Twist Bioscience (TWST) — V2 Revaluation and Post-Mortem: Why the January SELL Was Wrong, Sep 2026

Published Sep 20, 2026 · psind707 · Damodaran-framework DCF · Replaces Jan 25, 2026 model (SELL, PWFV $9.98 at $42.61)

1. What changed, Jan 2026 to Sep 2026

Item	Jan 2026 (V1 inputs)	Sep 2026 (verified)
Price	$42.61	$167.28 (+293%)
Revenue base	FY25 $376.6M	FY26 guide $456.5M (raised 3x from $430M)
Latest quarter growth	Q4 FY25 +17%	Q3 FY26 +23% (14th consecutive sequential)
Gross margin	50.7%	52.8%; LT target above 60%; 70% incremental flow-through
AI-enabled discovery orders	$25M FY25 (disclosed; not modeled)	$50M FY26, $100M FY27 guided (CFO, Aug 3 call)
Adj EBITDA	FY25 -$46.9M	Q3 -$11.3M; breakeven Q4 FY26 reiterated
GAAP operating loss (after SBC)	FY25 -$136M (-36%)	9M FY26 -$115M (-35%); SBC $50M in 9M
Diluted shares	61.8M	approx 68M (62.4M basic + 3.59M sold at $96 in Aug + RSU/options)
Cash and investments	$232M	approx $494M ($167M Jun 30 + $327M net raise); $68M Atlas stake
EV / forward sales	5.2x	approx 23x FY26, 19x FY27
Street	avg PT $43	Leerink $160, UBS $144 (Neutral), BWS Sell $45

Decomposition of the 293% move: revenue run-rate +21%, forward growth expectation from 15% to 23%, cash roughly doubled via equity raise. Call it 1.3x on fundamentals. The remaining 3.4x is multiple: 5x sales to 23x sales on the "wet lab for AI drug discovery" narrative (Lilly TuneLab data-services agreement Sep 16; Anthropic mini-binder pilot, which management says is already inside FY27 order expectations).

2. Corrected valuation (V2)

Cost of capital. Bottom-up: risk-free 4.3% + unlevered life-science-tools beta 1.15 x ERP 4.6% = 9.6%. Zero debt, so cost of equity = WACC. (V1 used a raw 5-year regression beta of 2.24 and got 16.8%. Indefensible for a net-cash company; see post-mortem.) Rf and ERP are stated assumptions, not pulled live.

Model. Owner economics: GAAP operating margin after SBC (Kong/Damodaran convention). Starting margin -31% FY26, linear ramp over 7 years to terminal. Sales-to-capital 2.5. Terminal: g 3%, ROIC 15%, tax 21%. Equity bridge adds $494M cash plus $68M Atlas Data Storage stake, less $15M royalty liability. 68M diluted shares.

Scenario	Prob	Rev CAGR FY27-29	FY31 revenue	Terminal op margin	EV	Value / share
Bull — AI wet-lab thesis compounds; Lilly-type deals multiply; MRD probes scale	25%	32%	$1.66B	30%	$3.95B	$66
Base — above LRP; AI orders convert to revenue at 1-2 quarter lag	50%	22%	$1.15B	25%	$1.87B	$36
Bear — management LRP (mid-teens); ASP pressure in NGS/protein	25%	14%	$0.82B	18%	$0.67B	$18
Probability-weighted fair value (PWFV)						$39 (-77% vs $167.28)

Reverse DCF: what $167 requires

5-yr revenue CAGR (FY27-31)	FY31 revenue	Value / share at 28% terminal margin
25%	$1.39B	$54
35%	$2.05B	$87
45%	$2.93B	$141
approx 48%	approx $3.2B	$167

Management's own Investor Day long-range plan (May 2026): mid-teens growth, "organic revenue more than doubling by 2031", i.e. roughly $950M. The market is paying for about 3x the company's plan. Pricing signals cut the other way too: Goldman/Blair/Leerink/Guggenheim placed 3.59M shares at $96 six weeks ago and the stock is 74% above that print; six Form 144s filed in the past month.

Rating and positioning

AVOID at $167. Not a short. No dated negative catalyst; momentum, AI narrative, and float turning at 3M shares/day make a naked short a bet on timing, not value. Defined-risk expression only: long-dated puts (Jan 2028) sized at 1% of book, thesis = FY27 guide (Nov 2026 call) lands at 20-25% growth rather than the 40%+ the price needs. Bull-case DCF ($66) is the level at which the stock becomes a HOLD; base ($36) a BUY.

3. Post-mortem: why V1 was wrong

#	Error	What I did	Effect
1	WACC 16.8%	Plugged raw 5-yr regression beta (2.24) into CAPM for a zero-debt, $232M-cash company. Damodaran's rule is bottom-up industry beta precisely because regression betas on volatile small-caps are noise.	Replaying the V1 base case at 10% instead of 16.8% gives $20, not $9.60. Doubled the error on its own.
2	Double-counted cost of growth	Started from a -12% margin that already contains the R&D and SG&A spent to grow, then charged a separate sales-to-capital reinvestment on every new revenue dollar. For a platform business the losses are the reinvestment. Damodaran capitalizes R&D; I penalized the same dollar twice.	Negative FCFF for 4-5 years that never existed economically.
3	Terminal margin 20%	Anchored on peer median. Company's own economics (60%+ LT gross margin, 70% incremental flow-through) imply 25-30% operating margin at scale; Illumina peaked at 28-30%.	Terminal value understated by a third.
4	Ignored the order book	$25M of AI-enabled discovery orders were disclosed on the same call I modeled from. I treated TWST as a DNA-synthesis-plus-NGS vendor and never asked what happens if it becomes the wet lab for AI protein design. Orders lead revenue by 1-2 quarters; the leading indicator was in the transcript.	Missed the narrative shift that drove 3.4x of the 4.5x multiple expansion.
5	Non-operating assets	$54M Atlas Data Storage equity stake (now $68M) sat on the balance sheet, unmodeled.	Small, but sloppy.
6	Published a DCF / relative divergence instead of resolving it	My own relative section called 5.2x EV/S vs 5.0x peer median "reasonable". My DCF said -77%. When two methods disagree by 4x the DCF inputs are wrong, not the market.	Should have stopped the report.
7	Anchored on the post-earnings dump	-16% on a $0.04 EPS miss framed the narrative negatively when revenue beat and guidance was raised.	Same anchoring pattern as UPST and FROG V1.
8	Confused value with price	A DCF says what something is worth; it says nothing about when price converges. I issued SELL on a net-cash 20% grower with an accelerating order book and no negative catalyst.	A pricing call dressed as a valuation call.

Honest scorecard: errors 1-3 and 5 were input errors and are fully fixable (they explain roughly $10 to $20 of the gap). Error 4 is a framing error and explains most of the rest of the move. Errors 6-8 are process errors that would have prevented the wrong rating even with the wrong model.

4. Standing rules added to the framework

1. Beta. Never use raw regression beta when beta is above 1.5 or market cap is below $5B. Bottom-up industry unlevered beta, relevered, always. Hard bound on WACC of 7-12% for any US company with net cash; anything outside is an input error, not a finding.
2. Loss-making growth companies. Capitalize R&D (Damodaran method) and set reinvestment to zero in loss years, or run sales-to-capital from an R&D-capitalized margin. Never both a negative GAAP margin and a full reinvestment charge.
3. Terminal margin. Anchor to the company's stated long-term gross margin and incremental flow-through, cross-checked against best-in-class peer at maturity, not peer median today.
4. Order-book test. Any company that discloses orders, bookings or backlog gets orders modeled explicitly. Any segment with order growth above 50% is flagged as a narrative-shift candidate and gets a fourth scenario at 10-15% weight.
5. Consistency gate. If DCF and growth-adjusted EV/Sales comps diverge by more than 2x, the report does not publish until the divergence is explained in one paragraph.
6. Non-operating assets. Balance sheet line by line: equity investments, spin-out stakes, royalty and revenue-sale liabilities.
7. Value vs price. DCF output labels are OVERVALUED / FAIR / UNDERVALUED. A trade rating (BUY / SELL / AVOID) requires a separate pricing section: catalyst, momentum, float turnover, underwriter and insider activity, short interest. No SELL on a net-cash company growing above 15% without a named, dated negative catalyst.
8. Post-mortem cadence. Every published DCF is re-run at the next two earnings. Any price divergence above 50% triggers a versioned V2 with attribution split between fundamentals and multiple.

5. Catalysts and what would change the rating

• Nov 2026, Q4 FY26 print and FY27 guide. If FY27 revenue guide is above $600M (32%+) with GAAP operating loss narrowing below -20%, bull case probability rises to 40% and PWFV moves to roughly $50. Below $560M (23%) and the price has no cash-flow support at all.
• AI order conversion. $50M FY26 orders should show up as roughly $40-50M of FY27 therapeutics revenue. Watch DSPS therapeutics line, which grew 49% in Q3.
• Lilly TuneLab volume disclosure and any second pharma platform deal.
• Insider and underwriter behaviour. Form 144 cadence, any secondary from the August book.
• SBC. $72M run-rate on $457M revenue (16%). Adjusted EBITDA breakeven at Q4 still means roughly -$18M GAAP operating loss per quarter; do not let "breakeven" headlines substitute for owner economics.

Sources: SEC 8-K Ex.99.1 (Aug 3, 2026; Feb 2 and May 4, 2026), Q3 FY26 earnings call transcript, pricing release for Aug 4 2026 offering, Morningstar and Robinhood quotes Sep 17-18 2026, analyst target compilations via TipRanks. Rf and ERP are assumptions. Not investment advice. Prior model: Jan 25, 2026 DOCX report (SELL, PWFV $9.98).