Somewhere between a Nobel Prize for protein folding and a $305 million bankruptcy auction for a consumer DNA company, health tech found its identity crisis, and its future. The global AI-in-healthcare market surpassed $20 billion in 2025 and is accelerating. The FDA cleared 295 AI-powered medical devices last year alone - more than the entire decade prior combined. An AI-designed drug published Phase IIa results in Nature Medicine showing it can reverse lung disease. And a body-scanning startup co-founded by Spotify's CEO hit a $1.8 billion valuation on a 300,000-person waitlist. This is the state of health tech and AI in February 2026 - where the breakthroughs are real, the failures are instructive, and the gap between hype and evidence is finally narrowing.
The Race to AI-Discovered Drugs
For decades, drug discovery followed a brutal formula: 10-15 years, $2.6 billion average cost, and a 90% failure rate in clinical trials. AI promised to rewrite those numbers. In 2026, the first proof points are arriving, and the results are extraordinary.
Isomorphic Labs: DeepMind's Pharma Bet
Isomorphic Labs, the drug discovery spinoff built on DeepMind's Nobel Prize-winning AlphaFold technology, is making its biggest moves yet. In January 2026, the company announced its first AI-designed cancer drug entering Phase 1 clinical trials, announced in January 2026. The underlying technology keeps compounding: IsoDDE (Drug Design Engine), released in February 2026, doubles AlphaFold 3's already-leading accuracy on novel protein-ligand structures. AlphaFold 3 itself achieved 76.4% docking accuracy, 1.8x better than any prior computational method.
The commercial side is substantial. Isomorphic has locked in roughly $3 billion in partnership value across multi-year deals with Eli Lilly, Novartis, and Johnson & Johnson. Each deal follows the same structure: pharma provides disease targets, Isomorphic's AI platform designs the molecules. With over $600 million in total funding, the company is positioning AlphaFold not just as a research tool but as a drug design engine at industrial scale.
Insilico Medicine: First AI Drug in Phase II
If Isomorphic represents the frontier research approach, Insilico Medicine is the speed-to-clinic story. Its lead compound, rentosertib (ISM001-055), is historic: the first drug where both the biological target and the molecule were discovered entirely by AI. Phase IIa results, published in Nature Medicine in June 2025, showed that patients receiving the high dose (60mg daily) gained +98.4 mL of lung function, while the placebo group declined by 62.3 mL. That's a meaningful, dose-dependent reversal of idiopathic pulmonary fibrosis, a disease with limited treatment options.
Insilico went public on the Hong Kong Stock Exchange in December 2025 after closing a $110 million Series E earlier that year. Its Pharma.AI platform, comprising Chemistry42, Biology42, and Medicine42, has nominated 22 drug candidates since 2021 and maintains 30+ assets in its pipeline. In January 2026, a $120 million partnership with Qilu Pharmaceutical for cardiometabolic drugs further validated the platform model. The company is now engaging regulators on Phase IIb/III pivotal trials for rentosertib, with approval anticipated in 2026-2027.
Recursion Pharmaceuticals: Data at Scale
Recursion takes a different approach: less about designing individual molecules, more about building the largest biological dataset on the planet. The company sits on 50 petabytes of biological, chemical, and patient data, processed by BioHive-2, biopharma's most powerful owned supercomputer: 504 NVIDIA H100 GPUs, ranked #35 on the global TOP500 list. Its November 2024 merger with Exscientia for $688 million created the largest AI drug discovery company by pipeline breadth, with 10 clinical readouts expected and $200 million+ in milestone payments over 24 months.
The pipeline tells a mixed story. REC-4881 for familial adenomatous polyposis showed positive Phase Ib/II data in December 2025, with FDA registration pathway discussions planned for H1 2026. But REC-994 (cavernous malformation) and REC-2282 (neurofibromatosis) were both discontinued due to unfavorable efficacy. A reminder that AI doesn't eliminate clinical failure, it just gets you there faster.
The first AI-designed drug published Phase IIa results showing it can reverse lung disease. Fifteen to twenty AI-originated programs are expected to enter pivotal trials in 2026. The question has shifted from "can AI design drugs?" to "how fast can they reach patients?"
The Broader Landscape
Beyond the three leaders, the field is maturing rapidly. Generate Biomedicines has GB-0895 in Phase 3 for severe asthma, with two additional programs clearing IND in December 2025. Relay Therapeutics earned FDA Breakthrough Therapy Designation for zovegalisib in HR+/HER2- breast cancer. Chai Discovery raised a $130 million Series B at a $1.3 billion valuation in December 2025, backed by OpenAI and General Catalyst, claiming 100x improvement in antibody design success with its Chai 2 platform.
The investment numbers tell the story: over $100 billion invested in AI drug discovery over five years, with 50% deployed in the last two alone. There are now 173 AI-discovered drug programs in clinical development. The market, estimated at $2.9-24.5 billion in 2026 depending on methodology, is growing at 24.8% CAGR. And in January 2026, a joint FDA-EMA guidance laid out 10 principles for AI in drug development , signaling that regulators are building frameworks for approval, not barriers to it.
The Preventive Health Revolution
The most radical shift in health tech isn't happening in hospitals. It's happening before patients ever get sick. A new generation of companies is betting that AI-powered screening, done early enough and often enough, can fundamentally change healthcare economics.
Neko Health: The Body Scanner
Daniel Ek, the Spotify co-founder, is applying the same logic to healthcare that he applied to music: make it cheap, make it easy, and let the data do the work. Neko Health offers a 50-minute full-body scan using thermal cameras, 3D imaging, ECG, lasers, LIDAR, and real-time blood analysis - all processed by AI to detect early signs of heart disease, skin cancer, and Type 2 diabetes.
Neko raised a $260 million Series B in January 2025 at a $1.8 billion valuation. The global waitlist exceeds 300,000 people. The company delivered 10,000 scans across six clinics in London, Manchester, Birmingham, and Stockholm in 2025, 6x more than the prior year. The retention rate is staggering: 80% of customers prepay for their next annual scan. A flagship New York clinic opens in spring 2026 with capacity for 30,000 scans per year. At £299 per scan, the unit economics work, and Neko is proving that consumers will pay out-of-pocket for preventive AI when the experience is right.
The FDA's AI Diagnostic Explosion
The FDA has now authorized 950 AI/ML-enabled medical devices cumulatively, with 295 cleared in 2025 alone. For context, only 33 were cleared in the entire period from 1995 to 2015. Radiology dominates: 723 devices, or 76% of all approvals - but the accuracy numbers across specialties are notable.
In breast cancer screening, AI achieves 90% sensitivity versus 78% for radiologists alone, with early detection rates of 91% versus 74%. AI mammography could cut radiologist workload by roughly 50%. In pathology, a meta-analysis across 152,000+ slides showed 96.3% sensitivity and 93.3% specificity. Paige's PanCancer Detect earned Breakthrough designation in April 2025 as the first AI system covering both common and rare cancer variants. In ophthalmology, four FDA-approved systems for diabetic retinopathy detection now operate autonomously, with AEYE-DS achieving 92.98% sensitivity and 91.36% specificity. In dermatology, multi-model AI reached over 92% accuracy for melanoma detection as of January 2026.
The median review time for AI devices at the FDA dropped to 142 days in 2025, with 25% cleared in under 90 days. The regulatory machine is accelerating, not constraining.
Genomics After 23andMe
23andMe's Chapter 11 bankruptcy in March 2025 shook the genomics world (more on that below). But the underlying technology is thriving. Illumina launched the Billion Cell Atlas in January 2026, the world's largest genome-wide genetic perturbation dataset - in partnership with AstraZeneca, Merck, and Eli Lilly. The three-year initiative aims to build a 5-billion-cell atlas for AI-driven drug discovery and target validation. The precision medicine market hit $125.7 billion in 2026 and is projected to reach $237.3 billion by 2031, with AI/ML as the fastest-growing segment at 17.6% CAGR.
950 FDA-cleared AI devices. 295 in 2025 alone. AI outperforming radiologists on breast cancer detection by 12 percentage points. The evidence base for AI diagnostics isn't a pilot program anymore. It's a flood.
Wearables Go Clinical
Health wearables in 2026 aren't fitness trackers with a pulse sensor. They're FDA-cleared medical devices generating clinical-grade data, integrated into electronic health records, and covered by insurance. The transformation happened faster than the industry expected.
Oura Ring: The $11 Billion Ring
Oura's trajectory is the wearables success story of the decade. The Finnish company raised a $900 million Series E in October 2025 at an $11 billion valuation, making a ring the most valuable wearable startup on the planet. Revenue doubled for the second consecutive year, hitting $500 million+ in 2024 with projections of $1.5 billion+ for 2026. Over 5.5 million rings sold as of September 2025.
The Ring 4 introduced Smart Sensing Technology that adapts to individual physiology, recessed sensors in a sleeker titanium design, and up to 8 days of battery life. Clinical validation backs the hype: studies show strong agreement with polysomnography for sleep staging across 421,045 analyzed epochs, and the National University of Singapore ranked Oura as the most accurate consumer wearable for sleep tracking. Research partnerships with Mayo Clinic (adrenal disorders) and Stanford (menstrual health across diverse populations) are generating peer-reviewed evidence.
The strategic play is convergence. A $75 million investment from Dexcom made Dexcom G7 the only CGM compatible with Oura Ring - creating an integrated health ecosystem that combines metabolic and physiological data in a single platform.
Whoop: From Athletes to FDA Clearance
Whoop has evolved from a niche athletic tracker to an FDA-cleared medical device. The Whoop 5.0 earned FDA clearance for both ECG and blood pressure monitoring, the first wearable to achieve the latter. The new Healthspan metric quantifies "Pace of Aging," while Whoop Advanced Labs integrates periodic blood biomarker testing through Quest Diagnostics, combining continuous wearable data with blood work. Over 350,000 people joined the Advanced Labs waitlist within months of the May 2025 preview.
The partnerships tell the story of mainstream legitimacy. Ferrari Formula One (January 2026) deployed Whoop across drivers, pit crew, and engineers. The Ryder Cup (September 2025) provided unprecedented real-time golfer biometrics during NBC broadcasts. British & Irish Lions made Whoop the official supplier for their 2025 Australia tour. Valued at $3.6 billion with $260 million in 2025 revenue, Whoop is bridging the gap between elite performance and clinical health monitoring.
Apple Watch: Blood Pressure Goes Mainstream
Apple Watch Series 11 brought the biggest health feature since ECG: FDA-cleared hypertension notifications, rolled out to 150+ countries in September 2025. The optical cardiac sensor analyzes blood vessel response to heartbeats over a 30-day rolling window. The FDA validation showed 92.3% specificity - meaning very few false alarms - though sensitivity sits at 41.2%, a deliberate tradeoff since blood pressure cuffs are cheap and the condition develops slowly.
Non-invasive glucose monitoring remains the white whale. Despite persistent rumors, no Apple Watch glucose feature has shipped as of February 2026, with expert timelines pointing to 2027 at earliest. The workaround: third-party CGM integrations (Dexcom G7, Abbott Libre 3) pipe glucose data into the Apple Health ecosystem.
The CGM Revolution Beyond Diabetes
Continuous glucose monitors are the sleeper hit of health wearables. The CGM market hit $15.3 billion in 2026, growing at 15.4% CAGR, with a total addressable market estimated at $47 billion. Abbott Libre holds 56.7% revenue share, Dexcom 35.2% - but the real story is the non-diabetic consumer segment.
Non-diabetic users now represent 41.5% of the CGM market. Dexcom's Stelo OTC CGM, cleared for non-prescription use, brought AI-enabled food logging via smartphone photos and personalized glucose-food correlation insights. The US OTC CGM market - $48.6 million in 2024 - is projected to nearly double to $93.5 million by 2033. Wellness, not disease, is driving adoption. The Dexcom G7's 15-day sensor achieved 8.0% MARD - the lowest mean absolute relative difference ever reported in an FDA iCGM study, meaning clinical-grade accuracy in a consumer device.
A smart ring company is worth $11 billion. A fitness band has FDA-cleared blood pressure monitoring. Glucose monitors designed for diabetics are being worn by wellness consumers. The line between consumer wearable and medical device has effectively disappeared.
The Graveyard & Hard Lessons
Not everything worked. The failures of 2025 are as instructive as the successes.
23andMe: The Consumer Genomics Collapse
23andMe's Chapter 11 filing in March 2025 was the highest-profile failure in health tech history. The company that popularized consumer DNA testing - valued at $6 billion at its 2021 SPAC peak - was acquired for just $305 million. The lesson: consumer curiosity about ancestry and genetic traits doesn't convert to a sustainable health business. The data was the asset, but monetizing it proved harder than sequencing it.
Woebot: AI Therapy Hits a Wall
Woebot, the AI mental health chatbot, shut down entirely on June 30, 2025, with all user data anonymized by July 31. The clinical evidence actually supported Woebot - randomized controlled trials published in JMIR showed significant depression improvements over two weeks, with success among postpartum women and college students. But the business model couldn't survive. The industry is rapidly shifting toward hybrid models: AI-powered tools with licensed professional oversight, exemplified by Wysa, which earned FDA Breakthrough Designation in 2025 and has 30+ peer-reviewed studies showing efficacy comparable to in-person counseling.
Samsung Galaxy Ring: Patent Wars
Samsung's Galaxy Ring launched to sold-out preorders and expanded to 53 markets, but 2025 results disappointed commercially. Then in November 2025, Oura filed a patent infringement lawsuit, delaying the Galaxy Ring 2 to late 2026 or 2027. The smart ring category - growing 49% year-over-year versus 6% for smartwatches - is consolidating around Oura's early-mover advantage. Samsung's entry validated the form factor but couldn't overcome the incumbency gap.
The Regulatory Landscape
Two regulatory frameworks are reshaping health tech simultaneously, and companies must navigate both.
The FDA has accelerated its AI device pipeline dramatically - 295 clearances in 2025 with a 142-day median review time. The January 2025 draft guidance established a 7-step credibility framework for AI in regulatory decision-making. In January 2026, a joint FDA-EMA guidance laid out 10 principles for AI across the drug development lifecycle. But gaps remain: only about 50% of cleared devices have published clinical performance data, and less than a third provide sex-specific outcomes.
The EU AI Act entered force in May 2025, with penalties enforcement starting in August 2025. Medical AI systems - including SaMD, AI diagnostics, and regulated health devices - are classified as high-risk, requiring quality management systems, transparency documentation, human oversight protocols, and post-market monitoring. The compliance deadline for medical devices is August 2, 2026. Companies must satisfy both EU MDR and AI Act requirements simultaneously - a resource-intensive dual compliance burden that particularly strains smaller firms.
Market Reality Check
The numbers across health tech paint a picture of rapid, sustained growth across every segment.
The Funding Landscape
Capital continues to pour into health tech AI at record levels. The marquee rounds from 2025-2026 tell the story:
- Oura: $900M Series E at $11B valuation
- Recursion-Exscientia merger: $688M all-stock deal
- Neko Health: $260M Series B at $1.8B valuation
- Chai Discovery: $130M Series B at $1.3B valuation
- Insilico Medicine: $110M Series E + Hong Kong IPO
- Dexcom → Oura: $75M strategic investment
Market Projections
Every segment is growing at double-digit rates:
- AI drug discovery: $2.9-24.5B in 2026 (24.8% CAGR)
- AI diagnostics: $1.94B in 2025 → $10.3B by 2034 (20.4% CAGR)
- Wearable medical devices: $117B in 2026 → $500B+ by 2035
- CGMs: $15.3B in 2026 (15.4% CAGR)
- Remote patient monitoring: $88B projected by 2030
- Precision medicine: $125.7B in 2026 → $237.3B by 2031
The Integration Shift
The bigger story is wearable data flowing into clinical workflows. Epic, Cerner, and other major EHR platforms now accept data streams from Apple Health and Fitbit. Medicare's 2026 RPM changes reduced the minimum data collection period to just 2-15 days and management time to 10+ minutes, dramatically expanding coverage for post-surgical recovery and episodic conditions. Over 100 million Americans with chronic conditions are now monitored via remote patient monitoring. The business model is shifting from device sales to continuous health data services - and insurance is starting to pay for it.
What's Next
Here's what to watch over the next 12-24 months.
Near-Term (2026)
- Insilico will pursue Phase IIb/III pivotal trials for rentosertib - potentially the first AI-designed drug to seek regulatory approval
- Isomorphic Labs will advance its first AI-designed cancer drug through Phase 1, with molecules from Eli Lilly and Novartis partnerships entering trials
- Neko Health will open its New York flagship, testing whether American consumers embrace preventive body scanning at scale
- EU AI Act compliance deadline (August 2, 2026) will force every health AI company to meet high-risk classification requirements
- 15-20 AI-originated drug programs are expected to enter pivotal trials
Medium-Term Trends
- Wearable-clinical convergence will accelerate as insurance reimbursement expands and EHR integration matures
- AI diagnostics will push beyond radiology into primary care, with AI-powered triage becoming standard in telehealth workflows
- CGMs for wellness will continue mainstreaming, with non-diabetic users potentially overtaking diabetic users by 2028
- Hybrid AI mental health - AI tools with human oversight - will replace both pure chatbot and traditional models
- Preventive health economics will improve as companies like Neko prove that catching disease early costs less than treating it late
The defining question for health tech in the next two years isn't whether AI works - the clinical data is overwhelming. It's whether the healthcare system can integrate these tools fast enough to realize the economic and human benefits they promise.
The Bottom Line
Health tech and AI in 2026 sits at an inflection point where evidence is replacing hype. The first AI-designed drug has published Phase IIa results in a top medical journal. The FDA has cleared nearly 1,000 AI devices. A smart ring company commands an $11 billion valuation on $500 million+ revenue. A body scanner startup has a 300,000-person waitlist. Wearables have FDA clearance for blood pressure and ECG. Glucose monitors are worn by wellness consumers, not just diabetics.
But the gap between technology and adoption remains. Only half of FDA-cleared AI devices publish clinical performance data. 23andMe's bankruptcy showed that consumer genomics couldn't sustain itself. Woebot proved that clinical evidence alone doesn't guarantee a business. And the EU AI Act's compliance deadline will test whether health AI companies can navigate dual regulatory frameworks without breaking pace.
What's different about this moment is the convergence of clinical evidence, regulatory clarity, and economic incentive. AI drug discovery has published results, not just press releases. AI diagnostics outperforms human specialists on validated metrics. Wearable data is flowing into medical records and getting reimbursed by insurance. The infrastructure for AI-driven healthcare isn't theoretical anymore - it's being built, tested, and scaled.
The next 24 months will determine whether health tech follows the trajectory of AI drug discovery - slow, rigorous, and ultimately transformative - or the trajectory of consumer genomics - fast, hyped, and ultimately unsustainable. The evidence so far suggests the former. The technology works. The money is flowing. The question is whether the system can keep up.