7.00 — +26% to spot.

Range $6.20 – $7.90. Anchored on 6.94x P/S, the 25th percentile of the post-restatement window (current 3rd percentile), applied to NTM revenue of $187m.

Why the 25th percentile and not the median. Two reasons, both stated rather than applied silently: (i) the mix shift is degrading gross margin, and a lower-margin revenue mix earns a lower sales multiple — so partial rather than full mean reversion is the coherent assumption; (ii) the open SEC enforcement matter and five unremediated material weaknesses justify a discount to the window's own centre. The base multiple sits above the current multiple and above the window's p10, so no haircut has been stacked on an already mean-reverted anchor — the NTRA defect is not repeated.

Expect targets above spot to be common (item B16). This one is +26%. That is normal and it is not a house view about the market; it is a name trading at the 3rd percentile of its own post-restatement range with guided +20–23% revenue growth.

2.5 External target check

No professional's target for EVLV is on file in this project, so the sanity band is unavailable and no gap is reported. This is a missing check, stated as missing.

---

3. The two outputs side by side — and why they disagree

They disagree, and that is not a contradiction — it is the point of running both. A name can be cheap against its own recent trading range over twelve months and require more than it can deliver over five years. The first is a mean-reversion statement about sentiment; the second is a statement about terminal value.

The book should read this as: a tradeable twelve-month re-rating candidate that does not clear the ownership test. Which of those two facts governs is a strategy question, not a memo question — and the memo does not answer it.

---

4. Reconciliation to the screen

The screen's arithmetic is right and its inputs are right. The demonstrated CAGR is measuring the wrong thing — reported revenue for a subscription business mid-way through a recognition-model change. XBRL contains no ARR tag and no RPO growth rate; the screen could not have seen this. That is the calibration finding, and it generalises to every subscription business changing its delivery model.