Most of us have been doing this long enough to remember when software stocks trading at 40x forward earnings felt cheap.
That era appears quite .. over.
Software forward P/E multiples have now fallen below the overall S&P 500. Not at parity. Below.
This has never happened before.
Not in 2022 when rates spiked. Not in 2008. Not even in the dot-com crash, which was a speculative bubble unwinding — not a structural attack on the business model itself.
The Numbers Tell the Whole Story
Here’s how the multiple has collapsed since the peak:
- May 2020 – May 2022: 84.1x forward P/E. Zero interest rates, COVID digital acceleration, SMB going online overnight. Software was trading at nearly 4x the S&P 500.
- June 2022 – June 2024: 43.2x. The rate hike correction hit hard. But software still commanded roughly 2x the market multiple. The premium held.
- July 2024 – June 2025: 33.6x. Seat-based headwinds as enterprise hiring froze. Still a premium, just a smaller one.
- July 2025 – December 2025: 31.2x. AI starts changing the top of the funnel for application software. Investors start getting nervous about LLM providers eating into the stack.
- January 2026 – March 2026: 22.7x. App software disruption by AI becomes the main narrative. Software is now at or below market.
IGV, the iShares software ETF, is down over 21% year-to-date. Since its September 2025 peak, it has fallen roughly 30%. That is $2 trillion in market cap gone.
Why the Premium Existed in the First Place
For two decades, software companies deserved to trade at a premium to the market. The logic was straightforward:
- 70-80%+ gross margins vs. the S&P 500 blended average of ~45%
- Recurring revenue with predictable, contractual cash flows. Annuities, in many ways.
- Negative net revenue churn in the best businesses
- Growth rates that made the index look like a utility stock
You pay more for better economics. That is not a complicated investment thesis.
The premium also reflected the scalability of the model. A dollar of new software revenue costs almost nothing to deliver. That is simply not true for most businesses in the S&P 500.
What Broke It
Two things per Bloomberg: “app software disruption by AI” and “private credit concerns for software.”
Both matter. But the AI disruption thesis is driving the larger repricing.
The core fear is seat compression. If a single AI agent can do the work of multiple human employees, enterprises stop buying 500 seats and start buying 100. Or 50. Or renegotiate entirely.
Orlando Bravo, who has spent 20+ years buying and building software businesses at Thoma Bravo, said publicly this month that some of the software companies being disrupted by AI are facing “very warranted” decreases in their valuations. That is a significant thing for him to say out loud.
The seat-based model built the traditional B2B industry. Salesforce, Workday, Atlassian, Monday.com — all of them grew by selling per-seat licenses to expanding workforces. When headcount grows, revenue grows. Automatic. Durable. Predictable.
AI itself is breaking that link. As is incremental budget all going to AI.
If agents replace seats rather than complement them, the revenue model doesn’t just slow down — it reverses. That is not a cyclical problem. That is a structural one. And if all incremental budget is going to Anthropic and AI Agents, that leaves very little left for traditional productivity software.
Has This Ever Happened Before?
Not like this. The 2000-2002 crash saw tech multiples collapse, but that was speculation unwinding, not business model disruption. The 2008-2009 financial crisis hit everything, but software recovered fast because the underlying economics were still intact.
This time, the market is not saying “software is temporarily overvalued.” It is saying “we’re not sure the earnings growth assumptions embedded in even 22x are correct.”
That is a fundamentally different kind of repricing.
So Is This a Buying Opportunity or a Warning?
Both cases exist, and they’re both serious.
The bull case: Software companies still have the data, the enterprise relationships, the integrations, and the distribution. They have 10-20 years of customer lock-in baked into workflows. AI doesn’t erase that overnight. At 22x forward earnings — now below the market — you are not paying for a growth premium anymore. If these businesses stabilize and grow at even moderate rates, the stocks are cheap.
The bear case: If you strip out seat expansion as the growth engine, and pricing power gets squeezed as buyers renegotiate, the earnings trajectory for many companies looks genuinely uncertain. A “fair” multiple on decelerating or declining earnings is not 22x.
The businesses winning right now are the ones that have become infrastructure for AI, not those competing with it. The ones losing are trying to reframe the same product as “AI-powered” without fundamentally rethinking how value is delivered.
That gap is going to widen over the next 18 months.
Software Has Never Been on Sale Like This Before. Ever.
Software has never traded at a discount to the S&P 500 in the modern era. That is either a historic buying opportunity or a sign that the market has correctly identified a structural break in the model.
Either way, it will resolve one direction or the other — and probably not slowly.
For founders: your next fundraise will be priced off public comps. If you are growing fast and capital-efficiently, you are fine. If you are counting on SaaS multiples recovering because they always have before, that assumption now requires more scrutiny than it used to.
The model was intact the last time they came back. That is the question worth asking today.
