Phase Space AI

Valuation

TransMedics Group [TMDX]

TransMedics Group [TMDX] — Valuation

Two outputs, two horizons. Neither replaces the other. Spot $76.35 (2026-07-28 close) · shares 34,560,911 · analysis date 2026-07-29

---

0. The corrected capital structure — everything below depends on it

The Tier-1 screen recorded +$298.989m of net cash. The filed balance sheet at 31 March 2026 shows net debt of $395.3m. The error is $694m and it drives the screen's headline result.

Ex the finance lease, EV is $2,690m and EV/Sales 4.23x — still not the screen's 3.68x. The finance lease is carried as debt in the base case: the purchase option is one the company itself judges "reasonably certain to exercise," which is precisely why it is capitalised at all. The ex-lease case is run as a sensitivity in §2.3.

---

1. The implied-path test — the Valuation Criteria

Instrument: assets/reverse_dcf.py, solving for the unknown rather than asserting a path. Terminal value is 100% of EV by construction in this instrument, so it is mandatory as the primary long-horizon output and a forward DCF is not run as the verdict.

1.1 Base parameterisation and result

THE MARKET REQUIRES: revenue CAGR of 11.7%

1.2 The margin — demonstrated − required

Comparator used: 21.2% — the slowest rate the business has actually demonstrated. That choice is deliberate and conservative: growth has decelerated monotonically across five consecutive quarters (48 → 38 → 32 → 32 → 21%), so the older, faster windows describe a business that no longer exists.

Result: PASS. The implied path sits below what the business has demonstrated, on the slowest available reading of "demonstrated."

Against the screen: the screen claimed +22.9pp (28.9% − 6.0%). The corrected margin is +9.5pp — 13.4pp thinner. Both terms were wrong: the required rate was understated by the $694m net-cash error, and the demonstrated rate was computed over the wrong window.

1.3 Implied multiple compression, stated as a number

The base case therefore already assumes the multiple contracts by more than a quarter over five years. It is not a re-rating story at the long horizon.

---

2. Sensitivity — over the exit multiple, never over scenario probabilities

2.1 Exit multiple (terminal margin held at 19.0%)

Flip point: ≈15.5x EV/EBIT. Below that the Valuation Criteria fails. 15.5x sits below every named anchor in the comparator set — the lowest is RMD at 16.5x — so failing on the exit multiple alone requires TransMedics to be re-rated beneath a set of slower-growing, higher-margin, unlevered device peers. Possible, but it has to be argued, not assumed.

2.2 Terminal operating margin (exit multiple held at 23.3x) — the parameter that actually decides

This is where the vertical-integration trap lives.

Flip point: ≈12.6% terminal operating margin. Below it the price requires more than the business has demonstrated and the argument becomes narrative.

Why this range is the whole question. From TMDX_Research.md §2:

• The service (transport/logistics) segment is 38.3% of TTM revenue at a 28.5% gross margin; product is 61.7% at 78.1%.
• Holding both segment margins constant, the blended gross-margin ceiling is arithmetically pinned near 59.1%. To convert that into a 19.0% operating margin, opex must fall to 40.0% of revenue from a TTM 44.2% — and Q1-26 opex ran at 50.5%, up 8.1pp year-on-year.
• Both named growth initiatives (European NOP, the PAD Aviation investment) are logistics expansions. At a 45% service mix the blended gross-margin ceiling falls to ~55.7%, and the required opex ratio for a 19% operating margin falls to ~36.7%.

A screen that applies an industry-median device margin to this revenue line misprices it, exactly as the brief anticipated. The screen's own computed operating margin was 346% — nonsense — so the 19% is entirely an industry number, and the criterion's verdict turns on it.

2.3 Capital-structure sensitivity — what each correction is worth

The screen's 6.0% is not a modelling difference. It is the arithmetic consequence of a $694m balance-sheet error, and correcting that error alone accounts for 5.7pp of the 13.4pp margin overstatement.

---

3. The 12-month target

Built per references/valuation.md: near-term revenue, named product-cycle events, and the name's own multiple history with the percentile stated. Not a DCF, and not a peer median projected forward.

3.1 NTM revenue

The FY2027 figure is a house estimate, not consensus. The Alpha Vantage quota is exhausted, so the Consensus Criteria is INDETERMINATE and no NTM consensus was obtainable. +17% is a deceleration from the FY2026 guidance midpoint of +22.5%, consistent with five quarters of monotonic slowing. It is flagged wherever used.

3.2 The multiple — own history, with the regime break declared

TransMedics' own EV/Sales history, computed quarterly from filed balance sheets and quarter-end closes, 23 observations, Q3-2020 → Q1-2026:

The full history is DECLARED UNIDENTIFIED as a target anchor. It spans a genuine regime change and using its median would be indefensible:

• No owned logistics before August 2023. Service revenue was 27% of the total in FY2023 and 38% now; the gross-margin mix is a different business.
• No convertible debt before May 2023, no finance lease before January 2026 — the EV bridge is structurally different.
• Growth in the anchor period was 50–150%. It is 21% now. A multiple observed at 150% growth carries no information about a 21% grower.

The anchor used is the post-regime sub-sample — every quarter since the company turned profitable and the service mix stabilised at ~38%:

Base multiple: 6.0x NTM EV/Sales — the 33rd percentile of the sub-sample, marginally below its most recent observation (6.02x) and below its median (6.44x), reflecting the deceleration and the funding overhang; and above its minimum, because valuation.md prohibits setting a base beneath every stated anchor without a separately argued reason, and no such reason exists here.

3.3 Target, and the convertible treatment

At any price above $94.00 the converts are in the money and become 4,893,834 shares ($459,996,000 ÷ 1,000 × 10.6388), while the $460m of debt extinguishes. The two cases are modelled consistently.

Share count used above $94: 34.561m + 4.894m converts + 0.7m of 12-month SBC issuance = 40.155m. Below $94: 34.561m + 0.7m = 35.261m, with the converts held as debt. No benefit is taken for the capped calls ($52.1m of premium purchased May 2023) because the cap price was not located in the filings read — conservative, and disclosed.

12-month target: $122 — +60% above spot

Range $39 – $162. Probability-weighted: $107 (+40%).

3.4 Sanity band against the external target

Explanation of the divergence, as required. Both sit far above spot because the stock trades at the 0th percentile of its own EV/Sales history. The 18% gap is attributable to two specific choices, not to a different view of the business:

1. The multiple. The house base of 6.0x sits below the post-regime median of 6.44x. Street's $148.5 implies roughly 7.4x NTM EV/Sales on the same revenue and share count — i.e. above the post-regime median, near the Q2-2025 high of 8.76x. That is a re-rating call; the house does not make it while operating margin is contracting.
2. The balance sheet. Reaching $148 requires the $395m of net debt to be either refinanced or extinguished by conversion. The house case gives credit for conversion above $94, but takes no credit for the Somerville purchase option being abandoned, which would release $374.6m of the obligation.

This is a check on the output, never a calibration target for the model. No parameter was moved to close the gap.

---

4. Criteria results carried to the manifest

---

5. What this valuation does not establish

1. The FY2027 revenue figure is a house estimate, not consensus. No NTM consensus was obtainable.
2. The terminal operating margin is the parameter that decides the criterion, and it is not observable. The result is PASS at 19% and 15%, and FAIL at 12%. The memo does not pretend to know which.
3. The exit multiple is corroborated but not measured. The named comparator set has n=10 and none of its members carries a 38%-of-revenue charter-aviation segment. There is no true comparable.
4. The capped calls are ignored, which understates the equity value above $94 by an unquantified amount.
5. The finance lease is treated as debt in the base case. If TransMedics abandons the Somerville purchase option, $343.8m of the liability reverses (at the cost of $23.9m of annual rent from 2028 and the writedown of a $334.5m ROU asset), and the required CAGR falls from 11.7% to 9.0%.
6. Q2 2026 results land on 4 August 2026, six days after this analysis. Every number here is provisional until then.