Spotify founder Daniel Ek argues that while most attention is on software (bits), the biggest opportunities are in hardware and physical industries (atoms), like green steel and healthcare, which are harder but have huge potential. He thinks Europe can lead in these areas rather than chasing the US and China in tech. Key holdings mentioned: Spotify (his company, creators grew from 3M to 8M); TikTok (cited as a sign platform innovation isn't over); Ginkgo Bioworks (compared to AWS for biology, meaning it provides infrastructure).
This episode of Invest Like the Best features Spotify founder and CEO Daniel Ek, discussing the differences between the bit world and the atom world, Europe's tech potential, and the evolution of the creator economy. Core insight: Daniel believes that early-stage atom companies (e.g., hardware) are
Daniel Ek (Spotify founder and CEO) discusses with host Patrick O'Shaughnessy the differences between the bit world and the atom world, Europe's tech potential, and the evolution of the creator economy. Core thesis: Daniel argues that early-stage atom companies (e.g., hardware) are harder to build than bit companies (e.g., software), but the collection and observation of data pools will unlock downstream opportunities—"The first step is to create an engine that collects data, find a reasonable data structure proxy, and only much later can simple inferences be made."
Daniel points out that most of the world's attention is currently focused on "bits" (digital companies), yet the biggest problems facing society are almost all "atomic" issues (climate change, healthcare, infrastructure, etc.). Bit companies have established standardized platforms—from financing (transparent Series A and B processes) to distribution (Google, Facebook, App Store, AWS), the entire ecosystem is mature, with extremely low experimentation costs. "Any internet idea today likely already has 100 people competing for it."
In contrast, atomic companies face entirely different challenges:
Daniel uses the steel industry as an example: Green steel technology has existed since the 1960s-70s but has never been attempted at scale. The issue is not technology but a "coordination problem"—requiring over $100 billion in project financing, top-tier talent, and long-term patient capital. If solved, it could reduce global emissions by 7-8% in one go, with production costs comparable to current steel at scale.
Daniel believes the healthcare industry faces similar challenges, but is more complex. U.S. healthcare is the largest employer; if current growth rates continue, it will become the sole employer—this is clearly unsustainable.
Core issue: The healthcare system is reactive—"the entire system is built on diagnosing problems only when symptoms appear, but by the time symptoms show, it is often too late." Daniel uses a car diagnostic analogy to illustrate the gap:
Solution: A new diagnostic system must be established to collect more data, enabling a shift from reactive to proactive care. Daniel emphasizes that this mirrors Spotify's evolution path—"the first step is to create an engine for data collection, find a reasonable proxy for data structure, and only much later can simple inferences be made."
Falsification condition: If the healthcare data collection system cannot achieve scaled parallel experiments within 10 years (similar to the shift from single experiments to hundreds of concurrent experiments in the bit world), then Daniel's assessment may be overly optimistic.
Daniel shared his zero-to-one learning framework:
1. Accelerate learning through podcasts: Listen at 2x speed, leveraging questions from skilled hosts to gather information
2. Be willing to travel and meet face-to-face: "I will travel, meet people, and spend the necessary time to gain real insights"
3. Learn by teaching: Explain concepts to others, identify gaps in your own knowledge, then dive deeper
4. Validate by inversion: Only truly master a concept when you can explain why the opposite scenario does not hold
Daniel emphasized that this process requires "being the dumbest person in the room for a very, very long time," and human psychology is not adapted to this state.
Daniel proposed three key decision-making tools:
1. One-Door vs. Two-Door Decisions (a concept popularized by Amazon)
2. Reducing Variables
3. Inversion Thinking (Charlie Munger's framework)
Daniel recommends learning several new frameworks from different disciplines each year, because "decision-making is the primary tool of executives."
Daniel points out that Europe has a population of 500–700 million, a GDP comparable to that of the U.S., and respectable rankings for its research institutions and universities. However, it has clearly lagged behind in the tech sector over the past 20 years. Reasons:
But Daniel believes Europe's real opportunity lies not in the bit realm (chasing the U.S. and China), but in the atomic realm (climate, healthcare, etc.):
Readers should note: Daniel is European (Swedish), so his judgment carries a regional perspective and should be cross-validated with other sources.
Daniel Ek believes the creator economy is still in its early stages, with audio being even earlier than video.
Daniel notes that Spotify’s creator base has grown from approximately 3 million three years ago to around 8 million today, and is expected to reach 50 million in the coming years. Key trends:
Findings on podcasts:
Daniel believes blockchain could play a role in scaling business development—specifically, problems that require extensive bilateral negotiations. Music IP licensing and usage fall into this category. However, the challenge lies in "too many stakeholders in the value chain, making it difficult to find a solution that satisfies everyone."
Daniel argues that the smartest record labels have already realized "they are serving creators." If he were a record label CEO, he would push "empowerment" to the extreme—"the more you empower, the better the outcome tends to be." He cites the gaming industry as an example: from paid games → free-to-play with ads → free-to-play with in-app purchases, the more open the model, the higher the revenue.
Falsification condition: If the licensing process for music IP does not significantly simplify over the next five years (via blockchain or other technologies), Daniel’s assessment that technology can solve copyright issues may be overly optimistic.
| Position | Analyst View | Key Data |
|---|---|---|
| Spotify | Bullish (Core Holding) | Creators grew from 3 million to 8 million, projected to reach 50 million; podcasts exceeded 300 episodes over 5 years, reaching nearly 30 million people |
| TikTok (ByteDance) | Neutral (Industry Case Study) | Demonstrates platform innovation is far from over |
| Bad Bunny | Neutral (Trend Case Study) | Most streamed artist globally in 2020, Latin music |
| Clubhouse | Neutral (Trend Case Study) | Proves audio format innovation continues |
| Patreon | Neutral (Trend Case Study) | Creator business model innovation during the pandemic |
| Ginkgo Bioworks | Bullish (Infrastructure Analogy) | Analogized as "AWS for biology" |
| Shopify | Neutral (Comparative Case Study) | "Empowerment vs. Aggregation" comparison with Amazon |
1. "The first step is to build an engine that collects data and finds a reasonable data structure proxy; only much later can simple inferences be made" (Daniel Ek) — Spotify took over 10 years to go from a unified homepage → by country → by demographics → partial personalization → full personalization; the healthcare industry needs a similar path.
2. "Green steel technology has existed since the 1960s-1970s; the problem is not technology but coordination — requiring project financing of over $100 billion" (Daniel Ek) — Steel accounts for 7-8% of global emissions; at scale, its price could match current steel.
3. "If you remove San Francisco and Seattle from the U.S., the U.S. doesn't look much better than Europe" (Daniel Ek) — The main reason for Europe's tech lag is the lack of superclusters, not a shortage of talent or capital.
4. "Europe's real opportunity is not catching up with the U.S. and China in the bit domain, but in the atom domain — where the playbook has yet to be written" (Daniel Ek) — European governments have invested decades in areas like climate and healthcare, potentially leading in these fields.
5. "One-way door decisions are hard to reverse; two-way door decisions are easy to enter and exit — most people treat all decisions as either one-way or two-way, both dangerous" (Daniel Ek) — Distinguishing decision types is the first step to efficient decision-making.
6. "Reduce variables: 'make up numbers' for other variables, create a decision tree — if all paths are low-cost, just pick one" (Daniel Ek) — In multi-variable problems, simplify first, then decide, rather than pursuing perfect information.
7. "The discovery problem for podcasts is harder than for music — a 20-60 minute commitment vs. a 3-minute song" (Daniel Ek) — But solving it could lead to a 10-20x improvement in discovery rates, driving new usage periods and music genres.
8. "The gaming industry went from paid → free + ads → free + in-app purchases; the more open, the higher the revenue — the music industry will follow a similar path" (Daniel Ek) — Empowering creators is more profitable than controlling IP.