Beyond the AI Hype: What Endures in Early-Stage Investing

In the first half of 2026, U.S. startups raised $412.7 billion, with AI capturing 86 cents of every dollar. That level of concentration is changing the market dramatically.

Median seed pre-money valuations reached a record $16 million last year. The numbers reflect a market where a relatively small number of companies can attract unheard of amounts of capital, often well ahead of meaningful revenue. At the same time, AI products can be built and brought to market more quickly than ever, making it harder to determine where durable advantage will come from.

That doesn't mean there are fewer great companies. It means the signals investors have traditionally relied on (product polish, fundraising momentum, prominent co-investors) can tell us less than they once did. Instead, it raises the bar on early-stage investing, getting back to the basics, and focusing on the things that can't be manufactured. At HearstLab, there are five things I keep coming back to:

1. The founder matters more, not less. When it becomes easier to build a compelling product demo, the founder becomes an even more important part of the investment case. I want to understand what gives a founder a unique view of the problem, how they make decisions with incomplete information, and whether they can recruit, sell, and adapt as the company changes. Those qualities are much harder to manufacture.

2. Industry expertise needs to inform the investment from the beginning. The question isn't simply whether the technology works. It is whether the product fits into the realities of how customers operate, buy, and allocate budget.

Buzz Solutions, a HearstLab portfolio company, sells AI-powered grid inspection and maintenance software to organizations including Dominion Energy, American Electric Power, and the New York Power Authority. Utilities are complex, regulated buyers with long sales cycles and significant operational requirements. Understanding that environment is essential to understanding the company's opportunity.

The difficulty of selling into those customers can also become part of the moat.

3. Capital needs to be differentiated. Last month we invested in Founderz, a Barcelona-based AI business school, and connected the team with Fitch Learning, a Hearst company. That partnership brings Founderz's AI training to more than 125,000 financial services learners.

More broadly, our founders can draw on a community of more than 220 Hearst Scouts and expertise across more than 360 Hearst businesses. That can translate into practical support with customer conversations, partnerships, product feedback, marketing, legal questions, and scaling a business.

In a market where capital itself is widely available to the companies attracting the most attention, strategic fit and access can matter as much as the check.

4. AI does not change the fundamentals. What problem is being solved? Who feels it most acutely? Who will pay for the solution? What evidence suggests the company can acquire and retain customers?

The technology may change, but the due diligence questions do not.

5. We need to price risk honestly. A high valuation does not reduce risk. Neither does a large round or an impressive set of investors.

Investing is inherently uncertain. Our job is not to eliminate that uncertainty, but to understand what we are underwriting, where conviction has been earned, what still needs to be proven, and what we cannot know.

HearstLab has backed more than nearly 100 women-led companies globally. In a market this concentrated, our edge is the ability to distinguish momentum from substance, develop conviction where it is warranted, and then be genuinely useful to the founders we back.