It's 1995 again

In 1995 the internet was slow and mostly useless. Pages took a minute to load, nobody trusted it with a credit card, and the largest companies in the world treated it as a curiosity. A few people could see that it would touch everything. They were right, and it still took thirty years.

We think AI is at the same point, with two differences that matter.

It is not useless

The 1995 web could show you a page. A 2026 model can read a contract, write the reply, look at the photo, plan the week and book the flight. The technology arrived already capable of doing work, not just routing it. That compresses the timeline. Products that would have needed a decade of infrastructure to become useful are useful on day one.

Everyone already has the terminal

The internet needed a computer in every home and then a phone in every pocket. AI inherits both. A billion people can use a frontier model from a device they already own, over a connection they already pay for. Distribution, the thing that took the web twenty years, is already done.

What this means for founders

The companies that defined the internet were founded in its first five years, when the technology still looked unfinished and the incumbents were not paying attention. Amazon in 1994. Google in 1998. The same window is open now, and it is shorter.

We think the biggest changes will show up in how we work, how we live, and how we turn intentions into results. Those are not sectors. They are the whole day. The founders who win will earn trust, not just usage, because an AI product fails on trust long before it fails on capability.

What this means for investors

Being early is uncomfortable. It means backing companies that look small and strange, in a market that has not agreed on what matters yet. It also means the price of being right is lowest exactly now. That is why Zoha exists, why it is small, and why it invests at the first check.

The analogy will break in places we cannot see. Plenty of funds raised in 1999 lost money. But the shape is familiar, and we would rather be early and wrong about the details than late and right about the story.