Brian Armstrong, the Chief Executive Officer at Coinbase, has publicly pushed back against the growing narrative that the artificial intelligence (AI) boom demands a complete shift away from cryptocurrency.
On Sunday, July 26, 2026, Armstrong took to Elon Musk’s X network to call the popular slogan “If you’re in crypto, pivot to AI” an example of “zero-sum, scarcity thinking.”
Coinbase’s vision for agentic finance
He argues that crypto is not a competitor but a foundational technology that actually becomes more crucial as AI advances, especially for what he terms “Agentic Finance.”
This firm stance from the head of the largest U.S.-based cryptocurrency exchange provides a counter-narrative to a visible trend within the crypto industry. Many Bitcoin mining operations are already repurposing their extensive computing infrastructure to serve the burgeoning AI sector.
Armstrong’s assertion reclaims a vision where AI and crypto are symbiotic, with blockchain-based systems providing the necessary financial rails for a future dominated by autonomous AI agents.
Armstrong believes that AI’s rise will make crypto more important, not less. He frames cryptocurrency as a general-purpose technology, akin to electricity or the internet, designed to underpin other innovations rather than compete with them. This perspective is central to Coinbase’s strategic focus on “Agentic Finance” (AiFi), a new sector where AI agents autonomously manage funds, execute transactions, and participate in commerce.
The core idea behind AiFi is that AI agents cannot effectively use traditional banking systems. They face limitations like Know Your Customer (KYC) checks, slow transaction times, and geographic restrictions. Instead, Armstrong envisions these agents relying on real-time, programmable money — a role perfectly suited for cryptocurrencies.
The x402 protocol and its growth
A key component of Coinbase’s AiFi strategy is the x402 protocol, which launched in May 2025. This payment system utilizes the HTTP 402 “Payment Required” status code, allowing software to pay for software seamlessly. It represents a practical application of programmable money for machine-to-machine interactions.
The protocol has seen significant adoption, processing over 100 million machine-to-machine transactions in approximately nine months by July 2026. By April 2026, x402 had already processed more than 165 million transactions across roughly 69,000 active agents, generating around $50 million in cumulative volume.
Its 30-day payment volume recently hit $24.24 million, and in May 2026 alone, it handled 3.1 million transactions worth about $1.2 million. This growth demonstrates a tangible demand for automated, crypto-enabled payments.
Base network and USDC integration
Coinbase’s Layer 2 network, Base, plays a pivotal role in enabling agentic finance. Over 90% of on-chain agentic stablecoin payments currently take place on Base, benefiting from its lower fees and faster settlement times. This highlights the practical utility of Layer 2 solutions for high-frequency, automated transactions.
The stablecoin USDC is also integral to the x402 protocol on Base. It provides AI agents with a stable unit for pricing and budgeting, mitigating the volatility often associated with other cryptocurrencies. This stability is crucial for AI systems that need predictable costs and consistent value in their financial operations.
Coinbase for Agents and marketplace innovations
To further facilitate AiFi, Coinbase introduced “Coinbase for Agents” in June 2026. This initiative provides developers with specialized tools that enable AI systems to trade, spend, and manage funds within predefined rules. It’s a direct response to the unique financial needs of autonomous agents.
Additionally, Coinbase launched Agentic.market in April, creating a dedicated marketplace for AI agents. Here, agents can discover and pay for essential services such as data, search capabilities, computing power, inference tools, and trading utilities. This ecosystem streamlines the financial interactions of AI, powered by crypto infrastructure.
Miners shift from crypto to AI compute
Despite Armstrong’s optimism, a significant segment of the cryptocurrency industry has already begun a pragmatic pivot towards AI. Many Bitcoin mining companies are retooling their operations, transforming their extensive data centers and energy infrastructure to support AI and high-performance computing workloads.
This shift is driven by evolving economic realities within the mining sector and the surging demand for AI infrastructure. For many, repurposing existing data centers and energy infrastructure to support high-performance computing workloads for AI presents a compelling financial incentive.
Core Scientific leads the pivot
Core Scientific, a major player in Bitcoin mining, is at the forefront of this transition. The company plans to completely exit Bitcoin mining by 2028, converting its facilities to focus on GPU colocation and AI workloads. This strategic move includes a substantial $3.3 billion capital raise in 2026, aimed at funding the construction of AI-focused data centers.
Core Scientific is expanding five new data center sites and actively pursuing long-term AI infrastructure contracts. They are developing powered infrastructure equipped with liquid cooling systems, specifically optimized for AI GPU cloud workloads. Initial deliveries for a high-performance computing customer are expected in the first half of 2025, underscoring their commitment to the AI market.
Grayscale ETF rebrands
The shift isn’t limited to individual companies; investment vehicles are also adapting. The Grayscale Bitcoin Miners ETF ($MNRS) is set to be renamed the Grayscale AI Compute ETF around September 15, 2026. This rebranding signals a clear investment strategy change.
The fund, originally launched in January 2025 to track Bitcoin mining firms, will now target businesses providing technology for AI development, including data centers and computing hardware manufacturers. Its underlying benchmark, the Indxx Bitcoin Miners Index, will similarly become the Indxx High Performance Computing Index, reflecting the broader industry trend.
Other major players follow suit
Beyond Core Scientific and Grayscale, several other prominent Bitcoin miners are making similar strategic adjustments. CleanSpark announced a significant pivot towards high-performance computing and AI infrastructure in October 2025. They’ve acquired land and secured power for dedicated AI data centers, advancing a multi-gigawatt AI infrastructure platform with 1.8 gigawatts under contract by May 2026.
DMG Blockchain Solutions is also accelerating its strategy to convert its Christina Lake facility into a 50-MW, liquid-cooled AI data center. The company recently signed a letter of intent for a 12-year colocation service agreement with a single tenant, with an initial capacity of 50 megawatts. Meanwhile, IREN secured a five-year, $3.4 billion cloud contract with NVIDIA for Blackwell GPUs, further solidifying the shift.
Differentiating infrastructure from application
Armstrong’s argument hinges on distinguishing between foundational infrastructure and specific applications built upon it. He views crypto as the underlying layer that enables new forms of financial interaction, much like the internet enables various digital services. This perspective suggests that while AI might be the “next big thing,” it ultimately relies on robust, open, and programmable financial systems that crypto can provide.
The comparison to electricity or the internet isn’t merely rhetorical. Just as these technologies became invisible yet indispensable enablers for countless innovations, Armstrong believes crypto will serve a similar role for AI. It’s not about AI replacing crypto, but rather AI creating an even greater need for crypto’s unique capabilities in financial transactions.
AI’s demand for programmable money
Autonomous AI agents, by their very nature, require financial systems that are instantaneous, globally accessible, and not bound by traditional human-centric banking hours or regulations. They can’t wait three days for a wire transfer, nor can they easily open a bank account that adheres to national jurisdictions. This is where crypto’s core properties, such as real-time settlement and borderless transactions, become indispensable.
These agents will need to pay for API calls, access datasets, lease computing power, and even trade autonomously without human oversight. Such a dynamic environment necessitates financial rails that are programmable, automated, and capable of handling micro-transactions at machine speed. Crypto, particularly stablecoins and Layer 2 solutions, is uniquely positioned to fulfill this critical requirement.
Broader implications and competitive landscape
Coinbase’s aggressive push into agentic finance suggests a long-term strategic bet on the convergence of AI and blockchain. If Armstrong’s vision materializes, the demand for crypto infrastructure will surge, transforming how AI systems interact with the global economy. This creates both immense opportunity and significant competitive challenges for the exchange.
While Coinbase aims for a first-mover advantage, established tech giants and financial institutions are also exploring similar solutions. Google, Visa, and Mastercard are reportedly developing their own agentic payment solutions, signaling a nascent but highly competitive market. Additionally, Ripple has rolled out a developer toolkit specifically for agentic payments on the XRP Ledger, indicating a broad industry awareness of this emerging sector.
The market’s reaction to the AI pivot has been immediate, with some Bitcoin mining companies seeing substantial stock gains in 2026 as they shift focus. Hut 8 Corp. (HUT) gained nearly 140% year-to-date, while KEEL and WULF also saw significant increases, in contrast to Bitcoin’s roughly 25% decline.
This divergence underscores the market’s enthusiasm for AI-driven growth, even as Brian Armstrong reminds the industry of crypto’s foundational importance.