Tokenary Policy: Companies as Nations
The ways an organization can govern its token supply, and what the history of monetary policy suggests about each (Series 2/3)
If tokens are the money supply of an AI-enabled organization, as our previous post argued, then every company is something like a small sovereign nation that must decide how to manage its own currency. That decision, whether explicit or not, is the organization’s tokenary policy. Most companies have one already; comparatively few have chosen it deliberately.
The Default Choice Is to Peg
The most common approach is also the least examined: pegging the organization’s tokenary policy directly to a major provider. The company mirrors the usage-based pricing, rate limits, and billing structure of a company like OpenAI, Anthropic, or Google, and treats those terms as the environment in which it operates.
Pegging is appealing because it is simple and requires no internal machinery. Its weakness is that it imports external volatility. When a provider adjusts its approach, changes its pricing, or tightens capacity, the organization’s costs move immediately and without warning. The parallel in monetary history is familiar: a country that pegs its currency to a larger economy’s gains stability and credibility, but surrenders control whenever the anchor moves. The defining feature of a peg is that the controlling rule lives outside the organization. You are governed by someone else’s policy.
Four Regimes, and Their Historical Parallels
Where the peg adopts an external rule, the four internal regimes sort along two simple axes: how freely tokens are spent, and whether a deliberate rule governs that spending. Each has a clear analog in how nations have managed money.
Loose Policy
Access is unlimited or lightly governed, and teams are encouraged to use AI wherever it might help. This mirrors expansionary, discretionary monetary policy - the easy-credit booms that produce explosive short-term experimentation and genuine productivity spikes, followed by over-expansion, budget blowouts, and painful correction. The abundance-driven overruns described in the previous post are the textbook result; an annual budget exhausted in a few months is the characteristic ending.
Tight Policy
Rather than open access, the organization imposes hard token budgets, real-time AI gateways, and chargebacks that hold teams accountable. This is the corporate version of a currency board: very low “inflation” and high predictability in exchange for almost all flexibility. Costs become predictable and overruns rare, but genuine opportunities can feel shackled when the rule is too rigid to bend.
Rules-Based Policy
The organization builds its own internal AI central bank - automation that routes simple tasks to cheap models, applies semantic caching and persisted memory, and shifts high-volume workloads to self-hosted or open-source models. The goal is a stable token “inflation” rate tied to measurable business outcomes rather than raw volume. This parallels modern inflation-targeting central banks: rules-based, sovereign over its own decisions, and the most durable regime for most organizations.
Political Policy
Allocation is governed not by any rule tied to value but by executive enthusiasm, internal influence, or vendor relationships. Spending optimizes for secondary effects - rewarding favorites, protecting prior decisions - while the stated objective and the real objective quietly diverge. This produces the highest volatility of any regime: costs that cannot be traced to outcomes and budgets that no one fully trusts.
Choosing Deliberately
Pegging is easy and exposes you to shocks you do not control. Loose policy buys speed now and a correction later. Tight policy buys predictability at the cost of agility. Political policy optimizes for everything except the value you are actually after. Rules-based policy asks the most of an organization and, in return, offers the most control.
This does not mean every company should immediately stand up a sophisticated internal token bank. It means the choice should be made on purpose rather than inherited by default. The strongest performers will most likely adopt an independent, rules-based tokenary policy, letting token supply grow predictably with the value it creates while still using the major providers for flexibility and self-hosting where volume justifies it. The essential discipline is to write the rules down, so they outlast any single executive or hype cycle.
Key Takeaway
Every organization already has a tokenary policy. The only question is whether or not it chose one. Borrow from how nations manage currency - balancing stability, sovereignty, and discipline against your own risk tolerance and growth ambitions - and write the rules down so they survive the next budget cycle.
Our next and final post turns the recommended path into practice, examining how to actually run an independent tokenary policy without it quietly collapsing back into discretion.
This series — Tokenary Policy:
Post 1 — Tokens as the Monetary Supply of the AI Revolution: why cheap tokens lead to bigger bills.
Post 2 — Companies as Nations (you are here)
Post 3 — Running an Independent Tokenary Policy: a practical playbook built on a basket of measures and a rule that holds under pressure. Coming next.