Skip to content

Topic

Usage-Based Pricing

A billing model that charges customers for actual consumption—such as API calls, compute, or credits—rather than a flat subscription or per-seat fee.

Current stories

investOne report1 publisher

Per-seat pricing for AI agents ranks last in G2's survey, with 9% of 1,019 buyers calling it fairest

G2's survey of 1,019 B2B buyers finds only 9% call per-seat pricing the fairest way to charge for AI agents, against 57% for pay-as-you-go. Vendors that kept agents inside seat plans go into 2027 renewals on the model buyers rank last, according to saastr.com's write-up.

Publishers:saastr.com

Reality

Evidence50
Adoption66
Hype gap+30
Incentives
Not scored
Confidence55
buildOne report1 publisher

Auditors reproduced 45-plus billing bugs across 110 AI usage tools, most often retries counted as new work

An audit of 110 AI usage tools reproduced more than 45 metering bugs, most often retries counted as new operations. The authors did not publish an overbilling rate, so teams paying per usage should check their own logs before trusting an invoice.

Publishers:dev.to

Reality

Evidence45
Adoption10
Hype gap+20
Incentives55
Confidence50
buildOne report1 publisher

Lovable now bills hosting and backend from the same credit balance that pays for prompts

Lovable has billed hosting and its built-in backend from the same credit balance as prompts since June 1, 2026. A dev.to guide reports that an empty balance pauses the backend and locks users out of the deployed app.

Publishers:dev.to

Reality

Evidence40
Adoption
Not scored
Hype gap+20
Incentives75
Confidence45
investOne report1 publisher

Finance chiefs in Pigment's CFO Index found usage-priced AI costing more than they expected

Pigment's Q3 2026 CFO Index found 83% of respondents saw consumption-based AI costs come in above expectations. Budget owners should expect central sign-off and milestone gates, though CloudZero's survey has 87% still planning to spend more.

Reality

Evidence42
Adoption50
Hype gap+20
Incentives
Not scored
Confidence40
investOne report1 publisher

Anthropic's operating loss per dollar of revenue fell from $7.70 to $1.75 in 2025

Anthropic's revenue rose nearly 12-fold to about $4.6 billion in 2025 while its operating loss grew 2.7-fold to $8.06 billion. Most of the $42 billion net loss came from accounting charges tied to financing, so the operating line is the one that shows whether the business is improving.

Publishers:fool.com

Reality

Evidence55
Adoption70
Hype gap+10
Incentives
Not scored
Confidence55
investOne report1 publisher

Median AI support agent resolves 48% of conversations on Gorgias's live-store test

Gorgias tested 13 AI support vendors on 212 live ecommerce stores and found the top five resolve about 70% of conversations, the median 48%. Resolution is the pricing unit, so the rate to contract on is one measured on the buyer's own traffic, repeatedly.

Publishers:saastr.com

Reality

Evidence52
Adoption55
Hype gap+18
Incentives82
Confidence50
productOne report1 publisher

Vayu launches a customer-by-customer margin view for AI vendors' finance teams

Vayu launched a per-customer margin tool for AI vendors, citing a survey in which 58% of finance leaders run hybrid pricing across three or more systems. For usage-priced teams, the useful part is the metering layer that ties each customer's usage and contract to what it costs to serve.

Publishers:thenextweb.com

Reality

Evidence30
Adoption20
Hype gap+35
Incentives75
Confidence40
buildOne report1 publisher

Copilot Pro's AI Credits reset at 00:00 UTC on the 1st, whatever day the subscriber pays

GitHub resets Copilot Pro's included AI Credits at 00:00 UTC on the 1st of each month, so paying mid-month does not refill a used-up allowance. Teams that forecast usage by invoice date are tracking a different clock from the one GitHub meters them on.

Publishers:dev.to

Reality

Evidence35
Adoption
Not scored
Hype gap+20
Incentives40
Confidence40
productConfirmed9 publishers

Microsoft moves Copilot's hand-off features onto usage billing in a rebuilt desktop app

Microsoft's rebuilt Copilot app bills Cowork, Code and Autopilot by usage while Chat stays on per-user licenses. Companies piloting it will pay in step with how much work staff hand to the agent, so the useful measure is tasks it finishes.

Perspective Coverage

9 publishers
Builder
Builder 27%
Operator
Operator 49%
Investor
Investor 24%

Reality

Evidence70
Adoption12
Hype gap+35
Incentives70
Confidence68
buildOne report1 publisher

A workflow's step count decides whether Make or n8n costs less to run

Make bills a credit for every module a scenario executes and n8n bills once per run, so a 20-step pipeline costs 20 units on Make and one on n8n. For short flows the two prices converge, and the choice turns on who maintains the flows and where the data may live.

Publishers:dev.to

Reality

Evidence55
Adoption
Not scored
Hype gap+5
Incentives35
Confidence55
productOne report1 publisher

a16z's 50 technical buyers preferred credits over tokens by nearly two to one

The firm's rule is to meter at the highest unit you can measure and defend, which sounds obvious until you try to instrument a completed unit of work and find your billing system only knows how to count tokens.

Publishers:a16z.com a16z.news

Reality

Evidence25
Adoption
Not scored
Hype gap+25
Incentives
Not scored
Confidence60
productConfirmed15 publishers

Microsoft caps the cloud hours of the 4% who stream Xbox games the most

The unlimited streaming bundled with Game Pass ends in November. Microsoft says the change touches 4% of subscribers because longer sessions cost it more, and it has not yet said what extra hours will cost.

Perspective Coverage

15 publishers
Builder
Builder 13%
Operator
Operator 59%
Investor
Investor 28%

Reality

Evidence78
Adoption20
Hype gap+5
Incentives65
Confidence72
investOne report1 publisher

GitLab's $16 million guidance raise excludes a Flex revenue headwind of up to $13 million

GitLab raised its FY27 revenue guide by $16 million at the midpoint after 24% billings growth, leaving out a Flex shift worth up to $13 million. Fast conversion would leave cash intact and push reported revenue back toward the old range.

Publishers:saastr.com

Reality

Evidence55
Adoption30
Hype gap+25
Incentives40
Confidence50
investOne report1 publisher

Stripe's payments data shows its top AI customers accelerating from 120% to 175% growth

Stripe says its top AI customers grew 175% in 2026, up from 120% a year earlier, with 48% of revenue earned outside their home markets. Whether that breaks the rule that B2B growth slows with scale depends on how Stripe picked the group.

Publishers:saastr.com

Reality

Evidence30
Adoption55
Hype gap+35
Incentives75
Confidence60
buildOne report1 publisher

Anthropic's usage-based Enterprise plan bills all tokens at API rates on top of the seat fee

The seat fee covers access and no tokens, so every prompt in Claude, Claude Code and Cowork bills to the organisation at API rates. The one control Anthropic's help page describes is an admin-set spend limit.

Publishers:support.anthropic.com

Reality

Evidence74
Adoption
Not scored
Hype gap+5
Incentives80
Confidence66
buildOne report1 publisher

Reading a JSON array as a string held every Facebook Page to three posts

A scraper asked Facebook for 40 posts a Page and got three, because a False isinstance test returned the same value as an exhausted feed. The unit fixtures had been hand-written to match the buggy parser.

Publishers:dev.to

Reality

Evidence48
Adoption
Not scored
Hype gap+15
Incentives85
Confidence42
buildOne report1 publisher

ElevenLabs' $29 receptionist tier answers one call at a time

Reception bundles the voice model, a phone number and a calendar write into a self-serve plan with 75 included minutes a month. That works out at about 39 cents a minute, the rate a home-built inbound agent has to beat.

Publishers:runtimewire.com

Reality

Evidence48
Adoption
Not scored
Hype gap+12
Incentives72
Confidence55
buildOne report1 publisher

H3 Max renders five seconds of 768p video in about three seconds, soundtrack included

The speed figure comes from a single dev.to write-up and describes generation, not delivery. The design decision underneath it is that frames and stereo audio leave in one pass, so the smallest thing you can retry is the whole clip.

Publishers:dev.to

Reality

Evidence32
Adoption34
Hype gap+35
Incentives80
Confidence45

Earlier coverage

  1. AI revenue headlines now rest on definitions the buyer has to ask for

    Leadership · September 8, 2026 · One report1 publisher

  2. Four Actors topped a cost table on RAM they never used

    Build · September 5, 2026 · One report1 publisher