The New Speed of Light: AI Valuations, Margins, and the Race to $100M ARR

When it comes to private AI-native SaaS companies, the speed limit has been broken.

For nearly a decade, the "gold standard" for a best-in-class SaaS startup was reaching $100M in Annual Recurring Revenue (ARR) within ten years. Today, that playbook is being rewritten in real-time.

AI-native companies are shattering records that were once thought untouchable. Cursor, an AI coding assistant, reportedly sprinted to $100M ARR in roughly 12 months, beating the previous records held by Wiz (18 months) and Deel (20 months). Not long after, reports suggest that Swedish AI unicorn Lovable may have hit this milestone in just 8 months.

This velocity signals a massive appetite for innovation. It’s also fueling aggressive valuations. As private AI companies accelerate their growth, a critical question emerges: Is this growth sustainable? And how much do margins matter?

The "Margin Doomsday" Myth

There is a growing concern that certain AI companies are growing too quickly, without a prudent focus on margins. This is because they are not typically delivering the same margins as traditional SaaS companies.  

Best of class SaaS companies historically have enjoyed gross margins north of 80% because the cost to serve one additional customer is near zero. AI companies, however, have to pay a "compute tax" for every query, image generated, or line of code written. This has pushed their margins down to the 50-60% range, making them look more like low-margin services businesses than high-flying tech stocks.

But is questioning margins a "boring cliché"?

History is ripe with companies that were initially dismissed for low margins: Amazon, Salesforce, and Uber all faced similar skepticism. The reality is that lower margins in the early days of a business are not necessarily static. For AI-native companies specifically, they can be improved over time through:

• Model Routing: Automatically sending simple queries to cheaper, smaller models and only using expensive "frontier" models for complex tasks.
• Hardware Optimization: The cost of inference is dropping rapidly (some estimates say 10x-100x in 18 months).
• Pricing Power: As AI tools become essential, companies can introduce "Pro" and "Enterprise" tiers that command higher prices, offsetting compute costs. Therefore, a company’s ability to convert free users to paid users is an increasingly important metric to watch.

Moat Over Margins

There is another reason why smart investors shouldn't necessarily be scared off by lower initial margins: Moats are expensive to build.

Enterprise giants like ServiceNow and Workday started with gross margins in the 50-60% range. Why? Because they had to invest heavily in professional services and implementation to get their software deeply embedded in customer workflows.

Today’s AI agents aren't just chatbots. They are replacing complex human workflows. To do that effectively, they often require a serious investment to set up integrations, guardrails, and context. This human-in-the-loop approach lowers margins in the short term but builds sticky, bespoke workflows that competitors can't easily rip out. OpenAI’s investment in Thrive Holdings is a prime example of this investment in a distribution moat. They aren't just selling software to accounting and IT service firms. They are taking an ownership stake in them. OpenAI is embedding engineers directly into these "boring" industries to fundamentally rewire how they operate. 

Once that lock-in is established, margins can climb. ServiceNow, for example, now boasts gross margins of nearly 79%.

How to Navigate the Boom

So, where is the equilibrium? And how should investors think about making informed decisions?

• Hunt for the "Up and Comers": While "Centicorns" (companies valued over $100B) like Databricks and Stripe may offer stability, the steepest growth curves are often found in the "up and comers", companies valued under $10B that are just hitting their hypergrowth stride. These are the companies more likely to replicate the "Cursor trajectory."
• Look for "Strategic Shields": As noted in our report on Big Tech's Strategic Alliances, many top AI startups are partnering with giants like Microsoft, Google, and NVIDIA. These alliances provide the capital and infrastructure shield necessary to survive the high-cost growth phase, giving them a distinct advantage over standalone competitors.
• Look Out for Churn vs. Burn: Hypergrowth can sometimes mask poor product-market fit. An AI company might race to $100M ARR by spending aggressively on customer acquisition, but if those customers leave after a few months because the product is a "nice-to-have" rather than a "need-to-have," the business is unstable. Look for companies that are showing signs of successfully keeping and upselling their customers, proving the product is sticky to customers.

When it comes to AI, we are not in a doomsday scenario; we are in the deployment phase. The companies that will define the next decade are being built right now, and they are reaching scale faster than any in history.

Yes, margins matter, but not as much in the earlier innings. Great investments will be found by identifying the companies that are using compute spend to build something durable that becomes indispensable to its users.

Interested in exploring these opportunities? Check out the leading private AI companies trading on EquityZen today.