AI Break-even Price: Formula and Margin Examples

Question

What monthly price breaks even on an AI feature and preserves target margin?

Quick answer

Formula: break_even_price = cost_per_user_month (0% margin baseline)

Target margin formula: required_price = cost_per_user_month / (1 - margin)

• Assumption: cost model includes all variable and infra cost components.
• Assumption: pricing is per active paying user per month.
• Assumption: market willingness to pay is validated separately.

Example: if cost is $12, break-even is $12; at 75% target margin, required price is $48.

In practice, break-even is your floor, not your launch price. Once you know the floor, the next question is whether your target margin still fits what the market will pay.

This page calculates the customer price floor. Implementation ROI and payback require separate revenue or labor-savings inputs.

Fastest Working Method

1. Estimate cost per user/month with realistic request volume, token, and infra assumptions.
2. Use that cost as the 0% gross-margin floor.
3. Apply your target gross margin to compute the actual required price.
4. Compare the result against current packaging and comparable products before changing price or scope.

What Usually Moves Break-even the Most

• Request frequency per active user.
• Retrieved context size and chunk count.
• Model choice for generation and fallback paths.
• Cache hit rate and any repeated-answer savings.

Margin Ladder Example

Start with one cost number and expand it into price floors. If cost per active user each month is $12, break-even is $12, a 70% gross-margin target implies $40, an 80% target implies $60, and an 85% target implies $80.

This is the practical reason small cost changes matter so much. Cutting cost from $12 to $10 only saves $2 at break-even, but it lowers the 80% target price from $60 to $50.

Common Pricing Mistakes

• Using only token cost and ignoring retrieval, vector, or infra overhead.
• Treating break-even as a healthy price instead of a bare minimum.
• Mixing monthly cost with annual pricing without normalizing units first.