AI agents could change the internet faster than the market expects. Cardano founder Charles Hoskinson said at Consensus Miami 2026 that by 2035, most search queries, purchases, and online actions will be performed not by people, but by autonomous programs. For Google, Amazon, and Facebook, this threatens the very foundation of their business.
The problem is not the technology itself. AI agents do not behave like regular users: they do not click on ads, are not attached to brands, and choose actions based on efficiency, not marketing. That is why the advertising model of major platforms could be at risk.
AI Agents Are Changing the Internet Economy
Hoskinson believes the internet is shifting to a new usage model. Today, services are optimized for humans, their attention, emotions, and habits. In a few years, an increasing share of actions will be performed by agents that search, compare, buy, and conduct operations automatically.
For large platforms, this changes the rules of the game. Google earns from search advertising, Amazon from product promotion within its marketplace, and Facebook from selling users’ attention to advertisers. An AI agent does not need this logic.
It does not choose a product because of brand recognition. It compares terms, price, delivery, fees, and reliability. This reduces the value of advertising and increases the role of direct machine logic.
Big Tech Fears Losing Control of Attention
According to Hoskinson, the largest technology companies already understand the scale of the threat. They are actively investing in agent systems because the old model may stop working.
If people stop being the main participants in the internet economy, the advertising market loses its usual point of application. You can show a user a banner, a recommendation, or a sponsored result. For an agent, this is almost useless.
Such a shift could hit companies that have built their business on managing user attention for decades. In the new model, value shifts to infrastructure, data, payments, and trust.
Cryptocurrencies Gain a New Use Case
Hoskinson considers AI agents one of the strongest drivers for the crypto market. The reason is simple: agents need fast, programmable, and global payments.
Traditional banking infrastructure is poorly suited for autonomous micropayments between programs. Cryptocurrencies and stablecoins offer a more flexible model, where an agent can pay for access to data, APIs, computing, or financial operations without human intervention.
That is why Hoskinson mentioned Google’s interest in x402. This protocol, supported by Coinbase, allows applications and AI agents to make direct payments over the internet using stablecoins and crypto infrastructure.
AI Agents Could Take Over DeFi Operations
The next stage is tied to decentralized finance. Hoskinson expects AI to analyze projects, check risks, send transactions, and interact with DeFi protocols.
This could dramatically simplify the user experience. Today, a person needs to understand wallets, networks, fees, bridges, and smart contract risks. An agent could take over most technical actions.
For DeFi, this is a major turning point. If interaction becomes easier, the market will have a chance to reach beyond the narrow audience of advanced users.
Simplification Should Not Deprive Users of Control
At the same time, Hoskinson warned of the main risk. Convenience should not lead to transferring control to intermediaries.
He recalled the basic principle of cryptocurrencies: users should own their data, identity, and money. If everything is again handed over to custodial wallets, closed networks, and third parties, the industry returns to the model it tried to leave behind.
This thesis becomes especially important in the era of AI agents. If the agent manages the user’s assets, the issue of control, permissions, and security comes to the forefront.
Blockchain Fragmentation Hinders Development
Hoskinson also criticized the fragmentation of crypto ecosystems. According to him, about 11 million tokens have been issued over the years, and the market no longer needs new assets for the sake of assets.
The main problem is the lack of consistency. Users and liquidity are split between networks, applications, and tokens, which complicates infrastructure development.
Mass adoption requires compatibility. The less users have to think about which network they are on, the higher the chance the technology will become part of everyday financial experience.
User Experience Remains a Weak Point
Even in 2026, entering cryptocurrencies is still too complicated. A user must create a wallet, save a secret phrase, choose a network, buy an asset, send a transaction, and not make a mistake in the address.
For a broad audience, this is unacceptable. One wrong step can lead to loss of funds, and support is often lacking.
Hoskinson believes that account abstraction and network abstraction will help solve the problem. These technologies should hide technical complexity while preserving user control over assets.
Financial Institutions Are Changing Their Attitude Toward Blockchain
Separately, Hoskinson noted the change in position of major banks. In the past, JPMorgan restricted the activities of clients associated with cryptocurrencies, but now it is developing its own blockchain products.
This shows the market’s maturity. Technology that was once seen as a threat is gradually becoming part of the financial infrastructure.
For the industry, this is an important signal. Large institutions can no longer ignore blockchain, especially as stablecoins, tokenized assets, and payments for AI agents appear on the market.
What This Means for the Market
Hoskinson’s forecast shows that cryptocurrencies could gain a new role. They are becoming not only an asset for investors, but also a payment infrastructure for the autonomous internet.
If AI agents start conducting millions of operations, they will need a neutral and programmable settlement environment. Here, blockchains could gain an advantage over traditional payment systems.
However, the industry needs to solve old problems for this to happen. Simplicity, security, compatibility, and user control remain key requirements.
What’s Next?
By 2035, the internet may become less human in the usual sense. Most actions will be performed by programs, and people will set goals and control results. For Big Tech, this threatens the advertising model. For the crypto market, it is a chance to become the settlement layer for the new economy.
The main question now is implementation. If blockchains can simplify access and preserve direct user control over assets, AI agents could become one of the strongest use cases for cryptocurrencies.
Read More: The White House Wants to Pass the CLARITY Act by July 4, the Market Awaits the Vote
