SaaS Is Not Dead — But Investors Are Sorting Winners From Losers
Software-as-a-service is not dead, but the market has entered a phase of careful discrimination. Investors are sifting through the field, trying to identify which companies will emerge as winners and which will fall behind. The defining question they are asking is: where are the moats?
The strongest moats belong to companies that own their data — the data that powers large language models like Claude and OpenAI's products. A company that serves as the underlying data infrastructure for an entire ecosystem occupies a defensible position. A useful contrast illustrates this. On one side is a company like Data Dog, which functions as the data rails for essentially any company that uses it; it has emerged as a clear winner, evident both in its share price and in the strength of its reported numbers. On the other side is a company like Adobe, which still has a great deal to prove to the market. The reason is that AI players — Figma, Claude, and other strong AI companies — are now able to disrupt much of what Adobe does, and the major model providers are rolling out their own competing products. Anything that a model provider like Claude or OpenAI can quickly disrupt, or replicate with its own offerings, puts the incumbent in a difficult position. This dynamic is playing out across the sector: some companies are struggling while others are flourishing.
The Next Evolution: Owning the Outcome, Not Just the Workflow
A central argument is that the next evolution in software is owning the outcome, not just the workflow. What does that mean?
Up until roughly a year ago, software was largely static. A user or employee would input data into a system, glean the insights themselves, and then manually execute on the action that followed from that data. The software was essentially a read-only system — it held and displayed information, but the human did the work of acting on it.
What the best companies are now doing is moving from "read" to "read, write." Software no longer merely accepts data input; it can now agentically operate on the workflow and execute the action itself, without the user having to perform it manually. The future lies in software that can both ingest data and then take action on the user's behalf.
This is what is currently in vogue and what is "hot," and the best companies are beginning to deliver on it. It is, however, still early days. Not every company is executing on this strategy — and it would be dishonest to claim that every top-performing public software company is executing on it. But the early data is showing very positive trends in the ability to automate a great deal of the work that human labor was doing up until about a year ago.
The "Haves and Have-Nots" Divide
The market has split into a "tale of the haves and have-nots." The dividing line is clear: companies that can demonstrate AI-accelerated growth are rewarded, and those that cannot are being beaten down in their stock price.
Crucially, this is not limited to the model providers. Many established public software companies — those that are not OpenAI or Anthropic — are now showing that they can genuinely accelerate growth by leveraging the data and workflows they have ingested over 15, 20, or more years as large public companies. That accumulated data and workflow knowledge becomes a powerful asset when combined with AI. The very best of these companies are demonstrating AI-accelerated growth; the others are not, and they are suffering for it. If you want to do well as an investor, you very much want to be in the "haves" category.
Why OpenAI and Anthropic Are the Exception
There is aggressive investor appetite for OpenAI and Anthropic and their two upcoming IPOs, despite the fact that both companies have ongoing losses. What makes them the exception, when the rest of the field is being judged on a focus on profitability?
The answer is hypergrowth. The market is clear: if you show AI-accelerated growth, you command significant price premiums; if you do not, you don't. Very few companies are showing growth at the level and scale that OpenAI and Anthropic are achieving — both the sheer scale and the rate of that growth are remarkable. These are described as "two in a generation" companies; you do not see this kind of growth often. That rarity is precisely why investors are becoming more and more excited about them. A striking data point underscores the scale: the combined private market capitalization of OpenAI and Anthropic is larger than pretty much every other software company out there combined. This is a trend expected to intensify over time.
The AI Price War and the Profitability Question
Even before these companies have gone public, an AI price war is beginning to take shape. There are reports that OpenAI is considering dramatic price cuts to win over customers — particularly customers away from Anthropic. (It is worth emphasizing "considering," because nothing has actually happened or been formally announced yet.) How far can they push this before it becomes a profit problem?
It is a profit problem. There is genuine eagerness to see these companies' S1 filings to understand the underlying economics. The situation closely resembles the early days of Uber and Lyft, when both were young hypergrowth companies and venture capital was subsidizing their gross-margin-negative losses. OpenAI and Anthropic appear to be doing the same thing now. It is fundamentally a land grab — a race to see who can capture the most market share. Neither company seems to have much regard for its current margins, even though both are rumored to be going public soon, because the market is clearly signaling that AI-accelerated growth is what matters most.
With this much pent-up demand — both private and public — for both platforms, the contest right now is simply about who captures the most market share. The strategic goal, in the Uber-versus-Lyft analogy, is clear: you want to be Uber, not Lyft. Both companies are benefiting from the tailwind of having access to what amounts to nearly unlimited capital.
Where AI-Driven Software Growth Is Happening Fastest
The conversation around AI has shifted. It is no longer primarily about adoption; it is now about beginning to see returns and offering concrete use cases. So where is the growth fastest?
Low-hanging fruit — software that is not deeply embedded. About a month ago, Claude released its small-business product line. This points to where rapid growth is occurring: the "horizontal lift." Consider a small business that does not want to manage 20 or 30 different software systems. If Claude can do it all — sparing the business from talking to 30 different vendors and going through 30 separate implementations — adoption comes quickly. Software that is not deeply entrenched is the easiest to displace, and that is where a great deal of very fast adoption is happening.
Core verticals. Beyond the horizontal opportunities, growth is strong in specific verticals:
- Cybersecurity — a sector that has been hot for years but is especially hot at the moment.
- Healthcare — increasingly active, and viewed as positive not only for financial returns but also for general health and the goal of helping people live longer lives.
The Open Question: Customer Retention
While adoption of the low-hanging fruit is happening very quickly, time will tell how durable it is. The crucial unknown is how well these companies retain their customers — and that answer will not be clear for a while yet. The key thing to watch is whether retention rates come to mirror those of top-tier public SaaS companies. If they do, the current growth will prove sustainable; if they don't, the rapid adoption may not translate into lasting value. That remains the open question hanging over the entire trade.