What to Know
- Coinbase CEO Brian Armstrong expects artificial intelligence agents to account for well over 99.9% of all transactions within a decade.
- Armstrong said AI agents may outnumber humans and could need crypto-based payment systems for small, real-time transactions.
- Coinbase is building financial accounts, payment tools and marketplaces designed to let AI agents pay one another for services.
- Early activity is taking place on Coinbase’s Base blockchain using USDC stablecoins, according to Armstrong’s market framing.
- Armstrong said he is already using AI agents to trade crypto with his own money, with some strategies up and others down.
- Coinbase is developing tools that could let users create trading strategies, publish results and allow others to copy successful approaches.
- Armstrong believes bitcoin has probably reached the bottom of this cycle and said prices could begin trending higher next year.
- He maintained his view that bitcoin could reach $400,000 by 2030.
- Prediction markets and derivatives now account for about 25% of Coinbase’s trading fee revenue.
Coinbase Bets on AI Agents as the Next Payments Frontier
Coinbase CEO Brian Armstrong is positioning artificial intelligence agents as a major force in the future of global transactions, arguing that autonomous software systems could soon handle the overwhelming majority of payments. Armstrong expects AI agents to account for well over 99.9% of all transactions within a decade, a view that places machine-to-machine commerce at the center of Coinbase’s next strategic push.
The idea is straightforward but potentially far-reaching: as AI agents become more capable, they may not simply answer questions or automate routine workflows. They could hire services, complete tasks, pay vendors, execute commercial instructions and interact with other agents across digital marketplaces. In Armstrong’s view, there will be more agents than humans, and those agents may conduct smaller transactions in real time, creating a payments environment very different from today’s consumer-driven card and banking networks.
Coinbase is betting that crypto infrastructure can support that environment. The company is working on financial accounts, payment tools and marketplaces that would allow AI agents to transact with one another without a human approving each individual payment. This is not just a technological experiment for Coinbase; it is also a potential revenue strategy as the exchange looks to diversify beyond its traditional reliance on trading fees tied to market cycles.
Why Crypto Rails May Suit Machine-to-Machine Payments
AI-driven transactions could be frequent, automated and small in size. That profile may prove challenging for conventional payment networks, particularly if fees, settlement times or account access limitations make tiny payments uneconomical. Crypto rails, including stablecoins, are often presented by market participants as a possible alternative because they can support programmable settlement, global transferability and continuous operation.
Armstrong pointed to a business using AI to create an advertisement as an example of how these systems could work. Rather than relying on a single model for an entire project, an AI agent could coordinate with separate services for voiceovers, video production and other functions, paying each provider as work is completed. In that model, the agent becomes both a project manager and a payments coordinator.
Coinbase’s Base blockchain and USDC stablecoins are central to the company’s early activity in this area. Stablecoins may be especially relevant because they allow users and developers to transact in digital assets designed to track traditional currency values, rather than exposing every transaction to the volatility associated with assets such as bitcoin. For machine-to-machine payments, predictability can matter as much as speed.
Revenue Diversification Becomes More Important
The AI payments push comes as Coinbase continues to seek additional business lines beyond spot crypto trading. Trading revenue has historically been sensitive to digital asset prices, investor sentiment and retail activity. When markets are active, exchanges can benefit from higher volumes. When markets slow, revenue pressure can increase.
Coinbase has already expanded into areas such as stablecoins, staking and custody, which can help offset weaker trading conditions. The company is also broadening its presence in derivatives and prediction markets, two areas that increasingly overlap with crypto-native finance, brokerage services and event-based trading platforms.
Armstrong said prediction markets and derivatives now account for about 25% of Coinbase’s trading fee revenue. That figure underlines how the company’s fee base is changing, even as it remains closely tied to broader investor engagement with digital assets. As brokerages, crypto exchanges and prediction market platforms move into one another’s territory, Coinbase is looking for product categories that can keep users active across different market environments.
Competition Is Intensifying Across Crypto and Brokerage Products
Coinbase is not operating in isolation. Robinhood has been expanding its crypto business while offering stocks, options and digital assets in a single platform. Exchanges such as Kraken and Binance remain important competitors, particularly in derivatives trading. Prediction markets have also drawn more attention as platforms such as Polymarket and Kalshi attract traders interested in wagering on outcomes tied to elections, economic data and other events.
That competitive backdrop helps explain why Coinbase is exploring AI payments. If autonomous agents become a major source of transaction volume, payment infrastructure could become a durable business line that is less dependent on whether human investors are actively buying or selling bitcoin on a given day. The opportunity remains early, and adoption is not guaranteed, but the strategic logic is clear: if AI agents are going to transact at scale, companies that provide the accounts, settlement tools and marketplaces could become central financial infrastructure providers.
Armstrong Is Testing AI Trading Agents
Armstrong also said he is already using AI agents to trade crypto with his own money, though the results have been mixed. He described having three or four agents running with different portfolios, with some up and others down. That candid assessment reflects both the appeal and the risk of automated trading tools in volatile crypto markets.
Automated strategies can process information quickly, operate continuously and follow rules without emotion. However, they can also amplify losses if models are poorly designed, if market conditions change or if permissions are too broad. In crypto, where prices can move sharply and liquidity conditions vary across venues, safeguards are especially important.
Coinbase is developing tools that would allow users to create trading strategies, publish their results and let others copy successful approaches. Armstrong emphasized the need for controls such as spending limits, account permissions and the ability to test strategies before committing real money. Those guardrails could become essential if AI-assisted trading shifts from a niche practice to a mainstream product feature.
Bitcoin Outlook Remains Constructive Despite Selloff
Armstrong also offered an upbeat view on bitcoin despite the recent selloff. He said he believes the market has probably seen the bottom of the current cycle and suggested prices could begin trending higher next year. He maintained his view that bitcoin could reach $400,000 by 2030, describing that target as reasonable.
The Coinbase chief linked part of the recent weakness to investors shifting capital toward major technology listings. He compared the pattern to an earlier period when investors rotated out of bitcoin ahead of a SpaceX IPO before later returning to crypto. Armstrong suggested that a potential Anthropic IPO, with order books planned to close at 4 p.m. ET on Tuesday, could be driving a similar shift in investor interest.
That interpretation remains a market view rather than a certainty. Bitcoin’s price can be influenced by liquidity conditions, investor positioning, regulatory developments, broader risk appetite and the strength of competing investment opportunities. Still, Armstrong’s comments show that Coinbase’s leadership continues to see bitcoin’s four-year market cycle as intact, even as the company diversifies into other business lines.
Stablecoins, Tokenized Stocks and Onchain Finance Remain in Focus
Beyond bitcoin, Armstrong pointed to growth in stablecoins, tokenized stocks and blockchain-based financial services as signs that crypto adoption is broadening. This matters for Coinbase because a wider onchain economy could generate activity even when speculative trading slows. Stablecoin transfers, tokenized assets and AI agent payments could all create transaction flows that differ from the boom-and-bust profile associated with retail crypto speculation.
For FXCOINZ readers, the key takeaway is that Coinbase is framing AI and crypto as converging technologies. AI agents may need payment rails, permissions, identity-like account structures and access to marketplaces. Crypto networks, especially those designed for programmable settlement and stablecoin use, could provide part of that infrastructure. The opportunity is still developing, but Coinbase is clearly treating it as a significant strategic priority.
The larger question is whether users, developers and institutions will trust autonomous agents to initiate payments and trading activity at scale. Security, compliance, fraud prevention and risk management will likely shape how quickly the concept advances. If the market evolves as Armstrong expects, the next decade of digital payments may be defined not only by people buying goods and services, but by software agents transacting continuously in the background.
Frequently Asked Questions (FAQs)
What did Brian Armstrong say about AI agents and transactions?
Brian Armstrong said he expects AI agents to account for well over 99.9% of all transactions within a decade, arguing that there will be more agents than humans and that many of their payments may be small and real time.
Why does Coinbase see crypto as useful for AI agent payments?
Coinbase sees crypto rails as potentially useful because AI agents may need to make frequent, automated payments that are too small or too continuous for traditional card networks to handle efficiently.
Which blockchain and stablecoin are involved in Coinbase’s early AI payments activity?
Armstrong said much of the early activity is taking place on Coinbase’s Base blockchain using USDC stablecoins, which can help support programmable payments with a more stable unit of value.
Is Coinbase moving beyond traditional crypto trading?
Yes. Coinbase is expanding into areas such as stablecoins, staking, custody, prediction markets, derivatives and AI payment infrastructure as it looks to reduce dependence on trading activity tied to crypto price cycles.
How much of Coinbase’s trading fee revenue comes from prediction markets and derivatives?
Armstrong said prediction markets and derivatives now account for about 25% of Coinbase’s trading fee revenue, showing that these product categories have become a meaningful part of the company’s business mix.
Is Armstrong using AI agents for crypto trading?
Yes. Armstrong said he is using three or four AI agents to trade crypto with his own money, and he described the results as mixed, with some portfolios up and others down.
What tools is Coinbase developing for AI-assisted trading?
Coinbase is developing tools that would let users create trading strategies, publish performance results and allow others to copy successful approaches, while adding safeguards such as spending limits, permissions and testing features.
What is Armstrong’s current view on bitcoin?
Armstrong said he believes bitcoin has probably reached the bottom of the current cycle and could begin trending higher next year, though that remains a market view rather than a guaranteed outcome.
What is Armstrong’s bitcoin price target for 2030?
Armstrong maintained his view that bitcoin could reach $400,000 by 2030, describing the target as reasonable while continuing to frame bitcoin’s longer-term cycle in constructive terms.
