Start, Build, Scale: Anthropic's Tier Rename Is a FinOps Maturity Story

The Announcement

On July 2, 2026, Anthropic retired its deposit-based Tier 1-4 system ($5 / $40 / $200 / $400 cumulative spend thresholds) in favor of three usage-pattern-based tiers: Start, Build, and Scale, with monthly spend caps of $500, $1,000, and $200,000. Tier assignment moved from "how much have you deposited" to "how are you actually using the API."

This Isn't New — It's FinOps

The FinOps discipline solved this exact problem for cloud spend years ago. Every FinOps maturity framework the industry has produced treats cost management as a staged journey — crawl before you walk, walk before you run — not a flat function of dollars spent. Naming a tier "Small/Medium/Large" tells a buyer the wrong thing: that quantity is what matters. Naming it "Start/Build/Scale" tells them the right thing: that where you are in your adoption journey is what matters. Anthropic didn’t invent this idea. It rediscovered it — the same way every infrastructure category eventually does once flat pricing stops matching how customers actually mature.

PromptKing Was Built Around Maturity, Not Usage Buckets

PromptKing’s own maturity model has five stages — Blind, Aware, Optimising, Governing, Strategic — scored across four pillars: Visibility, Behaviour, Governance, and Value. An organisation that’s Blind (no visibility into AI spend or behaviour) needs a fundamentally different set of capabilities than one that’s Strategic (AI spend driving board-level ROI). That’s a different question than "how many seats" or "how much spend." This model existed in the product before Anthropic’s July 2 rename — the parallel isn’t a coincidence we’re claiming credit for, it’s evidence that the underlying shape of the problem is the same one FinOps has been solving all along.

Why This Matters for Buyers

If your AI governance tooling treats every seat the same regardless of maturity stage, you’re measuring the wrong thing. A five-person team just turning on Claude for the first time and a 500-seat engineering org running autonomous agents in production don’t have a "spend" problem in common — they have completely different governance problems. Vendors organising pricing around maturity stages instead of flat usage tiers is a signal that the market is catching up to what FinOps practitioners already knew: the stage you’re at determines what "good" looks like, not just what you’re spending.

THE PARALLEL

THE MODEL IN THE PRODUCT

The Maturity Score dashboard, scored across Visibility, Behaviour, Governance, and Value.

WANT YOUR SCORE?