How it works
The full methodology — sources, formulas, and default assumptions — published rather than held back. If you're going to disagree with a number, you should be able to see exactly how it was built.
From ticker to thesis in seconds
You search
Enter any biotech ticker and our AI agents activate.
We gather
Agents pull from SEC, FDA, ClinicalTrials.gov, and more in real-time.
We read
AI synthesizes data into valuations, bull/bear cases, and catalyst timelines.
You decide
Adjust assumptions, explore scenarios, and make informed decisions.
Where the data comes from
The same public sources institutional investors rely on—accessed in real-time, synthesized by AI.
SEC EDGAR
10-K, 10-Q, 8-K filings, proxy statements, and insider transactions
ClinicalTrials.gov
Trial status, endpoints, enrollment, sites, and completion dates
FDA databases
Approval history, PDUFA dates, AdCom schedules, warning letters
PubMed and literature
Mechanism of action research, clinical data publications, review articles
Company press releases
Pipeline updates, partnership announcements, financial guidance
Real-time market data
Stock prices, market cap, trading volume, analyst coverage
Valuation methodology
Institutional-grade models used by sell-side analysts and hedge funds. Every assumption is transparent and adjustable.
Risk-Adjusted NPV (rNPV)
Sum-of-parts pipeline valuation
- Each pipeline drug valued individually based on phase and indication
- Probability of success (PoS) applied by development phase
- Peak sales estimated from market size, penetration, and pricing
- Cash flows discounted at biotech-appropriate rates (12-15%)
- Terminal value calculated for approved/commercial drugs
Core Formula
Drug Value = (Peak Sales × Margin × PoS) / (1 + r)^tDiscounted Cash Flow (DCF)
10-year revenue projection model
- Revenue projections based on pipeline commercialization timeline
- Operating margins estimated from comparable companies
- R&D and SG&A expenses modeled based on development stage
- Working capital and capex assumptions from historical data
- Terminal value using perpetuity growth or exit multiple
Core Formula
Enterprise Value = Σ(FCF_t / (1+WACC)^t) + Terminal ValueComparable Company Analysis
Peer-based relative valuation
- Peers selected by therapeutic area, development stage, and market cap
- EV/Revenue multiples for commercial-stage companies
- EV/Pipeline Value for pre-revenue biotechs
- Adjusted for pipeline depth, cash position, and catalyst timeline
- Provides sanity check against fundamental valuation
Core Formula
Fair Value = Peer Multiple × Company MetricDefault probability of success by phase
These are FuzeBio's starting assumptions, not published industry figures. They are anchored on the phase-transition rates in the BIO/Informa/QLS dataset◆ (source: BIO / Informa / QLS, 2011–2020, opens in a new tab) and then adjusted by therapeutic area, where success rates differ materially.
Every one of them is editable per drug — if you think a Phase 2 oncology asset deserves better than 25%, say so and the valuation follows.
| Phase | Oncology | Rare Disease | Other Indications |
|---|---|---|---|
| Preclinical | 5% | 8% | 7% |
| Phase 1 | 10% | 15% | 12% |
| Phase 2 | 25% | 35% | 30% |
| Phase 3 | 50% | 65% | 58% |
| Filed/Under Review | 85% | 90% | 88% |
Why not just ask ChatGPT?
Generic AI tools hallucinate drug names and use stale data. We built purpose-specific infrastructure for biotech research.
| Feature | FuzeBio | ChatGPT |
|---|---|---|
| Data freshness | Real-time. Pulls latest SEC filings, trial updates, and market data on every query. | Training cutoff. May reference outdated trial status, old stock prices, or missed catalysts. |
| Source transparency | Every claim linked to source. See exactly where data came from. | Black box. Cannot verify claims or check original sources. |
| Structured output | Biotech-specific format: pipeline tables, catalyst calendars, valuation models. | Generic text responses. No standardized investment analysis structure. |
| Adjustable assumptions | Change PoS, peak sales, discount rates. See how it affects fair value. | Static answers. Cannot interact with or modify the analysis. |
| Hallucination risk | Grounded in real data. Cannot invent drug names or fake trial results. | Known to hallucinate drug names, clinical results, and company information. |
See it on a real company
Methodology is easier to judge on a finished report than in the abstract. Open one and follow the numbers back — which sources were read, how the fair value was assembled, and what every assumption was set to.
Try it yourself
Now that you've seen how the numbers are built, go build some. Seven days, full access, no card.