NewsCryptoSolana Bets Big on AI Agents, Calling Agentic Payments Crypto's 'Biggest Opportunity'

Solana Bets Big on AI Agents, Calling Agentic Payments Crypto's 'Biggest Opportunity'

Author: Decrypt·

Key Takeaways

  • •The Solana Foundation launched Solana.ai and new agent-focused payments and privacy tools at Breakpoint, where it said more than 120 teams are building AI products on the network across payments, trading, and infrastructure.
  • •Automated software is becoming the internet's dominant user: bots accounted for over 53% of global web traffic in 2025, and agentic API keys made up 76% of tokens processed through OpenRouter in October 2026, up from 8% a year earlier.
  • •The x402 HTTP-based payments standard has processed more than 400 million payments and roughly $100 million in volume in a year, with Solana-based BlockRun described as the largest x402 server after handling over 60 million transactions in two months.
  • •The foundation has launched payment channels it says can handle 1 million payments per second, the Pay.sh server integrating more than 100 payable APIs, and is working with Google, Alibaba, and ByteDance to let them charge for AI models through agent payments.
  • •Real Vision founder Raoul Pal argued AI agents are not yet real economic actors, but said that as assets move onchain and agents use blockchains to trade and settle, value will accrue to the underlying layer one.
Solana Bets Big on AI Agents, Calling Agentic Payments Crypto's 'Biggest Opportunity'

AI agents are quickly becoming the internet's main users, and the Solana Foundation wants its blockchain to be how they pay for things.

"The biggest opportunity in crypto is already underway," Rishin Sharma, who leads AI at the Solana Foundation, said during a presentation at the Breakpoint conference in London. His remarks accompanied a slate of agent-focused launches from the foundation.

Sharma used his presentation to showcase Solana.ai, a new homepage for the network's AI ecosystem that the foundation launched last week. More than 120 teams are now building AI products on Solana across payments, trading, and infrastructure, he said. The foundation has also rolled out new payments and privacy tools alongside the hub. The underlying bet, per Sharma's framing, is that once software is the customer, the question of what money it can functionally use becomes as important as the models themselves.

Automated traffic takes the majority

Bots and other automated software already generate the majority of global web activity. According to cybersecurity firm Imperva's latest Bad Bot Report, which Sharma cited, bots accounted for more than 53% of global web traffic in 2025, up from 51% the year before.

Agents have also come to dominate AI usage itself. API keys classified as agentic accounted for 76% of the tokens processed through model marketplace OpenRouter in October 2026, up from 8% a year earlier, according to the company's monthly data recap — meaning software agents, not humans, now consume the bulk of AI model usage. For payment networks, that flips the traditional customer profile: the entities that will need accounts, balances, and spending authority are programs, not people.

At the same time, running AI is getting dramatically cheaper. Sharma's slides showed the price of a million tokens falling from $37.50 for OpenAI's GPT-4 to $0.14 for GPT-5 nano — a 270-fold drop — while monthly token use has risen around 330-fold.

"Which model is smarter matters a little bit less, and what matters more is what the model can actually do," Sharma said. "What are the tools it's using? What are the data it's able to access? And what money can it functionally use?"

Paying for the agentic internet

Payments are where that shift is showing up first, according to Sharma. The x402 standard, an HTTP-based payments protocol that lets AI agents pay for online services, has processed more than 400 million payments and roughly $100 million in volume, up from zero a year ago — a growth curve showing agent payments moving from concept to production within a single year. Solana-based BlockRun, which he described as the largest x402 server, has handled more than 60 million transactions in the past two months.

The foundation has also launched payment channels that it says can handle 1 million payments per second, a private AI service offering attested private inference on Solana, and Pay.sh, a server integrating more than 100 APIs that agents can pay for. Sharma added that the foundation is working with companies including Google, Alibaba, and ByteDance to let them charge for their AI models through agent payments — a sign that agent payments could eventually route through some of the largest platforms in consumer AI, not just independent services.

Next, the foundation plans to publish benchmarks measuring how well AI models perform financial tasks, and to launch an on-ramp that lets users convert cash into stablecoins and hand them to their AI assistants, he said. Both are worth watching as early measures of whether agentic payments move beyond conference showcases into routine use.

'They're like kids'

Macro investor and Real Vision founder Raoul Pal, speaking earlier at the conference, argued that agents are not yet real economic actors. "They're like kids," he said. "These are young kids who spend nothing on a few sweets and stuff."

But as stocks and other assets move onchain, he said, agents will start using blockchains to trade and settle, and "the value accrues to the underlying layer one."

Solana has been building toward that scenario for months. The foundation teamed up with Google Cloud on a stablecoin payments service for AI agents in May, while the x402 Foundation, backed by Coinbase and the Linux Foundation, is developing the payments standard. Asset manager Franklin Templeton has described agentic AI as a killer use case for crypto. With its Breakpoint launches, the foundation is working to ensure the payment rails are in place if that scenario arrives.