Time’s Up for SaaS (Grow Faster or Vanish)
A review of the recent public SaaS drawdown
The public software (SaaS) index is down 37% since the end of Q3’25, and the median company trades at a paltry 3.5x NTM (next-twelve-months) revenue. We have had 2 consecutive quarters of negative share price returns; the only other time this has occurred in the last 10 years was in 2022. The public markets want high growth, yet growth rates for public SaaS have been lackluster for years.
Put simply, time is up for public SaaS to grow revenue (through AI) or vanish. Investors have seen AI companies such as Anthropic grow from $0 to nearly $10B in annualized revenue, and Cursor grow from $0 to $1B in annualized revenue, both in very short periods. Public markets are demanding high growth, and unfortunately, SaaS isn’t delivering; share prices are paying the price.
While the recent releases of Anthropic’s plug-in and Cowork, and OpenAI’s Frontier signal a deeper push into business applications, which could be to blame, continued slow revenue growth may be the primary issue. IT budgets are also realigning toward consumption-based models, and seat-based pricing is under pressure, adding further uncertainty. The markets seem to be realizing that the widely expected AI growth re-acceleration in public software is not materializing.
The following chart shows the average LTM (last-twelve-months) revenue growth rate across all public software over the past 10 years. Average growth has slowed to 16% year-over-year last year, its lowest level in a decade.
For many years, the most reliable assumption about public SaaS was that it would continue to grow durably even with a larger revenue scale. Quarter after quarter, companies would ‘beat and raise’, and the average company grew at 30-40% for almost 10 years, culminating in a growth pull-forward and high multiples during 2021, the ZIRP (zero-interest-rate policy) period. Post-ZIRP, rates rose, demand slowed, and revenue growth rates fell. AI was supposed to come to the rescue and reinvigorate growth rates. That has not yet happened, and the market appears to be betting that it won’t.
ServiceNow: The Barometer of Success in SaaS
ServiceNow ($NOW) is easy to highlight, as it is arguably one of the most (if not the most) successful SaaS/Cloud 1.0 companies. They went public in 2012 at a $2.7B market cap and reached a peak market cap of over $240B in January, up ~65x on a per-share basis from their IPO price. Just last week, ServiceNow reported their Q4 and full-year 2025 results; the company exceeded top and bottom-line targets and did $3.6B of revenue in Q4, growing 21% year-over-year with $2B of free cash flow. ServiceNow is a Rule of 56 company that consistently grows by 20%+ each year and generates billions in free cash flow. By any comparison, the company is performing exceptionally well and is the benchmark of success. However, the stock is down dramatically over the past year and by 30% over the past month as of this post.
The following chart shows ServiceNow’s references to AI in their 10K and annual revenue growth over the past 4 years. ServiceNow is telling a compelling narrative around AI, but it’s not yet showing up in the numbers. Although the company reported their “Now Assist” outperformed expectations and surpassed $600M in ACV in Q4. Even so, this is not enough to move the needle on overall revenue growth and market sentiment.
ServiceNow is a $14B+ ARR business, growing 20%+, generates billions of free cash flow annually, and has a consistent 98% customer renewal rate; the company is valued at 7.0x ARR! The bar for best-in-class growth in an AI world is very different from that in the prior SaaS world. While private markets have experienced a dramatic shift in expectations for revenue growth, the public markets are now catching up.
Time to ‘Burn the Ships’ on Product and Margins?
While ServiceNow has a compelling AI vision, their moat is likely their Achilles’ heel. As one Global 50 CIO told me, “I have a battalion of people that manage our ServiceNow deployment.” ServiceNow has spent over 2 decades developing a broad and deep software surface area, a complex data model, and a highly secure and performant platform. Moreover, a significant number of people have spent their entire careers, or much of it, developing expertise in managing complex deployments at the world’s largest companies. How can a company like ServiceNow go and disrupt itself and its customers and champions?
Many private AI-native companies also operate with margins significantly below those of traditional SaaS (20-40% gross margins), despite astronomical revenue growth. Theoretically, with the right product offering, growth could come quickly but would be accompanied by much lower margins. How would the Street react to higher revenue growth but significantly lower free cash flow margins? And what are the expectations on the ROI of such a change? There are also no precedents on how AI-native software companies will trade or be valued in the public markets. So companies not only need to innovate on product but also take (for at least the moment) significant margin risk to grow faster for unknown valuation multiples. In reality, these decisions are obviously much more nuanced than I described, but taking both product and margin (and value) bets simultaneously as a public company is no easy feat.
NVIDIA’s Parabolic Growth
Compare that to NVIDIA, the stalwart barometer of AI in the public markets. Their revenue growth has been parabolic over the past few years. The following chart shows NVIDIA’s LTM (last-twelve-months) data center revenue over the same time period. It’s growing at a rate and scale that no company has likely ever experienced. The share price has followed suit.
Data might not be a Moat
The data moat surrounding incumbent SaaS / systems of record is likely overstated. I would argue that the most important context doesn’t reside in systems of records, regardless of their size, but in the minds and multimodal work of millions of users over time. This is now moving to agents. Even if systems were rearchitected to capture this critical data, how would companies leverage or train on customer data? And capture data not only in their apps but also in a multimodal fashion? We are likely to see a tectonic battle of agents competing for this critical “task” context. It’s also unlikely that companies will rebuild Salesforce, Workday, or ServiceNow with vibe-coded internal applications at scale; these applications are not going to die, but will instead slowly vanish as they become databases for agent pings. OpenAI’s new announcement around Frontier puts this in plain sight (see below). Note systems of record sit at the bottom of this graphic.
In some ways, this shock to the system could be beneficial, as it serves as a rallying cry to further innovate, and the prize for success is bigger than ever. While every SaaS company has had a compelling AI narrative for some time (i.e., every SaaS company is now an AI company), it’s truly wartime to deliver on AI products and revenue growth (and consider margins). It’s not to say that all is lost; this debate could be moot if companies report Q4 earnings with strong AI revenue growth and not just narratives, yet market sentiment is certainly betting against it. While we are very bullish on the best teams/companies innovating – and most importantly re-accelerating revenue growth through AI – the public markets are telling companies that the time is up to deliver on it.
Thank you to my partners Tanner, Austin, and Will for their help on this post.
Source: CapitalIQ, OpenAI blog, Wall Street Equity Research, and public company filings as of February 5, 2026.