Interview, Fireside Chat, Conference Presentation
Coinbase CEO's Top 3 Crypto Trends for 2026 + More from Davos!
- Coinbase anticipates completed US crypto market structure legislation in upcoming sessions, predicts five of the top 20 global banks will build on its infrastructure, and foresees financial institutions increasingly moving on-chain.
- Armstrong expects the Trump administration to maintain clear crypto rules, with the Genius Act defining 30-day US treasuries as the maximum asset duration, prohibiting interest classification on stablecoin rewards, and remaining immune to undoing despite potential banking group opposition.
- Stablecoin trends include Tether splitting into US-compliant and non-US versions, US treasuries not failing within 30 days, and Silicon Valley Bank becoming non-viable for 100% fund storage due to run risks.
- Major market trends predicted involve all assets coming on-chain for trading, rapid growth in prediction markets, explosive stablecoin payment expansion, and B2B cross-border payments remaining the top growth area over the next year.
- Coinbase Business expects a large customer backlog requiring staffing, while tokenization is predicted to democratize private market access for investments as low as $100 or $1,000, with top funds like BlackRock seeking to tokenize all products.
- Armstrong projects private companies may eventually go public entirely on-chain and that AI agents within the Coinbase app will teach financial literacy and analyze internal communications as strategic advisors.
- California is expected to face insolvency or bankruptcy within two to three years as the budget rises and services worsen, driving approximately 20% of billionaires to exit the state.
- US macroeconomic outlook suggests a 5.6% GDP growth is difficult to sustain, while inflation is expected to fall from the 2.8%–2.9% range to the 2% target.
- Regarding AI, Armstrong views job displacement as a positive transition to abundance, while Feldman anticipates massive AI compute demand in early stages and predicts China could trigger a "winner-take-all" scenario if the US falls behind in chip making.
- Feldman states US AI policy should focus on empowering allies and establishing uniform federal laws, noting that exporting H100 chips to China remains a complex issue and that Canada may view US tariffs as unreliable.
- Cerebras Systems is predicted to deliver a blisteringly fast wafer-scale engine, with deep research tasks dropping from five minutes to 4–10 seconds, driven primarily by latency reduction and a shift to power-based data center capacity measurement.
- Energy infrastructure predictions include natural gas as the cheapest power source after hydro, water cooling efficiency in closed loops, and small modular nuclear reactors becoming necessary only after a 3–4 year delay.
- Data center timelines estimate space-based facilities are 8–10 years away, while OpenAI has reportedly ordered 750 megawatts of power from Cerebras.
- Feldman predicts job displacement in middle management will be a delayed impact of SaaS tools, while entire job categories will become vastly more efficient, and notes Chinese EV manufacturers like BYD may capture significant European share via subsidies.
- Lucerarian predicts robotics will shift from consumer tasks to high-ROI industrial applications, with one welder supervising 10 robots to reduce hazardous work hours in environments like deep-sea welding.
- Industry adoption timelines include robotics companies dominating ROI extraction within five years, and the ability to train humanoid robots on complex tasks like bricklaying without specific training expected in three years.
- Data requirements for robotics involve collecting proprietary datasets via sensors for specialized tasks like welding, though future models are expected to eventually learn from internet video libraries.