Daniel Ek: Life After Spotify, Broken Healthcare Incentives, Catching Disease Early & AI's Potential

Summary

Daniel Ek joins the All-In hosts to revisit Spotify’s founding amid rampant piracy and no practical legal streaming market, then map the same “prove it in one market, expand country by country” playbook onto Neko Health. He argues that US healthcare is structurally reactive, expensive, and misaligned with prevention because incentives are tied to employment and short insurance horizons, while chronic disease can often be addressed earlier if better longitudinal data exists. Neko’s product combines bloodwork, high-resolution skin imaging, cardiovascular measures, grip strength, and clinician review into a roughly hour-long annual visit priced at $499, with AI used to triage skin findings before human escalation. Ek also says the company has already completed more than 100,000 scans, sees serious undiagnosed conditions in about 1% of members, and believes the economics work because of vertical integration. The conversation closes with AI strategy, including compute as a possible proxy for capability and risk, plus Spotify’s mix of open standards and proprietary features.

Notes

• Spotify began in 2006, launched in late 2008, then expanded to the UK and the US in 2011 after years of licensing negotiations.
• Ek frames Spotify’s origin as a response to piracy and the absence of good legal music options in Sweden.
• He says he and his team had to guarantee labels their bonus and budget for a year to secure a trial deal.
• The launch sequence was Sweden first, then the UK, then the US.
• Jason Calacanis notes Spotify’s current scale at over 700 million active users and over 300 million premium subscribers.
• Ek says he moved into executive chairman and is no longer running Spotify day to day.
• He says he only started thinking deeply about healthcare years ago but had been discussing it publicly as early as 2012 or 2013.
• Ek says he had no money until Spotify went public, so he was not an active angel investor early on.
• He says the US spends 18% of GDP on healthcare and heart disease alone costs hundreds of billions of dollars.
• He identifies cardiology, skin cancer/melanoma, and chronic disease as the biggest problem buckets.
• His core thesis is that many chronic diseases are preventable if discovered early.
• Ek says the missing ingredient is enough data to discover disease early.
• He says predictive healthcare needs better, multimodal, longitudinal datasets.
• He explicitly ties modern sensors and machine learning to the healthcare thesis.
• Neko Health’s current product is priced at $499 and is vertically integrated from facilities to clinicians to software.
• The visit measures 53 blood markers, takes more than 6,000 high-resolution skin images, and indexes every mole and lesion.
• The appointment includes same-visit clinician review of all results.
• Ek says the company has done more than 100,000 scans and publishes annual outcome data.
• He says about 1% of members have a serious undiagnosed underlying condition discovered by the scan.
• The company says the highest-risk members improve the most after receiving a clear health baseline and clinician guidance.
• Dermatology is a key AI use case because the average person at Neko has 950 moles.
• Ek says a normal physician cannot realistically inspect all of those lesions, so Neko uses AI first, then a human clinician, then dermatologists if needed.
• Longitudinal indexing lets the company compare lesions year over year and catch abnormal growth that doctors may not remember.
• Ek says Neko is not his sole focus; he remains executive chair of Spotify and also works with Prima Materia.
• He says the US healthcare system was built for infectious disease and acute symptom treatment, not prevention.
• He argues employer-linked insurance creates a prevention problem because people and insurers turn over too quickly for long payback periods.
• He says lowering upfront prevention costs from millions to tens of thousands could improve adoption even with a long payback period.
• Neko is releasing annual data and running clinical trials to build evidence across countries and populations.
• The company has four completed clinical trials, two underway, and four more planned.
• Ek says healthcare datasets are small by tech standards and that 10x to 100x more data would materially improve conclusions.
• Neko imports Apple Health data and can ingest wearable data from devices that sync to Apple Health.
• Ek says the product is already on Gen 2 diagnostic infrastructure, not a frozen one-off test.
• Ek argues AI’s value should be judged on positive applications in health and music, not only on downside risks.
• He says AI could eventually soundtrack every moment of life with better personalization.
• Ek supports open-source models and thinks open and closed systems will coexist.
• Spotify uses frontier models plus its own fine-tuned models to improve cost, efficiency, and task-specific performance.
• Ek says compute scale may signal model power and even cyber risk, contrasting large GPU clusters with running an open-source model on a home PC.
• David Sacks links compute to defensibility and argues access to the most compute may need guardrails.
• The discussion compares AI compute controls to old Cray supercomputer restrictions.
• Ek says Spotify’s podcast strategy mixes proprietary platform features with open-standard distribution.
• He says Spotify already helps podcasters distribute to other platforms through open standards.
• Calacanis pushes for live RSS notifications and value-and-donations support in podcasting.
• Ek says he will take the feature request back to the chairman/team.
• Ek recounts Stardoll’s origin as Paper Doll Heaven in Turku, Finland, later rebranded with Index Ventures and Matthias.
• He says Stardoll’s page load time was about four minutes before he re-architected it and cut it to under a second, which drove traffic and a Sequoia investment.
• Neko Health launched its first New York location at 300 Lafayette and plans expansion to Miami, D.C., and more U.S. cities within 12 to 24 months.