Why Net Revenue Retention Is the AI Investors’ Metric of Choice

The artificial intelligence (AI) revolution has been long on buzz but short on reliable performance measures. GenAI is a new world with new business models. How do investors know if a great idea will result in long-term growth?

To cut through the noise, VCs are increasingly looking at net revenue retention (NRR). It measures loyalty and captures a startup’s ability to keep customers coming back while growing their average spend. Will a new model be sticky or become another victim of try-and-abandon? NRR metrics help tell the tale.

We look at net retention’s significance in the AI age, explain how it’s measured and where it fits into startup financials.

Table of Contents Table of Contents

1. Key Takeaways
2. What Is AI's ROI?
3. Why Do So Many AI Startups Fail?
4. What Is Net Revenue Retention?
5. Calculating Net Retention for AI Firms
6. How to Calculate NRR
7. What VCs Are Looking For
8. The Bottom Line
9. FAQs
10. References

Show Full Guide

Table of Contents

1. Key Takeaways
2. What Is AI's ROI?
3. Why Do So Many AI Startups Fail?
Show Full Guide
4. What Is Net Revenue Retention?
5. Calculating Net Retention for AI Firms
6. How to Calculate NRR
7. What VCs Are Looking For
8. The Bottom Line
9. FAQs
10. References

What Is AI’s ROI?

Most startups fail, some spectacularly. New research shows the flame-out effect is even bigger in AI, where 90% of companies crash out within a year of launch.

VCs have sophisticated systems for capturing growth potential, but even with rigorous due diligence, sometimes a Theranos slips through. For investors trying to decide if you’re a budding OpenAI or a Humane-level sinkhole, the net revenue retention rate (NRR) can signal that you have what it takes to punch through a competitive and cost-conscious market.

Ace it and you could be resilient enough to beat unexpected challenges, sustain profitable growth, and maybe, just maybe, stay in the game long enough to secure a new funding round.

Why Do So Many AI Startups Fail?

Figures from Edge Delta suggest there are more than 70,000 AI companies operating around the world in multiple sectors. The vast majority began their life as startups, cumulatively pulling in VC investments totalling $75 billion – though that was in 2020. Today, more than half of global VC funding goes into AI-focused companies.

There’s a lot of money sloshing around in a still-unproven category that tends to produce duds. Of course, VCs want to find the next Anthropic or Perplexity. The trouble is that their potential strengths and weaknesses may be masked.

That weakens claims around differentiation. If an ‘AI-powered’ app is just a re-skinned version of Claude, a competitor could do something similar.

Worse, Anthropic might like the idea so much that they launch a native version of their own, quickly wiping you off the map.

What Is Net Revenue Retention?

Net revenue retention (‘net retention’ or NRR) helps VCs see through the fog and identify likely winners. It measures how much money current customers are spending on a software product or service over time.

In SaaS – which AI is a subset of – investors expect net retention to stay comfortably above 100%. The higher, the better, because if you’re bringing in new customers with a propensity to spend more on your product as time goes on, your sales and marketing strategy is clearly working.

More importantly, it suggests the product or service has underlying utility that customers value, a scenario that places the startup on a trajectory to future growth.

Traditional SaaS investors look for NRR between 110% and 115%. In AI, even the biggest firms are struggling to keep paying users.

Calculating Net Retention for AI Firms

NRR is calculated by adding up three measures:

• Revenue expansion: Money made by upselling or cross-selling new AI products and services to current customers.
• Revenue contraction: Money lost due to current customers spending less or using an AI model less frequently.
• Customer churn: Money lost when customers stop using an AI service or switch to a flashy new one (e.g., DeepSeek).

A high NRR signals customers find your model effective, the interface is easy to use, and you’ve achieved a level of stickiness factor that keeps them around.

How to Calculate NRR

For example, let’s say an AI model took in $100,000 in revenue last quarter from current paying customers. In the following quarter, they generated $20,000 in expansion revenue, lost $5,000 due to reduced spending, and lost $1,000 to churn.

Net retention would be calculated this way:

(100,000 + 20,000 – 5,000 – 1,000) / 100,000 = 1.14

NRR = 100 x 1.14 = 114%

What VCs Are Looking For

With AI’s huge compute and energy costs, customers come at a cost. If they aren’t generating enough revenue, the burn rate will force VCs to rethink their exit timelines while pouring more cash into the venture.

By monitoring NRR, investors are looking for operational and marketing outcomes that will keep their plans on track. For example:

Taken together, AI startups that demonstrate high net retention metrics are more stable, scale more easily, and have higher potential for growth. Saas firms that achieve an NRR of 115% or higher have been shown to win significantly higher valuations.

The Bottom Line

With huge compute costs and energy requirements waiting to eat into margins, AI firms need to make customer relationships last – and pay – otherwise they burn up cash forever. That favors a Microsoft-sized company that can build and run its own infrastructure.

A successful AI startup needs to be clever enough to outmaneuver the giants, or visionary enough to set out its stall in a niche market that Big Tech ignores. NRR can show you where an AI company fits on that spectrum.

FAQs