The AI investment boom (or bubble) is something Silicon Valley has seen many times before, a gold rush where venture capital money is thrown into big new things. But there is an aspect to this era that is completely unique. That's when a startup's annual recurring revenue can skyrocket from $0 to $100 million in a matter of months.
Word on the street is that many VCs won't look at startups that aren't on the ARR superhighway, aiming for $100 million in ARR before a Series A funding round.
But Jennifer Lee, a general partner at Andreessen Horowitz who oversees many of the firm's most important AI companies, warns that some of the ARR mania is based on myth.
“Not all ARR is created equal, and not all growth is created equal,” Lee said on an episode of TechCrunch's Equity podcast. She said she's especially skeptical of founders announcing impressive ARR numbers or growth in tweets.
There is now a legitimate and well-known term in accounting parlance called annual recurring revenue. This refers to the annualized value of contracted recurring subscription revenue. This is a guaranteed level of income, as it is essentially income from contractual customers.
But what many of these founders are tweeting is actually their “earnings run rate,” or the annualized amount of money paid over a period of time. It's not the same.
“There's a lot of nuance missing from that conversation about the quality of the business, the maintenance, the durability,” Lee warned.
tech crunch event
Boston, Massachusetts | June 23, 2026
A founder may have just had a strong sales month, but that doesn't necessarily mean it will repeat every month. Alternatively, a startup may have a large number of short-term customers running a pilot program, and there is no guarantee that the revenue will continue beyond the pilot period.
Typically, such boasts about growth via tweets should be taken for what they are. Don't believe everything you read on the internet at face value.
But since rapid growth is a hallmark of AI startups, such claims “cause a lot of anxiety” for inexperienced founders, she says.
Lee's answer: “No, of course that's a nice aspiration, but you don't need to structure your business that way to optimize just sales growth.”
A better way to think about it, she said, is to grow sustainably: once a customer signs up, they never leave and grow their spending with your company. This could lead to “5x or 10x year-on-year growth,” Lee said, meaning growth from $1 million to $5 million to $10 million in the first year and $25 million to $50 million in the second year.
Mr Lee noted that this is still an “unprecedented” level of growth. If this is coupled with customer satisfaction, or high retention rates, such startups will find investors willing to back them.
Of course, some portfolio companies in Li's a16z group (infrastructure team) have achieved these kinds of ARR numbers, including Cursor, ElevenLab, and Fal.ai. But that growth is tied to “durable businesses,” Lee said, adding: “There are real reasons behind each one.”
Lee also said this type of growth comes with operational issues, such as hiring.
“It’s not about speed, but how do we hire the right people who can really jump into this kind of speed and culture,” she said. The answer is “it's not easy.”
That means the first 100 people wear many hats, and mistakes are bound to happen. Cursor, for example, angered its customer base last year by poorly rolling out price changes.
Lee pointed out that other fast-growing startups are dealing with legal and compliance issues before putting systems in place to face new challenges in the AI era, such as countering deepfakes.
So while lightning growth may be a good problem to have, it's also something a little like. “Be careful what you wish for.”
Listen to the entire episode here: