05/23/2026
This is a great summary of the risk we face if we continue to lack perspective…
Meat Packers and AI Data Centers
Is anyone else seeing the parallels between what’s happening in rural America with meat packing and what’s now happening with AI infrastructure?
On one side, USDA Secretary Brooke Rollins and others are arguing that we need to break up concentrated meat packing and return processing capacity to local economies.
On the other side, many of the same voices — including Kevin O’Leary in the current Utah data center debate — are arguing that America needs massive, monolithic rural data centers to compete in the AI race.
So my question is: why?
Why is concentration dangerous in agriculture but suddenly necessary in computing?
China has roughly 300 major data centers. The United States already has more than 4,000. Yet the answer continues to be larger facilities, more concentration, and pushing infrastructure farther into rural landscapes.
AI is not magic. It is processing power applied to data. Not every AI problem requires hyperscale infrastructure and endless centralized computing. In many cases, smaller, specialized, distributed systems are more efficient, more economically sustainable, and capable of producing higher accuracy insights closer to the source of the data itself.
The parallels between meat packing and AI infrastructure are hard to ignore:
- centralized processing
- rural resource extraction
- few entities controlling outcomes
- many communities impacted
- heavy use of power, land, and water
- control of supply chains and data
We already understand what concentration did to food systems. Smaller regional systems strengthen local economies, reduce fragility, and keep accountability closer to the communities affected by them.
That philosophy applies to AI too.
And notice where many of these projects are being pushed — not dense urban environments with existing infrastructure, but rural regions with open land, water access, and lower infrastructure costs.
Meanwhile, old industrial buildings in cities could potentially be retrofitted into smaller distributed compute environments connected through modern networks and specialized processing layers.
Bigger is not always better.
If we do not rethink this now, we risk repeating the same mistake we made in food systems: concentrating processing, infrastructure, and control into fewer and fewer hands until local economies become dependent on systems they no longer control.
We are already fighting to regain localized food systems.
Do we really want to spend the next 30 years fighting to regain localized data systems too?
Same data, different meat.
