From Answers to Action: Where AI Is Headed and Why It Matters to Economic Developers
Every so often I like to step back and ask a bigger question: where is all this going? In more than 40 years of economic development I have seen many technology cycles, and this one is moving faster than any of them.
From prompt-driven to goal-directed
Last month I wrote that generative AI answers and an agent acts. The first wave of AI was predictive, finding patterns that signal what comes next. That is what Whittaker Associates has done for decades: spotting the signals that a company is likely to expand or move.
Generative AI added the ability to create. Now reasoning models add the ability to think a problem through, weighing options step by step before they respond. Agents put that reasoning to work toward a goal. We are moving from tools where you ask and it answers to systems where you describe the outcome and the AI figures out the steps.
The intelligence factory
All that reasoning must happen somewhere. Nvidia CEO Jensen Huang calls today’s data centers AI factories: electricity goes in, and tokens, the basic units of AI output, come out. Reasoning and agents consume far more tokens than a simple chat, which drives enormous demand for chips. J.P. Morgan estimates that memory chip (DRAM) prices will have risen more than 400 percent from the start of 2024 to the end of 2026. If you have priced a new laptop lately, now you know why.
For economic developers, the factory metaphor is literal. These are industrial projects that need land, power, water, and fiber, and communities are competing hard for them. It is one reason I am spending so much of my own time on data center site selection.
When the cost of intelligence goes to zero
Here is the paradox. Even as AI spending soars, the cost of each unit of intelligence is collapsing. Andreessen Horowitz found that, at equivalent capability, the cost of running a model has fallen roughly tenfold a year.
We have seen this movie before. When I started my career, a long-distance call was priced by the minute. When the cost of communication went to zero, ideas flowed and innovation followed. Intelligence is next. Analytical horsepower once reserved for large consulting firms and big-city EDOs will be within reach of a two-person office in a rural county. That is a real equalizer.
The shadow side
Cheap intelligence serves bad actors, too. In November 2025, Anthropic reported disrupting what it called the first AI-orchestrated cyber espionage campaign. A group it assessed as Chinese state-sponsored used AI to do 80 to 90 percent of the work against roughly 30 organizations. The manufacturers, utilities, and public agencies in our communities are targets.
Then there is artificial general intelligence, or AGI: AI that can match people across most kinds of thinking work. Experts disagree sharply about whether and when it will arrive. I would rather plan for a faster-than-expected future than be caught flat-footed.
What this means for us
- Delegate goals, not just prompts. Start practicing with agents on real, repeatable work.
- Treat power and data infrastructure as economic development infrastructure. The communities that can deliver megawatts and fiber will win the projects of this decade.
- Put cybersecurity on your BRE agenda. Ask your existing employers how they are preparing. Many small manufacturers have not started.
The technology is moving from answering our questions to acting on our behalf. Our job is to make sure our communities, and our own organizations, are ready to act too.