POV · Series 02
Funding AI from Inside Your Existing Budget.
Most CFOs are now solving for the same equation: how to fund a multi-year AI program without growing the IT budget. The answer is in the building. The legacy database license. The data warehouse renewal. The integration tier that bills hours to keep itself alive. Each line is paying interest on systems built for a world that has ended. Each line is a candidate for reallocation.
The Argument
The AI budget is trapped, not missing.
The corporate finance question every IT budget cycle now confronts: where does the AI funding come from. Net-new budget is the hardest line to defend in a flat-or-down IT envelope. The companies that solve it through reallocation will fund the AI program their competitors are still asking for budget approval to start.
This series walks through the reallocation calculus : what to retire, what compounds, what the math looks like over a three-year horizon, and how to defend the moves to the audit committee, the board, and the operating team that depends on the systems being modernized.
Reading Order
Three essays, in sequence.
01
Why most AI initiatives get unfunded after year one.
The pattern: year-one budget approved on optimism, year-two budget cut because the AI program competes with the same legacy renewal it was supposed to displace. How to break the pattern with the budget you already have.
Read essay 01 →02
The legacy database cost-structure problem, in plain English.
Why a legacy database with eighteen years of stored procedures is not a cost line : it is compound interest on a foundation that no longer earns interest. The math that makes the reallocation argument inevitable.
Read essay 02 →03
Reallocating database and ERP spend to AI initiatives that compound.
The practical playbook. Which legacy lines reallocate first. How to time the reallocation against the renewal cycle. How to defend the move to the systems team that owns the legacy today.
Read essay 03 →“Most of the AI budget you think you need is already in the building. It is paying interest on the old stack.”
Where the Budget Goes
The reallocation calculus.
Where the budget is trapped today:
- •Annual license renewals on legacy database and ERP nobody loves
- •Pipelines and integrations that bill maintenance hours to stay alive
- •Multi-quarter projects that produce decks instead of production systems
- •Change orders against modernizations that never reached the production goal
Where the budget should reallocate:
- •A cloud-native data foundation that AI can actually read against
- •Application and database refactors that retire the debt permanently
- •Production agentic platforms that run against real data and real workloads
- •Forward-deployed engineering that ships rather than staffs
Related