Coinbase CEO's Top 3 Crypto Trends for 2026 + More from Davos!

Episode Summary

Executive Summary: The transcript spans two Davos interviews: Brian Armstrong discusses crypto regulation, stablecoins, banking partnerships, tokenization, and California’s business climate; Andrew Feldman of Cerebras explains why AI demand is driving massive compute, power, and data-center buildout, plus the geopolitical, energy, and infrastructure implications. Both argue that technology, deregulation, and clear rules are accelerating major shifts in finance, AI, and robotics.

Main Topics: Crypto regulation and the Trump-era policy shift (Priority: 5/5): Brian Armstrong argues the Biden administration tried to suppress crypto, while Trump has embraced it and pushed clear rules, especially for stablecoins and market structure. Stablecoins, banking partnerships, and financial infrastructure (Priority: 5/5): Coinbase is partnering with major banks and financial institutions to build crypto infrastructure, while stablecoins are becoming a regulated, treasury-backed financial rail. Tokenization and on-chain capital formation (Priority: 5/5): Armstrong predicts equities, private funds, REITs, and other assets will move on-chain, making capital formation and investing more accessible and liquid. AI compute, inference, and power constraints (Priority: 5/5): Feldman explains that Cerebras is built for faster AI inference and training, where latency, memory, and power are the main bottlenecks driving demand. Energy, data centers, and infrastructure buildout (Priority: 4/5): The discussion covers hydro, natural gas, water cooling, nuclear, grid upgrades, and even space-based data centers as possible answers to AI’s power needs. Robotics, defense, and the industrial economy (Priority: 4/5): Gecko Robotics’ founder argues robots plus AI will transform inspection, manufacturing, defense, and maintenance of the physical world, not just digital work. AI, job displacement, and organizational change (Priority: 4/5): Both speakers expect displacement, especially of middle management and repetitive knowledge work, but frame it as a productivity and abundance transition rather than pure loss.

Key Arguments: Crypto is now moving from regulatory gray area to structured legal status, and clear U.S. rules will help domestic firms and consumers. Stablecoins under the Genius Act are backed 100% by short-term U.S. Treasuries, making them safer and more transparent than legacy bank deposits in some respects. Banks are not just competitors; many are becoming partners by integrating Coinbase infrastructure and moving financial services on-chain. Tokenization will democratize access to private markets, funds, and real-world assets by reducing settlement friction and lowering barriers to entry. AI adoption is constrained less by model quality than by inference speed, memory availability, and access to cheap power. Data-center economics are now determined by electricity, not square footage, which makes location and grid policy central strategic issues. Robotics becomes far more valuable when combined with AI and domain-specific sensor data from industrial environments like refineries, bridges, and submarines. AI and crypto will converge because AI agents will need wallets and payment rails to transact autonomously. Job displacement is already visible in organizational flattening, but the larger AI-driven labor shift is still ahead. Geopolitically, the U.S. should support allies with chips and technology or risk ceding infrastructure influence to China. The future of work and wealth creation depends on expanding access to high-quality assets, tools, and productivity-enhancing systems. Nuclear, storage, and grid modernization are necessary for the next phase of data-center expansion, though not all solutions will arrive immediately.

Data Points: Coinbase bank integrations: 5 of the top 20 global banks - Armstrong said five of the top 20 global banks are using Coinbase infrastructure. Users of crypto globally: 500 million - Armstrong cited this figure to show crypto’s scale and legitimacy. Americans who have used crypto: 52 million - Armstrong used this to argue crypto is a major political constituency. Stablecoin reserve requirement: 100% - Under the Genius Act, U.S.-regulated stablecoins must hold full reserves. Treasury maturity cap: About 30 days - Armstrong said reserves must be in short-term U.S. Treasuries, roughly 30-day maximum. B2B growth area: Cross-border payments - Armstrong said this was the biggest growth area over the last year for stablecoin usage. Gecko defense share of business: 30% - Gecko Robotics says roughly 30% of business is defense-related. Reduction in manufacturing time: Up to 90% improvements - Gecko referenced Admiral Houston’s comments on speed gains using its technology. NVIDIA/AI-related deal power: 750 megawatts - Feldman said the OpenAI-related cloud deal announced publicly was 750 MW. System cost: About $1 million to $1.5 million - Feldman gave the on-prem system price range for Cerebras deployments. Token pricing: $0.50 to several dollars per million tokens - He described cloud rental pricing based on model and usage. Company scale: 4 trillion transistor part - Feldman described Cerebras’ wafer-scale engine as a 4T-transistor chip. Chip size comparison: 56 times larger than a B200 - He said the wafer-scale engine is roughly 56x larger than an NVIDIA B200. Memory market digestion period: About 18 months - Feldman said supply/demand signals will take around 18 months to normalize. AI data-center cooling: Closed-loop water cooling - He said Cerebras uses water cooling in a recycled closed-loop system. China AI/industrial context: Open-model leadership - Feldman said China is ahead in open model categories while the U.S. leads in chips. Robotics future horizon: About 3 years - Gecko’s founder estimated some bricklaying/robotic task capability could arrive in roughly three years.

Pivotal Quotes: "The Biden administration really tried to unlawfully kill this industry in America, from my point of view." — Brian Armstrong: Armstrong describing the shift from Biden-era crypto regulation to the Trump administration's friendlier stance. "The biggest trends happening in crypto right now is: number one, it's the everything exchange." — Brian Armstrong: Armstrong on the future of tokenized trading, prediction markets, and on-chain assets. "We built this part so you can get all those answers back in four seconds. In 10 seconds." — Andrew Feldman: Feldman explaining Cerebras’ advantage in ultra-fast AI inference and deep research workflows.

Implications: Crypto is moving toward regulated mainstream finance, while AI is becoming a power-and-infrastructure race. The winners will be firms that combine compliance, speed, energy access, and real-world data to build scalable platforms.