The internet we have known for the last 30 years is at the conclusion of its evolutionary cycle. Since the inception of the World Wide Web, human beings have been the dominant navigators of the digital landscape.
We’re the ones firing up browsers, typing search terms into engines, skimming through social media feeds, analysing things on e-commerce sites, and finally, clicking on the “buy” button. The entire architecture of the modern internet has been carefully constructed around human psychology, to grab, hold and monetize human attention . But this human-centric web is coming to an end, according to a recent paradigm-shifting prediction from Charles Hoskinson, founder and CEO of Input Output, the development organisation behind the Cardano blockchain.
Hoskinson’s prediction, made in early May 2026, was stark: in the next decade, by 2035, the vast bulk of activity on the internet will no longer be carried out by human humans. Instead the web will be all about autonomous artificial intelligence agents. AI agents will be our relentless, hyper-efficient digital proxies, handling everything from basic information retrieval and intricate research to trading, interacting with decentralised finance, and conducting everyday digital commerce. This is not a routine software upgrade. This is a fundamental disruption that threatens to utterly upend the underlying business models of the world’s most powerful technology corporations while at the same time unleashing a new era of machine-to-machine economy fueled by cryptocurrencies.
The Existential Threat to the Attention Economy
To get a sense of how big Hoskinson’s projection really is, you have to look at how trillion dollar companies like Google, Meta (Facebook) and Amazon make their money. These companies have created a level of wealth never before seen in the “attention economy. Their main source of income is advertising, and specifically the ability to deliver highly targeted ads to human eyeballs. They play on our psychological vulnerabilities, our impulse buying and our susceptibility to brand marketing.
AI agents, however, are not human psychology. An AI used as a personal shopping assistant or data aggregator is not influenced by a bright banner ad, a creative marketing phrase, or the emotional appeal of a brand ambassador. If you tell your AI agent to find the cheapest, best, and most durable pair of running shoes, the agent will immediately search international inventory databases, compare material durability data, assess past pricing trends, and make the purchase without ever viewing a sponsored search result.
That’s a terrible truth for Big Tech. The traditional cost-per-click (CPC) and cost-per-mille (CPM) advertising strategies instantly break down if the searchers and browsers are machines. This impending “agentic revolution,” as Hoskinson expressly pointed out, is something that Amazon, Google and Meta are well aware of and quite concerned about. Now they’re in this huge arms race of capital expenditure aggressively investing in AI, not only to be part of the future, but as a defensive survival mechanism as their old ad-driven cash flows are in terminal decline. For financial analysts building models of the long-term valuations of these tech equities, the change marks a major inflection moment. They will have to devise completely new measures for revenue growth.
The Infrastructure Pivot and the Machine to Machine Commerce
The emergence of the agentic web is not merely a matter of creative software, but requires a major revamp of both digital protocols and the physical computing infrastructure. The physical manifestation of this change is already under way. Huge computational facilities and infrastructure providers, many of which had previously devoted their capacity to pure cryptographic mining, are aggressively pivoting their hardware and power resources to the specialised high-density data centres needed for continuous AI training and inference. The physical environment is physically being rewired to provide the compute power needed to keep billions of autonomous digital agents running.
At the same time, the digital protocols that govern how these agents interact are also being forged. “A major focal point is Google’s active interest in protocols like x402,” says Hoskinson. It’s a protocol for machine-to-machine commerce. If an AI agent wants to access a paywalled database, call a third-party API, or purchase a digital asset on behalf of a user, it can’t whip out a physical credit card. Traditional banking rails are too slow, too friction-filled, and can’t keep pace with the micro-second execution rates of artificial intelligence.
This is where the crypto space, and stablecoins in particular, come in as the native financial layer of the new internet. Agents need programmatic money – digital assets that can be streamed, validated and settled quickly on decentralised ledgers. The x402 protocol makes it easy for agents to pay for API calls and digital services using stablecoins, bridging the gap between autonomous AI and decentralised finance. This perfectly verifies the long-standing notion that blockchain technology is the ultimate settlement layer for the digital age, operating entirely independent of sluggish legacy financial infrastructure.
Reimagining Asset Management & Decentralised Finance
AI agents will be the main actors on the internet, and their impact on the financial industry, especially the cryptocurrency and decentralised finance (DeFi) spaces, will be revolutionary. “Deep-dive due diligence, smart contract auditing and the execution of high-frequency trading strategies are all complex, time-consuming tasks that will be almost entirely taken over by agents in the crypto ecosystem,” Hoskinson says.
This sentiment reverberates strongly among the industry’s leaders. The CEO of Coinbase, Brian Armstrong, has said that the volume of AI agents making on-chain transactions will soon surpass the volume of human transactions. Similarly, Binance founder Changpeng Zhao has said that computers don’t sleep, don’t hesitate and aren’t subject to emotional trading biases, meaning AI agents could one day drive transaction volumes millions of times larger than those currently generated by human traders.
In this reality to come, the human investor is less an active player and more a strategic director. Rather of reading whitepapers, analysing macroeconomic indicators or watching the correlation between commodities prices and equities tickers, an individual will simply set their risk tolerance and investment parameters. The AI agent will then deploy capital across several DeFi protocols, quickly reacting to global news events, yield swings and liquidity shifts, rebalancing portfolios with a level of precision and speed that is statistically impossible for a human being to achieve.
The Unqualified Necessity of Digital Sovereignty
The appeal of the convenience and efficiency of an AI-driven internet is undeniable, but Hoskinson gave a grave and sobering warning about the concentration of power. If we humans are going to give up our digital autonomy, our financial decisions and our daily contacts on the internet to AI agents, then who owns those agents and the data they consume is the single most important problem of the 21st century.
Absolutely, Hoskinson emphasised, consumers need to own their own data, their digital identities and their money. If a person uses an AI agent from a centralised Big Tech company, if that AI agent runs on a closed permissioned network and if the AI agent uses a custodial wallet, then the user is at the complete mercy of the company. The corporation might change the agent’s algorithm so it promotes its own products, prevent the agent from accessing competitor services, or lock the user’s digital assets without due process.
This is exactly why decentralised blockchain networks are the required antidote to the AI revolution. And true self custody of digital assets, and cryptographic ownership of one’s identity, means that an AI agent will operate only for the user, and not some corporate dictator. Traditional finance is finally coming to the realisation that blockchain is the only secure framework for this future. Hoskinson was particularly eager to point out the hypocrisy of traditional institutions such as JPMorgan, which spent years shutting down the bank accounts of crypto users and denigrating the technology, only to now go all-in on rolling out their own blockchain projects. They also see that the architecture of money and data is undergoing significant change.
In the end, the move to an internet ruled by AI agents by the year 2035 is not some nightmarish dream, but an unavoidable advancement in technology. It will kill the old business models based on human attention and create a hyper-efficient global programmed economy. But the real war of the next decade will be not on who takes over the internet with AI, but on who controls the AI. As we go farther into this agentic age, the principles of decentralisation, self-custody and cryptographic truth will be the only shields that preserve human sovereignty in a machine-driven world.
Read Also: How a Simple Trick Fooled an AI Into Handing Over $150,000 in Crypto
