Phase Space AI

Valuation

Kiniksa Pharmaceuticals [KNSA]

Kiniksa Pharmaceuticals [KNSA] — Valuation

As of 2026-07-29 · spot $78.145 · framework v1.5.1 · references/valuation.md

Two outputs are required and both are produced. Neither replaces the other.

---

1. Verified inputs

Filing recency asserted: the latest 10-Q was filed 2026-07-28, one day before this memo. Nothing here is stale.

---

2. The implied-path test — the Valuation Criteria

Instrument: assets/reverse_dcf.py, solving for the revenue CAGR today's price requires. Parameters held fixed and named: horizon 5 years; WACC 10%; terminal EBIT margin (stated per case); exit multiple on EBIT (stated per case). Terminal value is essentially 100% of the modelled EV, so the reverse DCF is the primary long-horizon output and no forward DCF is presented as a competing answer.

2.1 Anchoring the exit multiple — growth-matched

Comparator set: the 4,018 status: OK names in reports/scan_final, filtered to market cap ≥ $300m and 0 < EV/EBIT < 200, with revenue growth bracketing the subject's growth at the exit year.

The set has real growth dispersion, brackets the subject at every plausible exit-year growth rate, and is large. The multiple is IDENTIFIED, not defaulted. The four anchors span 21.4x–24.2x.

Base exit multiple: 22.5x EV/EBIT. It sits inside the anchor range and above the lowest stated anchor (21.4x), so the NTRA defect — a base below every anchor, with the haircut applied silently — is not repeated. Sensitivity is run over 14.0x–30.0x.

2.2 Terminal margin — three bases, all stated

1. Own TTM operating margin: 11.9%. What the business earns today.
2. Framework basis max(own, industry median): 16.2%. The basis the Tier-1 screen used. Reproduced so the comparison is like-for-like.
3. Structural base: 25.0%. Derived, not assumed — see KNSA_Research.md §3. The Regeneron 50/50 profit split caps Kiniksa's pre-opex contribution at 0.5 × (1 − COGS%) = 45.2% of revenue. At SG&A 16% and R&D 10% of revenue, EBIT margin is 24.7%; the ceiling at SG&A 13% / R&D 8% is 28.6%. 25.0% is the mid-case and it is the number this memo treats as base.

An important asymmetry: the structural ceiling of ~28.6% means the terminal margin cannot be rescued by optimism, and the framework's 16.2% is conservative rather than generous. Both directions are bounded, which is unusual and makes this parameter less dangerous here than on most names.

2.3 Required 5-year revenue CAGR — sensitivity over the EXIT MULTIPLE

(Never over scenario probabilities. EV = $5,571.3m, revenue₀ = $840.8m, WACC = 10%, years = 5.)

2.4 The required outputs

1 — The required parameter. At the base case (terminal margin 25.0%, exit multiple 22.5x EV/EBIT, WACC 10%, 5 years), today's price requires a 13.7% five-year revenue CAGR, implying FY2031 revenue of $1,598m and FY2031 EBIT of $399m.

At the screen's own parameter choices (terminal margin 16.2%, exit 22.5x) the requirement is 24.0%, implying FY2031 revenue of $2,465m. The screen published 28.9%; the 4.9pp difference is entirely attributable to its stale TTM revenue, overstated share count and understated net cash.

2 — The margin, demonstrated − required:

This memo ranks on +34.9pp — the conservative combination of the slowest demonstrated window with the screen's own (lower) terminal margin. The screen ranked KNSA at +16.6pp. The corrected margin is more than twice as wide.

3 — Exit multiple and implied compression.

The 60% EBIT-multiple compression is doing most of the work and it is not an aggressive assumption — it is the arithmetic of a business whose margin triples while its multiple normalises to a large growth-matched comparator median.

4 — Sensitivity over the exit multiple. Table §2.3. The demonstrated rate of 58.9% clears the required rate in every one of the twenty cells — including the harshest (own 11.9% terminal margin, 14.0x exit, requiring 45.0%). There is no plausible corner of the parameter space in which today's price requires more than the business has demonstrated.

2.5 Verdict

Valuation Criteria: PASS. Not "PASS WITH ARGUMENT" — no argument beyond demonstrated performance is needed. The implied path sits well below what the business has already delivered, on every parameterisation tested.

---

3. Does the penetration math support the required path? — the direct answer

The brief asked whether a demanding required CAGR sustained for five years off a single drug is supported. The question is answered arithmetically, in patients.

Method. Revenue = patients on therapy × revenue per patient-year. Both are observable (KNSA_Research.md §1). Revenue per patient-year is $335.4k on 1H2026 annualised, and is modelled forward at +3%/yr — price only, because the persistency lever (therapy duration 20 → 36 months) is exhausted by management's own statement. That gives $388.8k in 2031.

Against what has actually been delivered: penetration went 5% → 9% → 13% → 18% → 21%, i.e. +4pp (2023), +4pp (2024), +5pp (2025), +6pp annualised (1H26) — accelerating.

Running the observed pace forward:

The plain statement

Yes — the penetration math supports the required path, with room. The base case needs penetration to reach 29.4% by 2031, at +1.7pp/yr. Kiniksa has not delivered a year below +4pp since 2023 and delivered +6pp annualised in the last two quarters. Even a halving of the current pace to +3pp/yr produces an 18.4% CAGR, which comfortably clears the 13.7% requirement.

At the screen's own 24.0% requirement it is still supported but no longer comfortable: penetration must reach 45.3%, at +4.9pp/yr — a pace the company has hit only in its two most recent years. That requires the launch curve not to roll over for five straight years, which is a real assumption.

The screen's published 28.9% is the one figure the penetration math does not support. It needs +6.8pp/yr sustained for five years — faster than Kiniksa's best year ever. But 28.9% is an artifact of stale inputs, not a live requirement. On corrected inputs the number is 24.0% at the same parameters.

Three qualifications, stated rather than buried:

1. The 14,000 is not the ceiling. It is the promoted core segment. ARCALYST's label covers recurrent pericarditis generally; there are ~26,000 additional first-recurrence patients inside the label and ~40,000 total prevalent treated patients. Penetration above 50% of the core segment is not required if the promotion broadens — and it is not a step Kiniksa has signalled.
2. The pool refreshes. Kiniksa's own investor materials state ~50% annual turnover with ~7,000 patients entering the multiple-recurrence pool each year, and ~45% of patients who discontinue restart within ~8 weeks. The denominator does not deplete, which is what makes >40% point-in-time penetration reachable at all.
3. Penetration curves decelerate. They always do. Every figure above assumes linear extrapolation of a variable that is empirically S-shaped. The base case survives a halving of the pace; the 24.0% case does not survive much deceleration at all. That is the real risk in the name, and it is a rate-of-change risk, not a level risk.

---

4. The 12-month target

Built per valuation.md: near-term estimates, named product-cycle events, and the name's own multiple history with the percentile stated.

4.1 NTM revenue (August 2026 → July 2027)

Sanity: FY2026 consensus revenue is $940.1m (stockanalysis.com, 2026-07-29) — 5.0% below the guidance midpoint used here, and that consensus predates the raise. Q2 2026 consensus revenue was $230.0m against $243.6m actual, a 5.9% beat. Estimates are moving up, not down.

4.2 The multiple anchor — the name's OWN history, percentile stated

Daily EV/Sales on as-known TTM revenue (lagged to filing date), 857 sessions from 2023-02-28, share count and net cash held at verified current values so the series measures multiple movement:

KNSA has never been more expensive on trailing revenue than it is today. Any target that assumes further multiple expansion is unsupported by the name's own history, and this memo does not make that assumption.

Two facts that shape the anchor choice:

1. The re-rating started before the print. Over the ten sessions to 2026-07-27 the name traded in a 5.62x–5.93x band, peaking at 6.26x on 2026-07-07. The print took it from 5.88x to 6.74x. The move is not a single-day spike off a low base.
2. The print delivered new information, not just momentum: a $50m guidance raise (+5.2%) and KPL-387 Phase 2 data with Phase 3 initiation. Under references/regime-change-test.md this is Tier 3 (structural) evidence — real, and sufficient to argue the multiple should not revert to a pre-profitability median, but not sufficient to carry "the multiple holds at an all-time high for twelve months." There is no contractual or physical floor here. Mean reversion remains the default.

Base anchor: 5.88x — the multiple the market paid on 2026-07-27, the last session before the print (99th percentile of the name's full history). This assumes the re-rating persists but the gap does not, on a revenue base 38% larger.

4.3 The target

12-MONTH TARGET: $94 — +20.3% to spot, ABOVE spot.

Sanity band against external targets. Street (stockanalysis.com, 2026-07-29, post-print): 8 analysts, average $98.00, median $98.50, low $87, high $108. The house target of $94 sits 4.6% below the Street median and inside the Street range. There is no material divergence to explain — this is the intended use of an outside target, as a check on the output rather than a calibration input.

(A pre-print average of $64.00 circulates in secondary sources. It is stale — it predates both the guidance raise and the KPL-387 data, and Goldman Sachs alone moved from $60 to $75 on the print. It is noted and discarded.)

Named product-cycle events inside the 12-month window (all in KNSA_Catalyst_Calendar.md): Q3 2026 results and a probable third 2026 guidance raise (late October 2026); FY2026 results and initial FY2027 guidance (late February 2027); the January 2027 corporate update; KPL-1161 Phase 1 initiation (company-guided "by the end of 2026"); PASTORALE enrolment progress.

---

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

Type: MEASURED. Logged and scored; it blocks nothing.

Primary bear case — the penetration curve rolls over before KPL-387 arrives

Named cause. ARCALYST penetration plateaus in the low-to-mid 30s of the 14,000 pool as the readily identified, high-severity multiple-recurrence patients are exhausted and the remaining pool proves harder to diagnose, refer and convert. Simultaneously, the persistency lever is already spent — management has stated duration of therapy now equals the median duration of disease — so revenue per patient grows at price only. PASTORALE reads out late or misses, removing the 2028/2029 migration path.

Mechanism to loss. Revenue plateaus around $1.6–1.8bn in 2029–2030. EBIT reaches ~$400–450m at the structural ~25% ceiling. A single-asset, US-only, 50%-profit-shared franchise with no growth and an orphan exclusivity that lapsed in March 2028 does not hold 22.5x EBIT — it holds ~12–15x. At 14x on $420m EBIT, EV = $5.9bn, equity ≈ $6.6bn with accumulated cash, or roughly $84/share — above today's price, which is the important point.

Quantified drawdown. The permanent-loss case is not a wipeout; it is a de-rating. The path to $84 in 2030 runs through the growth multiple unwinding first, and a compression from 6.63x EV/Sales to ~4.0x on FY2027 revenue of $1,300m implies $74 — a −5% outcome — while a compression to the full-history median of 3.46x implies $59, a −25% outcome. Realistic peak-to-trough drawdown in this scenario: −30% to −40%, driven by multiple compression rather than revenue loss.

Probability: 25%.

Secondary named cause — a cheap oral competitor changes the payer conversation

Cardiol Therapeutics' MAVERIC Phase 3 (NCT06708299) of oral CardiolRx in recurrent pericarditis, FDA orphan designation granted, enrolment "approaching completion" per the company's 2026 6-K filings. If it succeeds and prices as an oral, payers gain a step-edit ahead of a $335k/year weekly injectable. This attacks net price and access, which is where the margin lives, not just share. No readout date is announced and none is invented here. Probability of a materially negative outcome for KNSA inside five years: 15%, overlapping with the primary case.

Tertiary — the March 2028 skinny-label pathway

Orphan exclusivity for pericarditis lapses 2028-03-18. A rilonacept biosimilar approved for CAPS/DIRA only and used off-label in pericarditis would carve around the 2038 method-of-use estate. No such programme was identified in this research, and the economics argue against one (complex Fc-fusion trap, ~$1bn branded market, BLA-scale development, a decade of patent litigation exposure). Probability: 10%, and it sits at the back end of the window.

Going concern

Explicitly not a going-concern case. $525.9m of cash and short-term investments, zero debt, GAAP profitable for six consecutive quarters, 1H26 operating cash flow of $97.4m, and management guiding that "its current operating plan [will] remain cash flow positive on an annual basis." Kiniksa can fund PASTORALE and the KPL-1161 Phase 1 from operations.

---

6. Consensus Criteria

INDETERMINATE. The Alpha Vantage daily quota is exhausted (verified — the API returned the rate-limit message on 2026-07-29), so no estimate-revision history was obtainable. Per criteria.md this blocks nothing.

What is documented from secondary sources, with attribution:

House versus Street on the metric the mechanism moves: the house FY2027 revenue estimate of $1,300m is a house number with no Street figure to compare against. That gap is stated, not filled.