Sapiom и $15 млн: теперь нейросети начнут тратить ваши деньги

Imagine your virtual assistant not just writing code, but independently purchasing access to a cloud server, paying for APIs and renewing subscriptions to services. Until recently, this was a technical dead end: AI has no passport, no credit card, and it can't solve a CAPTCHA that says "I'm not a robot" (though ironically, it actually is a robot). The startup Sapiom decided it was time to hand algorithms a checkbook and raised $15 million for the endeavor with backing from heavyweight investors at Accel. This is an important step toward what we call the agent economy, where programs become full-fledged market participants.

We have smoothly entered the era of AI agents—programs that don't just answer questions but execute complex chains of actions. But the moment an agent steps beyond the developer's comfortable sandbox, it hits a brick wall. The modern internet was built by and for humans: everywhere you need logins, passwords, and two-factor authentication via SMS. If you ask an AI to book a hotel or purchase training data for a model, you'll still have to manually enter your credit card details. This kills the very idea of autonomy and forces us to act as a "gasket" between two digital interfaces.

Technically, Sapiom's solution looks like an abstraction layer between AI and the outside world. The startup takes on all the dirty work: authorization and processing microtransactions. Investors from Accel are betting that in the next couple of years, the volume of transactions conducted by machines will exceed human traffic several times over. This isn't just another payment processor; it's a complex trust system. If an agent suddenly decides to spend your entire monthly budget on dubious tokens or endless requests to an expensive model, the system must stop it. Sapiom promises flexible limits and controls so that digital freedom doesn't turn into a financial sinkhole.

Why is this important right now? Because the market for language models is oversaturated, and the competition among giants is shifting toward "action." OpenAI, Anthropic, and Google are competing to see whose bot best manages computers and browsers. But without the ability to legally pay, these bots remain locked in the cage of free demo versions. Sapiom is laying rails for a new economy in which the players won't just be people and companies, but autonomous chunks of code. This fundamentally changes how we consume software: why do you need a personal subscription to an expensive graphics editor if your agent can rent it for five minutes to complete one specific task?

Of course, this raises a huge range of security and ethics questions. Giving artificial intelligence direct access to money—that's a scenario that sounds like the beginning of a cyberpunk horror story to many. Yet the alternative is stagnation. If we truly want agents to save our time, we'll need to give them the right to manage resources. Sapiom is trying to make this process transparent and manageable. Likely, we'll soon see the emergence of an entire class of "banks for bots," and current investments are just the first payment toward the foundation of a new financial reality.

The bottom line: AI agents are ceasing to be just chatbots and are becoming economic subjects. If Sapiom solves the trust and authorization problem, we'll see explosive growth in autonomous services that operate completely without human involvement. Are you ready to trust your bank account to an algorithm if it promises to do all the work for you?