On January 12, 2026, Anthropic launched Claude Cowork. A journalist built a kanban board in under 10 minutes and posted a video. Monday.com‘s market cap dropped $300 million before the session closed. That single tweet — not a product launch, not an earnings miss, just a demonstration — erased roughly a quarter-billion dollars from a company generating $1.3 billion in annual recurring revenue.
By the time February ended, approximately $1 trillion in aggregate market capitalisation had been wiped from enterprise SaaS. Media coined “SaaSpocalypse.” It generates heat, but not much light.
The full picture of what the SaaS reckoning means for technology leaders needs more than a headline. This article gives you the shared vocabulary and market context: a structured account of what happened, why it happened, and what it means — with real numbers, named triggers, and the frameworks to think clearly about it. By the end, you’ll understand why “SaaS is dying” and “SaaS is fine” are both wrong, and what the accurate framing actually looks like.
What actually happened to enterprise software stocks in January and February 2026?
The January–February 2026 SaaS dislocation erased approximately $1 trillion in aggregate market capitalisation from enterprise SaaS, with the S&P 500 software index shedding that amount since January 28, 2026. Broader estimates run closer to $2 trillion, but the enterprise SaaS figure is the more defensible anchor.
The individual stock moves tell the story. HubSpot declined approximately 51% from peak to trough — from roughly $880 per share to around $233, with market cap collapsing from $42 billion to under $10 billion. Monday.com fell approximately 44%. ServiceNow declined approximately 36%. Atlassian dropped 26.9% in eighteen trading days. Workday was down approximately 13% year-to-date.
The iShares Expanded Tech-Software ETF (IGV) — a broad SaaS proxy — declined approximately 22% year-to-date, marking the steepest software sell-off since the 2022 rate hike cycle. The losses weren’t a slow bleed. The majority were compressed into two sharp sell-offs in January and one in early February. Jefferies equity trader Jeffrey Favuzza dubbed it “SaaSpocalypse” and described trading that was “very much ‘get me out’ style selling” — language not heard since 2008.
While SaaS haemorrhaged, semiconductor plays surged — Lam Research up 30.3%, KLA Corp up 29%, Applied Materials up 27.3%. Capital wasn’t leaving tech; it was moving within tech. That’s the first signal this was structural, not a general correction.
Why did the SaaSpocalypse happen when it did?
The sell-off had two convergent triggers in January 2026, each compounding the other.
The first was the Anthropic Claude Cowork launch on January 12. Claude Cowork positioned Claude as an enterprise collaboration layer capable of executing multi-step SaaS workflows autonomously — not a new AI feature added to an existing tool, but a potential replacement layer for the per-seat model that underpins SaaS valuations. On January 30, Anthropic released 11 open-source Cowork plugins covering legal, finance, marketing, sales, and customer support. The plugins collapsed a significant portion of the SaaS stack into a single AI system: productivity, marketing, finance, and data workflows all replicable through one agent layer.
The timeline: January 16 — Anthropic launches multi-agent shared workspace, Atlassian and Asana drop ~4% immediately. January 20 — tech companies announce hiring freezes citing AI. January 28–29 — ServiceNow’s earnings with guidance language acknowledging AI substitution risk coincides with the OpenAI Frontier launch, which released a diagram showing value accruing to AI agents above the SaaS layer. ServiceNow drops 11% in the session.
CNBC host Deirdre Bosa’s Monday.com kanban tweet is the clearest illustration. She built a functional kanban interface using Claude Cowork and said she wanted to “try to recreate Monday.com.” Monday’s stock dropped 6% immediately and another 10% the next session — $300 million erased from a company generating over $1 billion in annual revenue.
The proximate triggers were real, but they landed on a market already repricing 18 months of enterprise AI data. Menlo Ventures documents enterprise AI spending growing from $1.7 billion in 2023 to $37 billion in 2025 — a 3.2× year-on-year growth rate. The triggers provided confirmation, not news.
For the mechanism-level breakdown of how AI agents actually threaten SaaS business models, that article covers the specifics.
Is SaaS really dying — or is something more nuanced happening?
The growth assumptions that justified 20–40× revenue multiples are no longer credible, and the repricing reflects that. SaaS products still work. Contracts still renew. But the budget that would have funded SaaS growth is now flowing into AI. That’s what this repricing is about — a contraction in growth expectations, not a collapse of the business model.
HarbourVest characterised the sell-off as “rational repricing” — AI is undermining three assumptions baked into SaaS multiples: that seat-based pricing will grow forever, that software margins are structurally fixed at 80–85%, and that recurring revenue is predictable. Morgan Stanley called it a “sentiment-driven dislocation.” Both are partially right. They’re not mutually exclusive.
Public SaaS growth rates had declined every single quarter since the 2021 peak. AI budgets are up 100%+ year-on-year while overall IT budgets are up ~8%. AI is absorbing the growth margin from total IT spend — and that margin was previously flowing into SaaS expansion.
The relevant question isn’t whether SaaS survives as a category. It’s which SaaS products are being starved, and at what rate.
What is the “uncertainty tax” — and why does it matter beyond the stock market?
The uncertainty tax is the valuation discount investors apply to SaaS businesses whose revenue model is perceived as structurally threatened by AI — a premium charged for unpredictability in ARR, margin, and net revenue retention when the 5-year model is opaque.
It’s rational. When a SaaS company’s revenue model depends on seat-count growth, and AI agents can perform the same workflows without seats, even a 10% long-term seat reduction assumption changes the entire discounted cash flow. HarbourVest documents the maths: 30% seat decline plus 10% price increase produces -23% revenue contraction; 50% seat decline plus 15% price increase produces -42.5%. AI forces a decoupling between value delivered and seats billed — value goes up but the monetisable unit goes down.
The Monday.com kanban incident illustrates this precisely: ARR didn’t change that day, but $300 million in market cap was erased — because investors updated their probability on future seats. Goldman Sachs analyst Ben Snider framed it clearly: “near-term earnings results will be important signals of business resilience, but in many cases insufficient to disprove the long-term downside risk.”
The tax falls most heavily on horizontal SaaS — project management, CRM, work-OS tools — and least on system-of-record platforms with deep integration lock-in.
For CTOs, this matters operationally. A vendor carrying a significant uncertainty tax faces constrained R&D, potential pricing aggression, and elevated acquisition risk. All of which surface in renewal conversations before anything dramatic happens on the exchange.
What does the data say about AI’s actual growth versus SaaS decline?
The enterprise AI spending data isn’t speculative. Menlo Ventures documents $1.7 billion in enterprise AI spending in 2023, rising to $11.5 billion in 2024, then to $37 billion in 2025 — a 3.2× year-on-year growth rate representing more than 6% of the entire software market within three years of ChatGPT‘s launch. AI-native startups captured approximately 63% of AI application layer revenue in 2025, up from 36% in 2024.
The Gartner 2030 prediction, cited through Deloitte Insights: at least 40% of enterprise SaaS spend will shift toward usage-, agent-, or outcome-based pricing by 2030. Seat-based pricing had already fallen from 21% to 15% of vendors as a primary model within twelve months, while hybrid pricing surged from 27% to 41%. These are current market data, not forecasts.
Gartner separately predicts 35% of point-product SaaS tools will be replaced by AI agents or absorbed within larger agent ecosystems by 2030. Deloitte found that 57% of enterprise respondents were allocating 21–50% of their digital transformation budgets to AI automation, with 20% allocating more than half.
Together, the datasets establish that the reckoning is structural — the demand side is genuinely migrating, and the supply side is already following. For a deeper treatment of the pricing model shift, that article covers the transition in full.
What is the K-shaped bifurcation — and which side of it are your vendors on?
The K-shaped bifurcation describes the post-dislocation divergence in SaaS: two arms separating rather than one uniform decline.
Platform incumbents — Salesforce, Oracle, Microsoft — are trending toward recovery. Horizontal point-solution SaaS — HubSpot, Monday.com, and in important respects Workday — continues to face downward pressure. HarbourVest’s analysis identifies the structural logic.
Systems of record hold proprietary customer-specific operational data, run deterministic mission-critical workflows, carry high switching costs, and are necessary for regulatory, financial, or operational continuity. AI augments these systems; it doesn’t replace them. If SAP goes down, factories stop. If Workday breaks, payroll fails.
Bolt-on tools have the opposite profile: they don’t own the data layer, perform tasks where “good enough” is acceptable, have low switching costs, and solve problems AI can replicate with general-purpose models. The tell: if Claude Cowork’s plugins can replicate a vendor’s core value proposition out of the box, that vendor’s moat was a feature wrapped in a subscription.
Workday is the nuanced case. Despite platform positioning, its HR and finance workflows are increasingly automatable at the process layer. Platform status is necessary but not sufficient — the workflows must also be irreplaceable by agents.
The valuation divergence between the two arms is pronounced. AI-native startups commanded 50× higher valuations than traditional SaaS at Series D. Harvey (legal AI) trades at 80× revenue. HubSpot trades at ~6×. The market is pricing growth assumptions, and those assumptions differ structurally between the two arms.
Understanding which arm your vendors sit in is the starting point for the practical work. The vendor evaluation framework covers that in full.
What does the SaaS reckoning mean for technology leaders right now?
The market repricing is a leading indicator of structural change, not just a financial event. It has operational consequences for every organisation managing a software stack.
The vendor financial health implication is direct. A vendor carrying a meaningful uncertainty tax faces constrained R&D, potential pricing aggression, and elevated acquisition or partnership risk. The high leverage of private equity-backed SaaS businesses inhibits reinvestment precisely when they need to spend aggressively to stay competitive. Vendor financial stress is an operational risk for buyers.
The pricing model transition is already arriving in contracts. Deloitte notes usage-based and outcome-based terms are already appearing in enterprise agreements. The Gartner 2030 prediction isn’t a distant horizon — it’s the direction renewal negotiations are already moving.
The build-vs-buy calculus has shifted. AI-assisted internal development has made custom tooling viable for more organisations. Build time now measures in days to weeks for point-solution tools, at a developer plus an AI subscription ($50–200 per month), compared to $500,000 or more annually for enterprise SaaS licences. That shift increases your negotiating leverage even where there’s no genuine intention to build.
For the mechanism of how AI agents attack SaaS models, that article covers the specifics. For vendor evaluation, the framework article covers that in full. For the full picture of what the SaaS reckoning means for technology leaders, that is the right place to continue.
Frequently Asked Questions
What is the SaaS reckoning?
The January–February 2026 broad repricing of enterprise SaaS equities. Triggered by Anthropic’s Claude Cowork launch on January 12, 2026 and compounded by ServiceNow’s earnings on January 28–29. Approximately $1 trillion in aggregate SaaS market capitalisation erased across six weeks.
What does “SaaSpocalypse” mean?
Media shorthand coined by Jefferies equity trader Jeffrey Favuzza and picked up by Forbes and TechCrunch. The term overstates the finality. “SaaS reckoning” is more precise — SaaS is being structurally repriced, not eliminated.
Why did HubSpot stock fall so much in 2026?
Two factors: Claude Cowork directly threatened HubSpot’s CRM and marketing workflow seat model, and HubSpot’s position as a horizontal SaaS tool with limited system-of-record defensibility made it a candidate for the highest uncertainty tax.
Is SaaS really dying because of AI agents?
SaaS isn’t dying — it’s being starved. AI agents threaten the seat-based growth assumptions that justify high-multiple valuations, but the operational software layer doesn’t disappear quickly. The question is which categories face the most acute substitution pressure.
What is seat-based pricing and why is it under threat?
Per-named-user licences where revenue scales with headcount. AI agents can execute the same workflows without occupying seats — breaking the growth assumption that more employees means more ARR. Per-seat pricing already dropped from 21% to 15% of SaaS companies as a primary model in twelve months.
What did Anthropic’s Claude Cowork actually do?
An enterprise collaboration layer capable of executing multi-step SaaS workflows using 11 open-source plugins covering legal, finance, marketing, sales, and customer support. Investors read it as a replacement layer for per-seat pricing across the SaaS category — not a feature, a structural challenge.
What does the Gartner 2030 SaaS pricing prediction say?
Via Deloitte Insights: by 2030, at least 40% of enterprise SaaS spend will shift toward usage-, agent-, or outcome-based pricing. Seat-based pricing is already declining from 21% to 15% of vendors in just twelve months.
What does the Menlo Ventures $37 billion figure mean?
Enterprise AI spending in 2025, from the Menlo Ventures State of Generative AI in the Enterprise report. Up from $1.7 billion in 2023 and $11.5 billion in 2024 — 3.2× year-on-year growth. Enterprise AI adoption is measured and accelerating.
How does the 2026 SaaS sell-off compare to the 2001 dot-com crash?
The 2001 crash was speculative valuations on businesses with no revenue model. The 2026 reckoning is a repricing of businesses with proven revenue models whose growth assumptions are genuinely in question. SaaStr’s characterisation: “2016 was cyclical — 2026 is structural.”
What is the K-shaped bifurcation in SaaS?
Post-dislocation divergence where platform incumbents with system-of-record status (Salesforce, Oracle, Microsoft) trend toward recovery, while horizontal point-solution SaaS (HubSpot, Monday.com) continues to face downward pressure. The split is structural — data ownership, integration depth, and workflow substitutability.
Why are all my software stocks going down in 2026?
Investors repriced the category when AI agents demonstrated they could replace the workflow tasks that per-seat licences support. The stocks that fell hardest are those most dependent on seat-count growth as their revenue engine.
What should technology leaders do about the SaaS valuation crash?
Understand which of your vendors sit in the downward arm of the K-shaped split. Anticipate pricing model changes in renewal conversations. Assess whether AI-assisted internal tooling has made build alternatives viable for point-solution tools. The vendor evaluation framework covers this in full.
