a16z VC wants founders to stop stressing over insane ARR numbers

The AI investing boom — or, as some see it, an AI bubble — is a familiar pattern in Silicon Valley: a rush of venture capital pouring into the Next Big Thing. But one aspect of the current moment stands apart from past cycles. Startups are racing from $0 to $100 million in annual recurring revenue in astonishingly short periods, sometimes within just a few months.

In some corners of the venture world, the prevailing narrative is that many investors won’t even consider a startup unless it’s already speeding down the ARR superhighway, with founders aiming to hit $100 million in annual recurring revenue before raising a Series A round.

But Andreessen Horowitz general partner Jennifer Li, who oversees several of the firm’s most significant AI investments, cautions that much of the current obsession with ARR is built on shaky assumptions.

“Not all ARR is created equal, and not all growth is equal either,” Li said. She added that founders and investors alike should be particularly sceptical when eye-catching ARR figures or growth claims are casually announced on social media.

There is, of course, a well-established accounting concept known as annual recurring revenue, which refers to the annualised value of predictable, contract-based subscription revenue. In other words, it represents relatively secure revenue because it comes from customers who are contractually committed.

What many founders are actually referring to in online posts, however, is closer to a “revenue run rate” — a figure derived by taking revenue from a short period and extrapolating it over a full year. That approach can paint a misleading picture.

“There’s a lot of missing nuance around business quality, retention, and durability that gets lost in that conversation,” Li warned.

A startup might have an exceptional sales month that inflates its run rate, but that performance may not repeat consistently. In other cases, companies may be generating revenue from short-term pilot programs, meaning the income is not guaranteed to continue once those pilots end.

Ordinarily, bold claims about growth on social media should be taken with a grain of salt. Not everything posted online should be accepted at face value.

But because rapid growth has become a defining trait of many AI startups, these high-profile ARR claims are “introducing a lot of anxiety” for less experienced founders, Li said. Many now wonder how they, too, can go from zero to $100 million almost overnight.

Her response is straightforward: “You don’t. Sure, it’s a great aspiration, but you don’t have to build a business that way, or optimise solely for top-line growth.”

Li argued that a healthier approach is to focus on sustainable expansion — building a business where customers stay, renew, and gradually increase their spending over time. That kind of model can still deliver extraordinary growth, such as “5x or 10x year-over-year,” she said. In practical terms, that could mean growing from $1 million to $5–10 million in the first year, then to $25–50 million in the second year, and continuing from there.

Even at that pace, Li noted, the growth would be considered “unheard of” by historical standards. When paired with strong customer retention and satisfaction, those businesses will still attract investors.

Some of the companies in Li’s infrastructure-focused portfolio at a16z have indeed achieved the kind of eye-popping ARR growth that dominates headlines, including Cursor, ElevenLabs, and Fal.ai. But Li emphasised that their growth reflects “durable businesses,” adding, “There are real reasons behind each of them.”

She also pointed out that hypergrowth brings its own serious challenges, particularly around hiring.

“How do we hire — not fast, but the right people — who can really jump into this kind of speed and culture?” Li said. “And the answer is: not easily.”

In many cases, the first 100 employees end up wearing multiple hats, and mistakes are almost inevitable. Last year, for example, Cursor drew criticism from users after a poorly executed pricing change.

Other fast-growing startups, Li noted, find themselves grappling with legal and compliance issues before they’ve built the systems to manage them, or facing new AI-era challenges such as dealing with deepfakes.

Ultimately, while explosive growth can be an enviable problem to have, Li cautioned that it also comes with risks. In her view, it’s a reminder that even in an AI gold rush, founders should be careful what they wish for.