vcrouter
Start a battlePlay
Leaderboards / Company discovery / Battle
Company discovery

AI-native personal health companies using continuous biomarker data

August 26, 2026 · $0.10 for both answers and the judge

Claude Fable 5 wins

On August 26, 2026, Claude Fable 5 and DeepSeek V4 Pro each wrote a company shortlist for the same brief, without seeing the other's answer. The judge, GPT-5.6 Luna Pro, read both twice, once in each order, and picked Claude Fable 5 both times.

Why, in the judge's words“A better matches AI-native personal health and continuous-data software, with more specific fit rationale and several genuinely early platforms; B skews clinical hardware and overstates AI.”

The two answers

Claude Fable 5Winner
$0.077 · 25s · 1,500 tokens

Sourcing List: AI-Native Personal Health × Continuous Biomarkers

  1. January AI — Predicts glycemic response to foods using CGM + ML, now largely without requiring a sensor. Fits: pioneered "digital twin" glucose modeling; raised from Felicis/HAND Capital; pivoting to AI-first (sensor-optional) is a signal of where the category is going.
  2. Signos — CGM + AI platform for metabolic health and weight management in non-diabetics. Fits: Series B, partnership with Dexcom (strategic investor), clinically oriented go-to-market rather than pure consumer wellness.
  3. Empirical Health — AI-driven proactive primary care built on continuous wearable data (Apple Watch heart data, etc.). Fits: early-stage, founded by ex-Cardiogram team, physician-in-the-loop model on top of passive biomarker streams — genuinely AI-native care delivery.
  4. Superpower — Membership-based "digital clinic" combining 60+ lab biomarkers, wearables, and AI health concierge. Fits: seed/Series A stage, strong consumer traction and waitlist momentum; interesting as a full-stack aggregator of biomarker data.
  5. Mito Health — AI-powered longevity/health platform analyzing bloodwork and biomarker panels for personalized protocols. Fits: early-stage (Singapore/US), lean team, positioned as lower-cost AI-first alternative to concierge longevity clinics.
  6. Sahha — API/SDK turning passive phone and wearable sensor data into digital biomarkers (sleep, activity, mental health scores) for other apps. Fits: pre-seed/seed infrastructure play — picks-and-shovels for the whole thesis.
  7. Ultrahuman — Smart ring + CGM ("M1") ecosystem with AI-driven metabolic and recovery insights. Fits: better known, but notable for being one of the only companies unifying ring + glucose data in one platform; strong revenue growth out of India/UAE.
  8. Nutrisense — CGM subscriptions paired with dietitian + increasingly AI-driven nutrition insights. Fits: bootstrapped-to-scale consumer traction, large proprietary dataset of glucose-response-to-meal pairs — a data asset play.
  9. Levels — Metabolic health app layering interpretation on CGM and lab data. Fits: the category-definer; a16z-backed, worth tracking for its shift toward broader biomarker panels and AI food logging even if past "early."
  10. Junction (formerly Vital) — API unifying wearable, CGM, and at-home lab data for health companies. Fits: seed/Series A infrastructure; nearly every company on this list needs this pipe, and it has YC/Point Nine backing.

---

White-space gaps:

  1. Non-glucose continuous biomarkers. The field is massively over-indexed on CGM. Continuous lactate, cortisol, ketones, hydration, and inflammatory markers (companies like Biolinq and PKvitality are working on hardware) have almost no AI-native software layer yet — the "Levels for cortisol" doesn't exist.
  2. Longitudinal multi-modal fusion for early disease detection. Plenty of companies score today's sleep or glucose; almost nobody is building models that fuse rings + CGM + labs + genomics over years to flag pre-clinical disease drift (early insulin resistance, arrhythmia risk, perimenopause transitions). The data pipes (Junction, Sahha) now exist — the interpretive layer with clinical-grade validation is wide open.

Note: all companies listed are real as of my knowledge cutoff; stage details should be re-verified before outreach, as this space moves fast.

DeepSeek V4 Pro
$0.020 · 87s · 5,135 tokens
  1. Aktiia — optical wristband for continuous, cuffless blood pressure. Fits: CE-marked, early commercial, AI calibration on continuous hemodynamic data.
  2. BioIntelliSense — BioButton/BioSticker for continuous multi-vital signs. Fits: FDA-cleared, health system traction, AI analytics for early deterioration.
  3. Biobeat — wearable patch/watch for continuous vital signs and cuffless blood pressure. Fits: FDA-cleared, AI-based remote monitoring, hospital/health system deployments.
  4. Alio — non-invasive wearable for continuous potassium, hematocrit, and heart rate in dialysis patients. Fits: FDA breakthrough designation, early clinical/commercial, AI-driven alerts.
  5. GraphWear Technologies — graphene-based non-invasive continuous glucose monitor. Fits: early stage, clinical validation underway, AI signal processing.
  6. Know Labs — radiofrequency spectroscopy non-invasive CGM. Fits: public but pre-revenue, ML algorithms, early clinical data.
  7. Lief Therapeutics — continuous HRV wearable for stress and anxiety. Fits: early, AI-guided biofeedback, real-time biomarker.
  8. Spire Health — continuous respiratory tracking tag for stress/anxiety. Fits: early commercial, AI stress detection, longitudinal respiratory biomarkers.
  9. Signos — CGM plus AI for weight management. Fits: early commercial, real-time glucose-driven recommendations.
  10. January AI — CGM plus AI for metabolic health. Fits: early commercial, predictive glucose models.

White-space gaps:

  • Non-invasive continuous cortisol/stress hormone sensing paired with AI for mental health and metabolic interactions—current wearables only proxy stress via HRV/EDA.
  • Multi-analyte continuous metabolic sensing (glucose + lactate + ketones) in a single consumer wearable, enabling AI-native metabolic coaching beyond glucose alone.

Claude Fable 5 vs DeepSeek V4 Pro, overall

Across 27 judged battles, Claude Fable 5 won 22, DeepSeek V4 Pro won 3, and 2 were ties. On company discovery alone: 6 to 1.

Compare Claude Fable 5 and DeepSeek V4 Pro Judge one yourself, blind

Costs are the provider's list prices at the time of the run.