Discovery › screen
Screen health tech software
Date: 2026-07-29 · Scout/Screener · US-listed only Field: Software business models sold into healthcare (recurring revenue, high gross margin, network effects, switching costs) Excluded: VEEV (covered), all names in the standing exclusion list
Field Verdict Up Front
This field is NOT the 2021-bubble graveyard the brief hypothesized — it is a 2026 crash, and that changes everything.
The framing assumption was that these names have been down 70-90% for four years. That is true of the broken half of the field (HCAT −95%+, DH −85%, TDOC −97% from 2021, EVH, PHR). But the interesting half is different: DOCS, WAY, and the RCM/provider-software complex were making new highs as recently as Q4 2025 and got destroyed in the first half of 2026 by a single, specific, brand-new narrative: agentic AI disintermediating healthcare software take-rates.
- DOCS: $76.51 → $21.24 (−72%), essentially all of it in 2026 (−56% YTD, TIKR, May 2026)
- WAY: $41.47 → $20.94 (−49%), −43% YTD, while raising guidance
- CERT: −44%. SDGR: −33%. TEM: −58%.
So the drawdown is six months old, not four years old. The "down from a bubble" objection genuinely does not apply here — but a different objection does, and it is the one that matters: is agentic AI actually a terminal threat to these business models, or is it a narrative? That is the single question this field turns on.
My answer: it is terminal for RCM take-rate businesses and for thin "workflow" SaaS, and it is largely a category error for Doximity. Details below. This yields one genuine gem, one interesting-but-flawed second, and a lot of wreckage.
The techbio sub-field is a clean reject. See §Rejects — no survivors worth capital.
Top 3
1. DOCS — Doximity · Conviction 7/10 · The one genuine gem
| Metric | Value |
|---|---|
| Price / Market cap | $21.24 / $4.23B (fully diluted, 199.05M sh) |
| % off 52w high | −72.2% ($76.51) |
| P/E ttm / fwd | 21.7x / 13.3x |
| EV | $3.53B (~$700M net cash + investments) |
| EV/FCF · EV/EBITDA | 11.1x · 15.4x |
| FCF (FY26) | $317.5M on $644.9M revenue = 49.2% FCF margin |
| FCF yield | 7.5% on cap · 9.0% on EV |
| FCF CAGR (3y) | +22.3% ($173M→$178M→$267M→$317M — never declined) |
| Revenue CAGR (3y) | +15.5%; latest FY +13.1% |
| Share count trend | Diluted −2.3%/yr (213.4M→199.05M). Buybacks $431.7M in FY26 alone |
| SBC % of revenue | 18.9% ⚠️ — up from 12.7% / 10.7% / 11.4%. The flaw. |
| Debt/Assets | 0.91% — effectively debt-free |
| ROE / Gross margin | 19% (understated — equity is mostly cash) / 89% |
| Cash runway | N/A — deeply FCF-positive |
Moat — named honestly: network effects. This is the real thing, and it is rare in this field. Doximity is the physician graph: ~80% of US doctors and ~90% of medical students verified on-platform. The asset is not the software, it is the verified identity layer plus the professional graph, which no competitor can rebuild because doctors will only join the network that already has their colleagues on it. Secondary moat: regulatory/data intangible — HIPAA-compliant verified-prescriber targeting is what pharma actually pays for, and that verification apparatus took a decade. 89% gross margin and 49% FCF margin are what a genuine two-sided network throws off.
Versus its own multiple band (roic.ai): P/FCF 14.6x today vs. FY26 average 32.8x and FY25 average 32.2x; EV/EBITDA 15.4x vs. 41.8x a year ago; P/S 6.8x vs. a 14-15x historical average. This is the lowest valuation in its public life, and it is currently pinned at the floor of its own band (FY26 low P/FCF: 13.7x). Caveat per instruction: DOCS IPO'd June 2021, so the band is five years long and the top of it is bubble-contaminated — I am anchoring on the low, not the average, and even that is barely below today.
Adversarial bear case (the real one): 1. AI disintermediation of pharma ads. If physicians get clinical answers from an LLM instead of browsing Doximity, the ad inventory shrinks. This is the thesis that took the stock down. My rebuttal: Doximity's product isn't a search box, it's a reach channel — pharma buys verified-NPI-targeted access to a specific physician, which is a compliance-gated distribution asset, not an information-retrieval product. An LLM doesn't give Pfizer a legally-targetable list of cardiologists. The threat is real to Doximity's content surface, not its identity surface. 2. The genuine bear point is demand, not disruption. Management said HCP digital pharma ad spend is "soft" with limited visibility, and pharma has shifted to shorter-duration commitments under policy uncertainty (Benzinga, May 2026). Revenue is ~85% concentrated in pharma marketing budgets, which are cyclical and policy-exposed. Jefferies cut Buy→Hold, PT $51→$19. 3. ⚠️ SBC is the legitimate red flag. SBC nearly doubled ($72M→$122M) to 18.9% of revenue while revenue grew 13%. FCF less SBC = $196M, so the "true" owner-earnings yield is 4.6% on cap / 5.6% on EV, not 7.5%. Management is paying for the "AI investment year" in stock. Mitigant: buybacks ($432M) ran 3.5x SBC, so diluted share count still fell — §1.4 is passed on a net basis. But it is passed by spending cash to mop up dilution, which is a real cost, not a free one. Watch this number. 4. Governance noise, mostly defused. CFO Anna Bryson resigned April 2026 — but the 8-K shows it followed February medical leave, and guidance was reaffirmed at the same time (8-K, Apr 2026). SEC correspondence on file is routine 10-K/10-Q comment review, not a formal investigation. The "legal investigation" headlines are standard plaintiff-firm trawling post-drawdown. Downgrade this flag, don't ignore it — a permanent CFO hire is the tell. 5. Reimbursement/policy risk: LOW — uniquely for this field, DOCS is paid by pharma marketing budgets, not CMS. No reimbursement-code exposure. That is a meaningful structural advantage over every other name here.
Valuation. Graham IV = $10.70 vs $21.24 — I am discounting Graham heavily and saying so: it is near-useless for an asset-light, 89%-gross-margin, net-cash network with essentially no tangible book. FCF-based is the right lens per §3. - Bull anchor: $317.5M FCF at 20-24x EV/FCF (justified by 15%+ growth, net cash) → $6.4-7.6B EV + $0.7B cash → $36-42/sh - Bear anchor: SBC-adjusted $196M at 16-18x → $3.1-3.5B + $0.7B → $19-21/sh - Middle (FCF less half of SBC ≈ $257M, at 18-22x) → $27-32/sh
Fair value range: $27-34. Entry zone: below $23 (in-zone today). Strong-add below $19. The bear anchor lands almost exactly at today's price — meaning the market is currently pricing DOCS as if SBC is a permanent 19% tax and growth stops. You are being paid to disagree with only one of those two.
2. WAY — Waystar Holding · Conviction 5/10 · Cheap, growing, structurally threatened
| Metric | Value |
|---|---|
| Price / Market cap | $20.94 / $4.02B |
| % off 52w high | −49.5% ($41.47) |
| P/E ttm / fwd | 31.3x / 11.3x |
| EV · EV/FCF | $6.03B · 21.3x |
| FCF (FY25) | $283.2M ($29.9M → $142.5M → $283.2M — steep ramp) |
| FCF yield | 7.0% on cap · 4.7% on EV (net debt matters here) |
| FCF CAGR (3y) | +49.2% (off a depressed LBO base — flatters) |
| Revenue CAGR (3y) | +16.0%; organic ~11% |
| Share count trend | Diluted +3.5%/yr ⚠️ (166.7M→184.8M) — IPO + Iodine acquisition stock |
| SBC % of revenue | 3.8% ✅ (5.8% / 1.1% / 1.1%) — genuinely low |
| Debt/Assets | 25.8%; $1.49B debt. Goodwill $4.02B = 69% of assets ⚠️ |
| ROE / Gross margin | 4% / 69% |
Moat — named honestly: switching costs, moderate. Not network effects. Waystar sits inside providers' claims/billing workflow across ~1M providers and 5B+ annual claims; ripping it out mid-revenue-cycle is genuinely painful. But this is embedded-vendor stickiness, not a network — the payer connections it brokers are a clearinghouse commodity, and it is a PE rollup (EQT/Bain/CPPIB, 69% of assets is goodwill) that bought its scale rather than compounding into it.
Adversarial bear case — and here the bear is probably right: - The AI threat to RCM is structural, not narrative. PitchBook Q2 2026 models US RCM industry profits falling ~22% by 2030 and ~54% by 2040 as agentic AI compresses cost-to-collect below 1%. Waystar's revenue is levered to a take rate on claim volume — precisely the variable AI deflates. This is not the same situation as DOCS. - Epic is the existential problem. EHR incumbents can bolt RCM onto a platform providers already own and already pay for. Mizuho cut PT $50→$42 explicitly on competitive risk. A feature a platform-owner can ship is not a moat (§6.4). - Diluting 3.5%/yr and carrying $2B net debt and 69%-goodwill assets. Buybacks are ~zero. - Reimbursement/policy risk: HIGH — revenue tracks claim volumes and payer behavior. - Counterpoint worth stating: Waystar raised 2026 guidance and ranked #1 in the Black Book RCM AI survey. The company is executing; the category is what's impaired.
Fair value: $24-30 (13-16x SBC-adj FCF on EV, haircut for terminal-value impairment). Entry below $18. Cheap on today's cash flows, but §0 asks whether it's a good business — and a business whose industry profit pool is credibly modeled to halve is a cyclically cheap, structurally challenged name. That is closer to the value-trap quadrant than the value quadrant. Watch, don't buy.
3. HQY — HealthEquity · Conviction 6/10 as a business, 3/10 as a buy today · Great company, wrong price
| Metric | Value |
|---|---|
| Price / Market cap | $100.99 / $8.47B |
| % off 52w high | −4.7% ($105.96) ⚠️ at highs |
| P/E ttm / fwd | 37.8x / 18.5x |
| EV · EV/FCF | $9.42B · 23.1x |
| FCF (FY26) | $408.3M; FY25 was −$166M on a $505M capex year |
| FCF yield | 4.8% on cap — thin |
| Revenue CAGR (3y) | +15.1%; EPS $-0.31 → $2.46 |
| Share count trend | +1.2%/yr, but $299M buyback started — inflecting the right way ✅ |
| SBC % of revenue | 5.6% ✅ and falling (8.0% / 7.7% / 7.3%) — cleanest of the three |
| Debt/Assets | 29.6%; $1.0B debt, goodwill $1.65B |
| ROE / Gross margin | 11% / 71% |
Moat: switching costs + efficient scale, real. HSA custodian with ~17M accounts distributed through employer benefits channels — employers don't re-paper their HSA administrator casually, and custodial float income scales with balances at near-zero marginal cost. Operating leverage is proven: revenue +15% CAGR turned a −$26M net loss into +$215M.
Why it is not a buy: it fails §0's second question outright. At 4.7% off its 52-week high and a 4.8% FCF yield, the market has already priced the turn. Bear case: float income is a rate bet (Fed cuts compress the highest-margin revenue line), and it carries a history of custodial-data breach and service-cost issues.
Fair value $95-115. Entry below $80. Genuinely good business — put it on the watchlist and wait for a rate scare. Do not chase.
Rejects — with reasons
Techbio / AI drug discovery — the entire sub-field is a reject. Story vastly exceeds revenue.
| Ticker | Why rejected |
|---|---|
| RXRX Recursion | P/S 24.3x, EV/S 14.9x on revenue that FELL 56% yoy. Gross margin 0.00. Op margin −1,986%. 524.7M shares outstanding — a serial-dilution machine. Pure pipeline story, no commercial engine. Textbook §6 avoid. |
| TEM Tempus AI | EV/S 6.0x, net margin −22%, ROE −82%, beta 3.46, 31% short float. Revenue does grow 36%, but it's largely low-margin sequencing services dressed as an AI platform. Insider 37% / institutional 55% — float dynamics, not fundamentals. |
| SDGR Schrödinger | Revenue −2% yoy, op margin −83%, ROE −30%. The physics-based software half is genuinely good IP; it is subsidizing a cash-burning drug pipeline. Valuation rests on the pipeline. 25% short float. |
| CERT Certara | The near-miss. Real regulatory moat (FDA accepts its biosimulation models). But revenue +1% yoy, FCF CAGR −4.5% ($80M→$70M), unprofitable, 30% short float, goodwill $773M/$1.56B assets. Growth is dead and pharma R&D budgets are shrinking. Cheap for a reason — value trap. |
| SLP Simulations Plus | $370M cap, 63% GM, but growth broke (+7% and decelerating) and it's only −13% off its high at 45x ttm. Wrong price for a broken growth story. |
Provider/payer software wreckage — genuinely broken, four years down:
| Ticker | Why rejected |
|---|---|
| HCAT Health Catalyst | $154M cap, revenue −11%, net margin −88%, ROE −103%, EV/EBITDA 257x. Dying. |
| DH Definitive Healthcare | $0.72/sh, $76M cap, revenue −6%, net margin −71%, P/B 0.52. 84% gross margin can't save a shrinking business. |
| PHR Phreesia | Cut FY27 revenue guidance 6.7% at midpoint, hit all-time lows; 68% GM but 2% net margin, D/E 25.8. Growth story that stopped growing. |
| EVH Evolent Health | 20% gross margin — this is not a software business, it's specialty care management taking medical risk. Revenue −31% yoy in Q2, D/E 249, ROE −65%. |
| TDOC Teladoc | −97% from 2021 highs and still losing money (net −7%, ROE −12%) with revenue declining. Broken. |
Not software / fails the business-model test: PGNY (24% gross margin — fertility benefits risk-taking, and at 52w highs), HIMS (73% GM but it's a consumer telehealth/compounding-pharmacy model, beta 2.34, 32% short float, revenue +4% after the GLP-1 unwind), NTRA (15.4x sales at a 52w high — diagnostics, not software).
Field Verdict
This is not a field of broken 2021 bubble names — it is a field with a broken half and a freshly-crashed half, and the distinction is the whole opportunity.
The 2021-vintage wreckage (HCAT, DH, PHR, EVH, TDOC) is exactly as dead as four years of price action implies. Do not bottom-fish there — those are declining-revenue businesses with sub-scale economics, and cheapness is not a thesis.
The 2026 crash cohort is different, and it is driven by one narrative: agentic AI compresses healthcare software economics. That narrative is correct for take-rate/workflow businesses (WAY, and the RCM complex generally — the PitchBook profit-pool math is credible and I'd rather respect it than argue with it) and substantially wrong for identity/network assets (DOCS). The market has applied the same multiple compression to both. That is the mispricing.
The single name most worth a full /analyze: DOCS — Doximity.
It is the only asset in this field that clears both of §0's questions simultaneously: 1. Good business? Yes — 89% gross margin, 49% FCF margin, genuine network effect, debt-free, FCF that has never declined and compounds at 22%, and management that has retired 7% of the diluted share count in three years. 2. Priced in? No — 11x EV/FCF and 13x forward earnings, at the absolute floor of its own five-year multiple band, with the bear-case valuation anchor sitting at the current price.
The two things that must be interrogated in a full analysis, and which keep this at 7/10 rather than 9/10: - Is the 18.9% SBC a one-year "AI investment" spike or a new permanent run-rate? This single variable moves fair value from ~$20 to ~$40. It is the crux. - How cyclical is the pharma HCP ad budget, really? Revenue concentration in one buyer-category is the structural weakness, and the FY27 guide implies a genuine slowdown, not just conservatism.
Secondary: HQY to the watchlist as a quality business awaiting a better entry (below $80). WAY: monitor only — I am not willing to underwrite a business whose industry profit pool is credibly modeled to halve, however cheap it looks today.
Data Gaps & Caveats
fin.pyreturns ~4 annual periods; §1's 5-8yr FCF CAGR could not be computed for any name here. All CAGRs above are 3-year. For DOCS this is a hard limit — it IPO'd June 2021, so a 5-8yr public history does not exist.- DOCS multiple band is five years and bubble-contaminated at the top. I anchored to the band low (P/FCF 13.7x), not the average, and flagged this rather than treating the average as a target.
- DOCS has a dual-class structure: Yahoo's "SharesOut 128.86M" is Class A only. I used 199.05M fully diluted throughout, which raises market cap to $4.23B from the screened $3.82B and correspondingly lowers the FCF yield. Any figure sourced from a screener that uses the 128.86M count will overstate cheapness by ~10%.
- WAY's 49.2% FCF CAGR is computed off a depressed 2023 LBO-era base ($29.9M) and materially flatters the trend; organic revenue growth (~11%) is the honest number.
- HQY's FY25 FCF is negative (−$166M) due to a one-off $505M capex year, making its 134.8% FCF CAGR meaningless. I used the FY26 absolute figure.
- No formal SEC investigation into DOCS was found; filings on record are routine comment-letter correspondence. Treated as unconfirmed/defused rather than clean.
Sources: TIKR — Doximity down 40% · Benzinga — Doximity FY27 outlook, May 2026 · StockTitan — DOCS 8-K CFO resignation · PitchBook — AI Kills the RCM Star, Q2 2026 · PitchBook — Q2 2026 Healthtech Comp Sheet · Finviz — Mizuho cuts WAY PT · Simply Wall St — Phreesia · roic.ai valuation multiples (DOCS, FY25-FY26) · Yahoo Finance MCP / fin.py (all fundamentals, retrieved 2026-07-29)