Notes on sovereign AI.
How regulated enterprises run agents inside their own jurisdiction, network, and keys — without giving up their intellectual property.
The Sovereignty Series
Five parts, weeklyEconomics, agenda, and open questions
The cost case for owning versus renting, a twelve month agenda, and an honest account of what we may be wrong about. Part V of the Sovereignty Series.
Read part 5The intelligence layer you actually own
What sits inside the boundary — context, memory, judgment and governance — and why Michael Baker built Titan. Part IV of the Sovereignty Series.
Read part 4What sovereignty is, and how to measure it
A working definition of AI sovereignty, and a maturity model for locating your enterprise on the path from rented to owned. Part III of the Sovereignty Series.
Read part 3Four forces that make rented intelligence fragile
Concentration, deprecation, repricing and drift: the four structural forces that turn a rented model layer into an operating risk. Part II of the Sovereignty Series.
Read part 2The rented intelligence layer
Enterprises quietly outsourced their capacity to reason and called it a subscription. Part I of the Sovereignty Series.
Read part 1