Relay Therapeutics, Inc. (RLAY) valuation

0 · TTM reconstruction (anchor: Q12026, 2026-03-31)

The latest filing is a 10-Q, so "base year" revenue / EBIT / OCF / CapEx are reconstructed as trailing-twelve-month values. Per-quarter facts (typical for income-statement items) get summed across four quarters; YTD-cumulative facts (typical for cash-flow items) use prior FY + YTD_now − YTD_{prior year same quarter}.

Revenue

Sum of the four most recent per-quarter values

• Q1 FY26 (2026-03-31): $3.0M
• Q4 FY25 (2025-12-31): $7.0M
• Q3 FY25 (2025-09-30): $0
• Q2 FY25 (2025-06-30): $677.0K
• = $10.7M

EBIT

Sum of the four most recent per-quarter values

• Q1 FY26 (2026-03-31): -$78.6M
• Q4 FY25 (2025-12-31): -$60.6M
• Q3 FY25 (2025-09-30): -$80.4M
• Q2 FY25 (2025-06-30): -$76.8M
• = -$296.5M

OCF

Prior FY + current-quarter YTD − same-quarter-prior-year YTD

• FY FY25 (2025-12-31): +-$235.5M
• Q1 FY26 (2026-03-31) YTD: +-$51.1M
• Q1 FY25 (2025-03-31) YTD: −-$73.2M
• = -$213.4M

Stock-based compensation

Prior FY + current-quarter YTD − same-quarter-prior-year YTD

• FY FY25 (2025-12-31): +$62.4M
• Q1 FY26 (2026-03-31) YTD: +$10.5M
• Q1 FY25 (2025-03-31) YTD: −$19.3M
• = $53.6M

Prior-year TTM revenue (growth-calc baseline)

Sum of the four most recent per-quarter values

• Q1 FY25 (2025-03-31): $7.7M
• Q4 FY24 (2024-12-31): $0
• Q3 FY24 (2024-09-30): $0
• Q2 FY24 (2024-06-30): $0
• = $7.7M

1 · Enterprise-value target (what the DCF must match)

Diluted shares = point-in-time basic shares outstanding — filer's diluted-vs-basic weighted-average pair not tagged in the company-facts API. No TSM dilution scale applied. See the diluted-shares methodology.

Market cap     = price × diluted shares
               = $9.95 × 189.2M
               = $1.88B

EV target      = market cap + total debt − cash & equivalents
               = $1.88B + $31.6M − $204.6M
               = $1.71B
            

2 · Starting NOPAT (base year 0)

Tax rate source: 21% US statutory default (fewer than 2 of the last 3 FYs had a positive effective rate — likely NOL utilisation or one-time items). See the tax-rate methodology for the precedence ladder.

GAAP EBIT          = -$296.5M   (-2776.6% of revenue)
× (1 − tax rate)  = × (1 − 21.0%) = × 0.7900
= NOPAT₀            = -$234.2M
            

3 · Invested capital & starting ROIC

Total debt above includes the operating-lease liability ($31.6M). The corresponding ROU asset sits on the balance sheet and flows through book equity via the accounting identity (E = Assets − Liabilities), so IC reflects the post-ASC-842 operating-capital base without an explicit add-back.

Invested capital = total debt + book equity − cash
                 = $31.6M + $642.2M − $204.6M
                 = $469.2M

Raw ROIC₀        = NOPAT₀ / Invested capital
                 = -$234.2M / $469.2M
                 = -49.9%
(no cap applied; raw value is within the 40.0% ceiling)
            

4 · WACC derivation

Cost of equity from CAPM , after-tax cost of debt from a synthetic credit rating built off interest coverage , weighted by market values of equity and debt. Inputs: Rf is FRED DGS10's 90-day mean (latest 2026-05-29); β is the 5-yr weekly regression vs VOO, floored at 1.00 for the cost-of-equity step (the empirical security market line is much flatter than CAPM predicts — Frazzini-Pedersen 2014, "Betting Against Beta" — so unfloored CAPM systematically under-estimates required return for low-β filers); ERP is the latest Damodaran US total ERP (2026-01-01). The synthetic rating is a Damodaran coverage-table heuristic, not an empirical S&P / Moody's letter. It maps EBIT-÷-interest into a letter-grade bucket and reads the spread out of a lookup table; an actual agency rating considers qualitative factors (governance, market position, jurisdiction, off-balance-sheet exposure) the coverage ratio can't capture. For most filers the gap is small; for capital-light, high-coverage names — software, IT services, consumer brands — this method tends to print AAA/AA where actual agency ratings sit at A/BBB+, and the resulting cost of debt is a few tens of bps light. WACC is more sensitive to β and weights than to credit, so the headline barely moves; flagged here so readers don't read the rating as an empirical agency assessment.

Cost of equity        = Rf + β × ERP
                      = 4.3% + 2.06 × 4.5%
                      = 13.5%

Cost of debt (pretax) = Rf + credit spread
                      = 4.3% + 1.5%   (synthetic A− (low leverage, coverage uncomputable))
                      = 5.8%
× (1 − tax rate)      = × (1 − 21.0%)
= Cost of debt (a/t)  = 4.6%

Weights               E/V = 98.4%, D/V = 1.6%

WACC (raw)            = E/V × cost_e + D/V × cost_d_after_tax
                      = 98.4% × 13.5% + 1.6% × 4.6%
                      = 13.4%
                

5 · Growth path construction

Source       = analyst consensus (absolute forecast, TTM-anchored): Y1 = 11.3%, Y2 = 60.0%

Detailed derivation:
  Current-FY analyst avg revenue forecast = $11.9M   (FY-over-FY vs FY2025 actual = -22.6%)
  Next-FY analyst avg revenue forecast    = $19.6M   (FY-over-FY vs current-FY forecast = 65.2%)
  Base revenue (TTM) = $10.7M
  Latest completed FY revenue (FY2025) = $15.4M   (denominator for the FY-over-FY check above)

  Y1 = current-FY forecast / TTM base revenue − 1
     = $11.9M / $10.7M − 1
     = 11.3%

  Y2 = next-FY forecast / current-FY forecast − 1
     = $19.6M / $11.9M − 1
     = 60.0%

  Note: the absolute analyst forecast is re-anchored
  against TTM (rather than the FY-over-FY consensus
  rate above) because the TTM base spans the current
  FY only partway. Re-anchoring keeps the growth rate
  consistent with the projection's starting revenue
  level.
Clamp        = [2.5%, 60%] (no sub-terminal or 60%+ starts)
Plateau rate = 60.0% (Y2 — held from year 2 through end of plateau)
Tier         = 7 years (rule: plateau rate < 15% → 3y, < 25% → 5y, else 7y)
Cap trim     = 5 years (compound cap: Y1 × plateau_rate^(n−1) ≤ 10× base)
Plateau      = 5 years
Fade         = linear from effective Y2 to terminal 2.5% across the remaining 5 years

Effective Y1 growth after solver bumps = 11.3%
Effective Y2 growth after solver bumps = 60.0%
Growth by year:
  Y1 = 11.3%
  Y2 = 60.0%
  Y3 = 60.0%
  Y4 = 60.0%
  Y5 = 60.0%
  Y6 = 48.5%
  Y7 = 37.0%
  Y8 = 25.5%
  Y9 = 14.0%
  Y10 = 2.5%
                

6 · Margin path construction

Starting margin (Y0) = -2776.6%   (source: TTM EBIT margin (GAAP))
Target margin (Y10)  = 80.0%   (solver output, widened band)
Year-t margin        = starting + (target − starting) × (t / 10)
Margin by year:
  Y1 = -2490.9%
  Y2 = -2205.3%
  Y3 = -1919.6%
  Y4 = -1634.0%
  Y5 = -1348.3%
  Y6 = -1062.6%
  Y7 = -777.0%
  Y8 = -491.3%
  Y9 = -205.7%
  Y10 = 80.0%

Filer-history margin distribution: insufficient history (1 FY available, 8 required); path-stretch flag suppressed.
            

7 · ROIC path construction

The capex heuristic compares latest-period CapEx (—) against the Normalized CapEx (3-yr mean) of $2.2M — mean of the last three annual CapEx values. When the latest is above 1.4× that mean and CapEx is at least 5% of revenue, we treat the filer as capital-intensive and mid-investment, hold ROIC flat for a 5-year harvest phase, and only then fade to terminal ROIC. The 3-yr mean does not feed the DCF directly — it only gates this flag.

Capex-heuristic inactive (latest CapEx vs the 3-yr mean of $2.2M — below the 1.4× / 5%-of-revenue gates).
Fade from Y1: ROIC_t = ROIC₀ + (ROIC_terminal − ROIC₀) × (t / 10)
ROIC₀ = 0.0%; ROIC_terminal = 13.4%

ROIC by year:
  Y1 = 1.3%
  Y2 = 2.7%
  Y3 = 4.0%
  Y4 = 5.4%
  Y5 = 6.7%
  Y6 = 8.0%
  Y7 = 9.4%
  Y8 = 10.7%
  Y9 = 12.0%
  Y10 = 13.4%
            

7a · Reinvestment formula

The Damodaran growth-firm closure Reinvestment_t = ΔRevenue_t ÷ salesToCapital collapses to noise whenever EITHER the 3-year revenue change OR the 3-year net reinvestment is small relative to the revenue base. The model picks the most-honest formula via a 3-tier ladder: tier-1 standard sales-to-capital fires only when both signals carry magnitude AND the resulting ratio sits inside the typical 0.5–8 band; tier-2 substitutes a TTM (CapEx − D&A) ÷ Revenue ratio applied to each projection year's revenue (not its delta); tier-3 falls through to the ΔNOPAT ÷ ROIC closure when neither signal is usable.

Inputs:
  ΔRev_3y                          = $9.3M
  netReinvest_3y (CapEx − D&A)     = -$7.7M
  avgRevenue_3y                    = $6.0M
  TTM CapEx − D&A                  = n/a
  TTM Revenue                      = $10.7M

Tier-1 gates (all three must clear):
  |ΔRev_3y| / avgRev_3y > 1.0%     154.19%        ✓ pass
  netReinv_3y / avgRev_3y > 1.5%   -128.31%       ✗ fail
  0.5 ≤ s2c candidate ≤ 8          1.418          ✓ pass

Tier-3 fires (no usable s2c or CapEx-of-revenue input — gate failed: denominator). Falling through to the ΔNOPAT ÷ ROIC closure.
  Reinvestment_t = max(0, ΔNOPAT_t ÷ ROIC_t)   when ΔNOPAT_t > 0
            

8 · Solver iterations

Each row is one configuration in the solver ladder; the "Solved margin" column is the result of bisecting the EBIT margin (up to 80 inner iterations) to match the EV target for that row's plateau / Y1 bump / phase. The solver sweeps Y1 growth bumps 0pp → +20pp across the plateau ladder inside the normal margin bracket, then — if nothing reconciles — repeats the same sweep in a widened margin band ([-10%, 80%]). The first feasible row is the one the page uses. If no combination reconciles, the page shows the row whose PV sits closest to the target EV so both levers are balanced.

9 · Terminal value derivation

NOPAT_{N+1}         = NOPAT_{10} × (1 + g_terminal)
                    = $146.8M × (1 + 2.5%)
                    = $150.5M

ΔNOPAT              = NOPAT_{N+1} − NOPAT_{10}
                    = $3.7M
Reinvestment_{N+1}  = ΔNOPAT / ROIC_terminal
                    = $3.7M / 13.4%
                    = $27.4M

FCF_{N+1}           = NOPAT_{N+1} − Reinvestment_{N+1}
                    = $150.5M − $27.4M
                    = $123.1M

Terminal value (TV) = FCF_{N+1} / (WACC − g_terminal)
                    = $123.1M / (13.4% − 2.5%)
                    = $1.13B

PV(TV)              = TV / (1 + WACC)^(10 − 0.5)
                    = $1.13B / 3.296
                    = $343.2M

10 · Reconciliation check (DCF PV vs. the market)

This isn't a fair value — it's the inverse check. The solver built the scenario so that DCF PV reproduces the current enterprise value; if the normal bracket worked the delta below is ~0 by construction. A non-zero delta only appears when the solver fell through to the widened margin band.

Σ PV(FCF_1..10) = -$3.05B
+ PV(TV)          = $343.2M
= Enterprise value = -$2.71B   (widened solve — may differ from EV target)
− Total debt      = $31.6M
+ Cash            = $204.6M
= Equity value    = -$2.54B
÷ Diluted shares  = 189.2M
= DCF PV / share  = -$13.40

Market price      = $9.95
Reconciliation Δ  = −234.7%   (widened band — residual gap the scenario could not close)