Wharton professor and AI researcher Ethan Mollick posted something recently that stopped me mid-scroll.

He observed that VC investments typically take 5–8 years to exit. Which means almost every AI VC investment being made right now is essentially a bet against the vision that Anthropic, OpenAI, and Google have publicly laid out for where AI is heading.

Read that again. If the big labs are right about the timeline — and they have more information than anyone — then a significant portion of today's AI startup ecosystem will be obsolete before their investors see a return.

The practitioners and investors who responded to Mollick largely agreed: the startups most at risk are those betting on capability gaps — building thin layers on top of foundation models, assuming the models stay limited long enough to justify the investment. One commenter called them "melting ice cubes." Another compared them to companies that built for dial-up right before broadband arrived.

Why This Matters for Your EDO

You might be thinking: "That's a Silicon Valley problem. What does VC math have to do with my economic development organization?"

More than you'd expect. Because the same flawed logic driving bad AI bets at the macro level is driving bad AI investments at the organizational level — and I see it every week inside EDOs across the country.

Here's what I've observed: there are exactly three ways any organization can invest in AI. Most organizations jump straight to the first two. Almost none start with the third — which is the only one that makes the other two work.

The 3 Ways to Invest in AI

Why the Third Way Has to Come First

Here's the logic most organizations miss: the education you need to make smart decisions about products and projects is taught through the people and processes investment.

Once your team understands the AI fluency landscape and has infused AI into how your organization actually operates, they will be far better equipped to evaluate which products are worth buying and which projects are worth building.

Without that foundation, you're making decisions with third-grade AI literacy about technology that's moving at a graduate-school pace.

Think about what Mollick's observation actually says: the riskiest AI investments are the ones that assume the models will stay limited. The same is true inside your organization. If you build your AI strategy around a specific tool or a specific capability gap, you're making the same bet the melting-ice-cube startups are making.

But if you build your strategy around your people's ability to work alongside AI — that investment compounds over time regardless of which model wins.

The Real Bet Worth Making in Economic Development

No frontier model is coming to solve the nuanced challenges of rural workforce development, site selection, or municipal incentive strategy. The labs aren't building for your community. You are.

The question isn't whether your organization will adopt AI. That ship has sailed.

The question is whether you'll build the internal capacity to adopt it well — or whether you'll spend the next three years buying products that become obsolete and funding projects that never quite deliver.

Invest in your people first. Let that investment guide everything else.