Three software companies posted double-digit revenue growth this quarter. Their stocks are down 80% or more.

Figma: revenue up 41% year-over-year to $1.06 billion, stock down 86.5% from its 52-week high. Duolingo: revenue up 39% to $1.04 billion, free cash flow of $360 million, stock down 83.3%. Monday.com: trailing revenue of $1.23 billion with $310 million in free cash flow, stock down 80.2%.

These aren’t broken businesses. They’re growing, profitable or near-profitable companies whose stocks have been gutted on the expectation that AI will eventually destroy their business models — before it actually has.

That’s a new mechanism. In earlier posts, we tracked the concrete stuff: seat compression at Atlassian, Block cutting 40% of its workforce, Oracle cannibalizing its own SaaS divisions. Real revenue impact. Real headcount reductions. This week, the market is doing something different: pricing in destruction that hasn’t happened yet, and punishing growth as if it’s already irrelevant.

IGV, the software sector ETF, closed at $76.64 on April 9 — down 35% from its October peak and 24% year-to-date. For context, the S&P 500 is roughly flat. Software stocks have collectively shed over $1 trillion in market value since the “SaaSpocalypse” began in February, and the rout has only deepened through April.

The Distribution Channel Breaks First

Monday.com’s collapse tells a more specific story than “AI kills SaaS.” The company’s stock crashed 79% after a February guidance reset that lowered 2026 revenue growth to 18-19%, roughly $40 million below the $1.5 billion threshold management had called “confident” at its September investor day.

The cause wasn’t AI replacing Monday.com’s product. It was Google’s search algorithm changes destroying Monday.com’s self-serve acquisition channel. Co-CEO Roy Mann told investors that “the cost to acquire and expand self-serve customers have increased over the past year, and the returns on those investments have been below historical levels.” The no-touch SMB funnel — the engine that made Monday.com a growth story — broke because Google’s AI-powered search results stopped sending the same organic traffic.

This is a second-order effect the hypothesis didn’t predict. AI isn’t replacing Monday.com’s project management tool. It’s replacing the distribution economics that made Monday.com’s growth model work. Google’s AI summaries answer the query that used to drive a signup. The product still works fine; the customer never arrives.

Meanwhile, Monday.com’s enterprise cohort — customers paying over $500,000 ARR — grew 74% year-over-year, a record. The bifurcation we’ve tracked in credit markets and public valuations now runs through individual companies: enterprise sales-led motion intact, SMB self-serve motion collapsing. Same product, two completely different stories.

The Platform Player Goes Proprietary

Meta launched Muse Spark this week, the first model from Meta Superintelligence Labs — and it’s proprietary. Not open-source. Not open-weights. The company that spent years positioning Llama as the open alternative to OpenAI has now built a closed consumer AI that will replace Llama as the default model powering WhatsApp, Instagram, Facebook, and Meta’s smart glasses.

Llama 5 shipped the same day as an open-source release — so Meta now runs a dual strategy. Open models for developers. Proprietary models for consumers. The separation is deliberate: Muse Spark is designed as a consumer agent layer, with subagents that plan trips, compare products, and automate multi-step tasks. Shopping mode “surfaces ideas from the creators and communities people already follow.”

For the SaaS hypothesis, this matters because Meta is building exactly the kind of dispatch layer we discussed last week — except this one sits on top of 3.3 billion monthly users. When Meta AI can plan your trip, compare products, and execute purchases inside WhatsApp, the individual travel SaaS, the comparison tool, and the booking platform all become backend services that the user never directly touches. The AI agent becomes the customer, and the human becomes the beneficiary.

Meta plans to spend $115-135 billion on AI in 2026 alone, including a $21 billion deal with CoreWeave. That’s not a research project. That’s a platform bet on owning the consumer interaction layer — the same layer that currently generates SaaS subscription revenue.

Hypothesis Update

The phantom repricing complicates the hypothesis in an interesting way. The claim was that AI commoditizes software and collapses the SaaS business model. What’s actually happening — at least in public markets — is that the expectation of commoditization is destroying valuations before the commoditization arrives. Figma grew revenue 41% and lost 86% of its market cap. The market isn’t waiting for the evidence; it’s front-running the thesis.

That means the hypothesis generates real economic consequences (companies lose access to capital, can’t make acquisitions, face hostile boards) even in the period before it’s fully verified. The repricing is the mechanism, not the replacement.

Monday.com’s channel collapse adds a wrinkle: AI may destroy SaaS businesses not by replacing their products but by disrupting their distribution. If Google’s AI answers replace the search results that drove self-serve signups, the entire SMB SaaS acquisition model — which depends on organic search — breaks from the demand side.

And Meta’s Muse Spark makes the dispatch layer concrete. When the AI agent sitting on 3.3 billion users goes proprietary, the question “who owns the customer when AI mediates the transaction” gets a specific answer: Meta does.

Verification signals strengthened: valuation compression accelerating (IGV -35% from peak); distribution channel disruption emerging as new mechanism; platform-layer consolidation (Meta proprietary agents).

Falsification signals noted: the companies being repriced are still growing 18-41% — the revenue hasn’t collapsed yet. Market sentiment may be overshooting.

What to watch: Q1 2026 earnings season starts in two weeks. The gap between revenue growth and stock price will either close (growth decelerates to match the repricing) or widen (growth holds, and the market was wrong). That’s the test.

Sources: 247 Wall St, TIKR, BlackRock, Meta Newsroom, Forbes, Anadolu Agency