What this site is

Independence. Coverage decisions are made by the editorial desk alone. No entry on the wire, ranking on the league table, or line of analysis can be bought, and no covered party reviews copy before publication.

Sourcing. Every tracker entry links its sources. Figures the parties have not confirmed are marked "reported" and stay marked until confirmed. Every number on the site carries an as-of date.

Corrections. Errors are corrected in place with a dated note, and material corrections are flagged in the next Briefing. To request one: corrections@privateequities.ai.

Shen Pandi — Founder & Editorial Lead. Shen Pandi covers AI deployment economics in private equity — joint vehicles, portfolio inference costs, and operating systems for value-creation teams. He also builds AI cost-control tooling at o10.io.

Contributors and specialists are commissioned for sector notes and data work as needed; bylines appear on the pieces they author. Tips for the wire are welcome from practitioners across sponsors, portcos, and advisors.

Prefer citing a specific page with its URL and the as-of date shown on that page. For league standings, note that capital is attributed in full to participants (not pro-rata) and link the methodology below. For model unit prices, cite the pricing index as-of stamp — not an undated screenshot.

Example: PrivateEquities.AI, "AI deployment league table," accessed [date], https://privateequities.ai/league (figures as reported; see methodology).

Press and permissions: press@privateequities.ai.

The tracker records joint ventures, portfolio deployments, and fund commitments from primary announcements and reputable financial press. Entries update when facts change; each shows its last-updated date. Open the full tracker.

The league table attributes the full committed size of a vehicle to every participating firm — not pro-rata — because participation, not cheque split, is what the table measures. Read the footnote before citing.

The calculator and pricing index use published per-token rates with an as-of date on every figure. Comparisons assume eval-gated routing: a cheaper model must clear a per-workload quality floor. Outputs are estimates, not quotes.