AI cost and market pay by role
Compare modeled AI run cost with market pay.
Browse representative role estimates, task classifications, source notes, and confidence on core estimates.
Each row uses the same representative Tier 2 and mid-experience inputs so roles can be compared. Build a Wagecard with your own tasks, hours, location, and experience for a personal estimate.
Company spending and disclosed AI revenue
Compare reported investment with reported revenue.
Read the figures with their accounting bases before comparing them.
Hyperscaler capex, 2025
Amazon, Alphabet, Microsoft, and Meta filings for FY2025
AI revenue they disclose
Two of the four do not report an AI-specific revenue line
Nvidia data-center revenue
Shown beside $6B of Nvidia's own FY2026 capital expenditure
Open the source ledger and accounting notes
Public hyperscalers: capital expenditure and disclosed AI revenue
Amazon
FY2025 capex (company-wide)
AWS AI revenue run-rate, Q1 2026
Amazon FY2025 capex & AWS AI run-rate (Yahoo Finance) (2026-04-09)
Alphabet (Google)
FY2025 capex
No AI-specific revenue line
Alphabet Q4 & FY2025 earnings release (2026-02-04)
Microsoft
Only company in this group reporting both AI revenue and AI capital expenditureFY2025 AI data-center capex
AI business run-rate, Q3 FY2026
Microsoft FY2025 AI data-center capex (CNBC) (2025-01-03) · Microsoft Q3 FY2026 newsroom (AI run-rate $37B) (2026-04-29)
Meta
FY2025 capex
No AI-specific revenue line
Meta Q4 2025 results (8-K, exhibit 99.1) (2026-01-28)
Nvidia
AI infrastructure supplierFY2026 capex (own property & equipment)
FY2026 data-center revenue
Nvidia Q4 & FY2026 earnings release (SEC) (2026-02-25)
Private AI labs: capital raised and reported revenue run-rate
OpenAI
Capital raised, Mar 2026 round
Revenue run-rate, Feb 2026
OpenAI $122B round (Bloomberg via Yahoo Finance) (2026-04-02) · OpenAI ~$25B run-rate (The Information via Yahoo Finance) (2026-02-15)
Anthropic
Series H, May 2026
Revenue run-rate, May 2026 (self-reported)
Anthropic Series H announcement ($47B run-rate) (2026-05-28)
xAI
Series E, Jan 2026
FY2025 revenue (consolidated, filed)
xAI Series E announcement (2026-01-06) · xAI FY2025 revenue, SpaceX S-1 (via Yahoo Finance) (2026-05-21)
How these numbers work
Figures come from company filings, official earnings releases, and named financial reporting available by June 2026. Sources appear in each row. The figures are not like-for-like: capital expenditure is cash committed during a fiscal year; a revenue run-rate is an annualized figure reported at a point in time; and capital raised is funding, not spending. We do not subtract one from another. If a company does not report AI-specific revenue, the table says “Not disclosed.” Figures tagged est.come from press reports or company statements rather than a filing. The layout was inspired by isaiprofitable.com; Wagecore compiled the figures and linked the sources above.
What AI run cost includes
The token bill is only part of the cost.
Common shortcut
Estimate AI cost from API usage alone.
Wagecore includes
Across these 20 representative roles, raw tokens account for 0% of modeled AI run cost. Review, retries, expected error cost, and integration make up the rest. Average exposure is 19/100, and 54% of hours are classified as human-led and AI-assisted or human-critical.
Across the published roles
See how the selected tasks are classified.
- 20
- Roles costed
- 19 / 100
- Average exposure estimate
- 4%
- Replaceable task share
Each at a representative Tier 2 / mid cell.
Lower means less of the representative task mix is economical to run with AI. 54% of hours are human-led and AI-assisted or human-critical.
Share of task hours across the representative roles in this tracker.
Average task-hour distribution
Role by role
Compare modeled AI run cost by role
Sort the representative estimates by exposure, market pay, or the difference between pay and modeled AI run cost.
Exposure is a 0–100 summary of the representative task mix. Lower means less of that work is economical to run with AI under the modeled assumptions. AI run cost includes tokens, review, retries, expected error cost, and integration for AI-handled hours. It is not a savings figure.
Computed with capability matrix v1 and model v1-mvp at the representative Tier 2 and mid-experience inputs. Median pay uses BLS OEWS and market compensation data. See the methodology at /methodology.
FAQ
How to read the tracker
Is this an AI-risk score?
Where do the numbers come from?
Why is the AI run-cost so much lower than the pay?
Why these roles, and why one representative cell?
Build a Wagecard with your own tasks and hours.
Choose your location, experience, salary, and task mix. See the preview before you create an account.