What this tool does
AI business valuation is the process of estimating the value of a business whose core product relies on machine-learning models, such as an AI writing assistant, an image tool, an AI agent or an API built on a large language model. Most small AI businesses sell subscriptions, so the tool values them like SaaS, on a multiple of ARR, and then applies adjustments specific to AI products.
Use it if you run an AI product and want a grounded estimate before selling, or if you're a buyer weighing an AI listing whose revenue has grown quickly.
Inputs you need
| Input | Why it matters | Required |
|---|---|---|
| Monthly recurring revenue (MRR) | Base of the calculation | Yes |
| Average monthly net profit | Shows the implied profit multiple | No |
| Revenue change, last 12 months | AI products can grow fast, and just as quickly lose ground | No |
| Monthly customer churn | Many AI tools see high trial-driven churn | No |
| Gross margin | Model inference and API costs can take a large share of revenue | No |
| Model provider dependency | Reliance on one provider's pricing, limits and policies | No |
| Largest customer share, age, owner hours | Concentration, durability and workload | No |
For gross margin, subtract direct costs from revenue: model API or GPU spend, hosting, payment fees and any per-request data costs. Then divide by revenue.
How the valuation is calculated
- ARR = MRR × 12.
- Base multiple. AI businesses start from 1.5× to 4.5× ARR (midpoint 3.0×), slightly below VALUERAQ's SaaS default. The lower starting point reflects two assumptions: small AI products typically have less operating history, and their gross margins and competitive position can shift when the underlying models change. Both are platform assumptions; see data sources.
- Adjustments for growth, churn, gross margin, provider dependency, customer concentration, age and owner time, as published in the methodology.
- Value range = ARR × adjusted multiples, with an implied profit multiple if you enter profit.
Worked example
The form opens with a hypothetical AI tool: $12,000 MRR, $4,000 monthly profit, 60% annual growth, 6% monthly churn, 62% gross margin, relying on a single model provider, 18 months old and needing 30 owner hours a week.
- ARR: $144,000
- Combined adjustment: × 0.74
- Adjusted ARR multiple: 1.1× to 3.3×
- Estimated value: $160,000 to $480,000, midpoint $320,000
- Implied multiple of annual profit at the midpoint: 6.7×
| Factor | Example input | Effect |
|---|---|---|
| Business age | 1.5 | −10% |
| Revenue trend | 60% | +15% |
| Owner involvement | 30 | −5% |
| Customer concentration | 5% | 0% |
| Churn | 6% | −10% |
| Gross margin | 62% | −7% |
| Model provider dependency | Relies on one third-party model provider | −10% |
Strong growth lifts the multiple, but short history, high churn, a below-software gross margin and single-provider dependence each pull it down. That combination is common in young AI products.
What buyers look for in an AI business
What the product adds beyond the model. If a competitor, or the model provider itself, could reproduce the product with a prompt and a simple interface, the revenue is fragile. Proprietary data, workflow integrations, a distribution channel or a trusted brand make it more defensible.
Unit economics under changing model prices. Buyers will ask what happens to margin if inference prices rise, or whether the business keeps the savings when they fall. Show cost per active user over time.
Provider terms. Check the model provider's terms of service, rate limits and usage policies. A product that could be switched off by a policy change is riskier. The ability to move between providers reduces that risk.
Churn after the novelty fades. Many AI tools attract sign-ups that don't last. Cohort retention after three and six months tells a buyer more than total sign-ups.
Quick value table
Default ARR multiples for AI tools and AI SaaS before adjustments. They start lower than for other SaaS because of model costs and provider dependence.
| ARR | MRR | Low (1.5×) | Midpoint (3.0×) | High (4.5×) |
|---|---|---|---|---|
| $60,000 | $5,000 | $90,000 | $180,000 | $270,000 |
| $120,000 | $10,000 | $180,000 | $360,000 | $540,000 |
| $250,000 | $20,833 | $375,000 | $750,000 | $1,125,000 |
| $500,000 | $41,667 | $750,000 | $1,500,000 | $2,250,000 |
| $1,000,000 | $83,333 | $1,500,000 | $3,000,000 | $4,500,000 |
Why small AI businesses don't get headline AI multiples
Reports of AI companies valued at very high multiples of revenue refer to large, venture-funded businesses with their own research teams and models. A small AI tool built on another company's model is a different asset: its costs rise with usage, its features can be copied, and its supplier can change prices. Buyers of small AI businesses therefore price them more like risky SaaS than like the companies in the news.
The AI-specific checks
| Check | Why it matters | How the tool treats it |
|---|---|---|
| Gross margin after model costs | Inference costs grow with every user | Under 50% −15%; 50% to 70% −7%; above 80% +5% |
| Model provider dependence | One provider can change prices, terms or quality | Single provider −10%; can switch providers 0%; own or fine-tuned models +5% |
| Churn | AI tools are often tried and dropped | Under 2% a month +10%; over 8% −20% |
| Customer concentration | A few large clients can leave together | Over 30% from one customer −12% |
| Proprietary data or workflow | Harder to copy | Not scored; note it for buyers in due diligence |
Common mistakes when valuing an AI business
- Using revenue multiples from venture funding rounds as if they were sale prices.
- Quoting gross margin before model and API costs.
- Ignoring how much of the product a model provider could add to its own platform.
- Counting free-trial or one-off credits as recurring revenue.
Limitations
- Venture-scale AI companies are valued on expectations this tool doesn't model.
- The tool can't assess technical defensibility, model quality or data rights; those need expert due diligence.
- The base multiple is an editorial assumption, and the AI market changes quickly.
- The estimate depends entirely on the figures you enter. It is not an appraisal or investment advice.
Frequently asked questions
How are AI businesses valued?
Small AI businesses with subscription revenue are usually valued like SaaS, as a multiple of ARR or annual profit. Buyers then discount for risks particular to AI: dependence on one model provider, thinner gross margins from inference costs, and high churn.
Why is the starting multiple lower than for SaaS?
It is a platform assumption reflecting that most small AI products are young and that their margins and competitive position can change when underlying models change. Strong fundamentals can lift an AI business above a comparable SaaS business once adjustments are applied.
Does using OpenAI, Anthropic or Google models lower my valuation?
Using a third-party model isn't a problem in itself. Relying on a single provider with no fallback adds risk, so the tool applies a small negative adjustment. Being able to switch providers removes it, and running your own or fine-tuned models adds a small positive adjustment.
Can I value a pre-revenue AI startup?
No. This tool needs recurring revenue. Pre-revenue startups are usually priced through negotiation with investors on team, technology and market size.
Does a buyer pay more if I use several model providers?
Being able to switch providers removes the single-provider discount in this tool. Owning or fine-tuning your own models adds a small premium, because it reduces the risk that one supplier changes your costs or product.
Should I include API credits I resell in revenue?
Include them in revenue, but make sure their cost appears in your gross margin. Buyers look at what is left after model and hosting costs.