AI Agents Redefine SaaS Valuation Logic: Why ServiceNow, Salesforce, and Adobe Have Faced a Reshaping of the Industry?

On July 24, 2026, the S&P 500 index closed at 7,411.98 points, still posting a positive return year-to-date. The Nasdaq 100 index is up about 15% for the year. The Philadelphia Semiconductor Index has risen by more than 60% cumulatively since 2026.

However, while the broader market and the hardware sector are charging ahead, a once highly sought-after segment is undergoing a brutal valuation reset—software-as-a-service (SaaS) stocks.

A software-as-a-service stock index is down 27% in 2026, while the broader software index is down 11%. iShares Expanded Tech-Software Sector ETF (IGV), which tracks the software industry, has fallen more than 27% year-to-date, closing at its lowest level since November 2023. Some SaaS stock indices’ cumulative declines have already approached 40%.

This is not a routine sector rotation. This is the market redefining the value of software.

AI Agent: from “auxiliary tool” to “replacement”

To understand the logic behind this selloff, you have to go back to the origin of the SaaS business model.

Over the past decade, the growth formula for SaaS companies was very straightforward: software subscriptions → long-term enterprise payments → stable revenue growth. Businesses buy software seats for each employee and pay annually, giving software companies predictable recurring revenue. The more users there are, the higher the revenue; the more features there are, the higher the pricing.

But the rise of generative AI and AI Agents is dismantling this formula.

In a report released on July 1, 2026, Gartner explicitly stated that agentic AI will fundamentally change how enterprise software generates revenue. From now to 2030, up to $234 billion in enterprise application software spending will face disruption from agent arbitrage; by 2030, this impacted portion will account for about 20% of total enterprise application SaaS spending.

What is “agent arbitrage”? Gartner’s explanation is straightforward: AI Agents can complete tasks across systems, reducing the need for users to interact with multiple traditional software interfaces. Enterprises no longer need to buy a separate set of independent SaaS products for each business process—one unified AI Agent can call different systems, read data, and execute tasks.

George Brocklehurst, Managing Vice President at Gartner, said: “Agent systems deliver outcomes directly, bypassing traditional user-experience-centric applications, making software invisible. This breaks the link between user growth and revenue growth for many enterprise software vendors.”

In other words, the biggest impact of AI may not be creating new companies, but redefining the value of old ones.

July 2026: OpenAI Presence triggers another round of selloff

This selloff didn’t start in July, but a key catalyst emerged in July.

On July 22, 2026, OpenAI released its enterprise AI Agent product Presence. This is not a tool that merely provides large-model capabilities—it’s a deployment product suite for AI Agents tailored to enterprise business processes. Enterprises can configure data permissions, business rules, operating scope, approval workflows, and human takeover conditions for AI Agents, and plug it into scenarios such as customer service, sales, insurance claims, and enterprise IT services.

OpenAI disclosed that Presence has been used for its English phone customer service, and after launch it can resolve 75% of incoming calls without human intervention. BBVA, SoftBank, and Australia’s insurer IAG are also testing similar capabilities.

Market reaction was rapid and severe. Since Presence was released, HubSpot’s stock price has fallen 12.7% cumulatively, Atlassian down 11.8%, Workday down 9.9%, and Salesforce down 7.7%. TD Cowen analysts believe Presence’s release is the “main reason” for IGV’s software index selloff of 3% and its further drop the next day.

In a note, TD Cowen analysts wrote: “Presence is promoting many AI Agent features that many SaaS vendors advertise, including integration of large language model reasoning, data access, and governance and policy controls.”

The market’s core worry isn’t that OpenAI will immediately replace all enterprise software; it’s that enterprises may no longer need to buy separate SaaS products for each business process.

Four tech giants, four kinds of trouble

ServiceNow (NOW): earnings hold up, stock price halves

ServiceNow’s stock price fell from its 52-week high of $210.20 to about $103, a decline of roughly 51%. In the first half of 2026, the stock is already down 36%.

But ServiceNow’s fundamentals aren’t bad. In Q1 2026, subscription revenue grew 22% year over year to $3.67 billion. Current remaining performance obligations (cRPO) grew 22.5% year over year to $12.64 billion. The company also completed 16 deals with net new annual contract value over $5 million each, nearly an 80% increase.

ServiceNow’s AI product Now Assist has annual contract value exceeding $1 billion. But investors are still worried—if AI Agents can automate enterprise workflows, how much value is there left in ServiceNow’s workflow automation product?

Intuit (INTU): panic over AI tax tools, market cap evaporates by more than half

Intuit’s stock price fell from its 52-week high of $813.70 to about $293.82, down roughly 64%. Year-to-date, it’s down about 55%. There are reports that its decline in the first half of 2026 is close to 60%.

The market’s logic is straightforward: AI can automate tax filing, so consumers and businesses may no longer need to buy TurboTax.

But Intuit’s financial data is also solid. In Q3 2026, revenue was $8.56B, up 10.4% year over year. TurboTax revenue grew 7% year over year, and overall company revenue grew 10%. Full-year revenue is expected to grow 13% to 14%. TurboTax Live (a service connecting professional tax preparers) is expected to account for more than 50% of total TurboTax revenue, with year-over-year growth of 38% for the fiscal year.

However, on July 21, Morgan Stanley downgraded Intuit’s rating from “overweight” to “equal to the market,” and cut its target price sharply from $580 to $335. Analysts believe the debate over whether AI disrupts the TurboTax business will persist.

Salesforce (CRM): the worst-performing stock in the Dow

Salesforce is down 40.9% in the first half of 2026. As of mid-July, it’s down 35.2% year-to-date, the worst-performing component stock in the Dow Jones Industrial Average.

Like ServiceNow and Intuit, Salesforce’s results have not collapsed. In Q3 2026, revenue was $10.26 billion, up 8.6% year over year. But what worries the market is whether its Agentforce AI product can drive an inflection point in organic growth. A Morgan Stanley analyst pointed out that its key performance indicators “have not yet driven an inflection point for organic growth, because headwinds from the traditional product mix remain.”

Wall Street’s bearish sentiment toward Salesforce has reached its lowest level since 2012.

Adobe (ADBE): the cost of AI transformation

Adobe’s stock price fell from its 52-week high of $376.16 to about $227.16, a decline of about 40%. Down about 36.5% year-to-date.

Adobe faces a unique dilemma: the company admitted it is “actively sacrificing short-term subscription growth in pursuit of a larger AI user base.” AI-priority annual recurring revenue grew three times year over year, exceeding $500 million. But investors question whether these free AI users can be converted into meaningful recurring revenue.

In July, Morgan Stanley downgraded Adobe, saying: “As disruptive generative AI controversies intensify and obscure the path for re-accelerating annual repeatable bookings revenue, Adobe’s simultaneous transition across its freemium model, leadership changes, and reinvestment increases execution risk.” Adobe’s overall recommendation rating has fallen to 3.3 (out of 5), the lowest level since the 1990s.

The bigger picture: enterprise IT budgets are being rebuilt by AI

The collective crash in SaaS stocks isn’t an isolated event. It’s part of a broader structural reshaping.

On July 14, 2026, IBM pre-released preliminary results for the second quarter. Revenue was about $17.2 billion, far below analysts’ expectations of $17.58 billion. IBM said it hadn’t anticipated that customers would shift their spending away from IBM products to focus on server, storage, and memory products used in AI computing. The company’s stock price plunged 25% in a single day, the biggest one-day drop since the 1960s. IBM cut its 2026 revenue growth outlook from above 5% to 4%–5%.

Bryan Malbury, chief market strategist at Zacks Investment Management, said: “This isn’t just a speed bump; it’s like a sledgehammer. It’s not that there’s no demand for other services—demand exists. It’s just that they can’t make it work because the total capital expenditures that can be spent are limited.”

In a report dated July 14, industry research analyst Anurag Rana wrote: “While the gap may be due to IT budget shifting, this fundamental weakness could be interpreted as AI disrupting the industry’s core business, which may bring further valuation pressure.”

Wall Street has been trimming expectations for the software and services sector. The industry is expected to achieve 16.5% earnings growth in 2027, and this market consensus has declined steadily for seven straight weeks.

Valuation reset in progress: from seat fees to pay-for-outcomes

At the heart of this selloff is the market redefining the valuation anchor for software.

Traditional SaaS valuation models are built on several core assumptions: the number of users keeps growing, average revenue per customer steadily rises, and customer lifetime value is predictable. But when AI Agents can replace human operation of software, these assumptions begin to loosen.

Gartner noted that enterprise buyers will “no longer emphasize buying more new tools or dashboards.” Instead, “they want better outcomes, and adding more AI capabilities often adds costs rather than better outcomes.”

That means the value of software is shifting from “features” to “outcomes.” In the past, enterprises paid for software seats; in the future, enterprises may only pay for the actual tasks completed by AI. Bloomberg’s analysis points out that AI Agents could compress software industry profit margins by 100 to 900 basis points.

Brocklehurst describes this shift as a “redefinition of the end of SaaS”—“not so much an end as a metamorphosis.” SaaS won’t be destroyed, but it will be reborn in a different form.

For traditional software companies, this means they must move from interface-based value to outcomes-based value—embedding agent capabilities at points of delivery to defend their position in the value chain.

Conclusion

The SaaS stock plunge in 2026 isn’t the market punishing bad performance—it’s the market repricing an old business model. ServiceNow’s subscription revenue grew 22%, and Intuit’s revenue grew 10%, but data that used to be seen as a “moat” has lost its persuasiveness in the face of an AI narrative.

AI’s biggest impact may not be creating new companies, but redefining the value of old ones. When software changes from “a tool that requires humans to operate” into “an Agent that completes tasks automatically,” the cornerstones of the SaaS industry—seat fees, subscription fees, and feature-module fees—are being dismantled one by one.

For investors, this valuation reset means the SaaS industry’s valuation anchor is shifting from “number of users × average revenue per customer” to “outcomes × efficiency.” For practitioners, it suggests that the next decade of “software as a service” may no longer be called SaaS—it could be called “agent as a service,” or something else that hasn’t been named yet.

But one thing is certain: the old map can’t find the new land. What the SaaS industry is going through isn’t a routine pullback—it’s an underlying paradigm shift.

FAQ

Q1: Why is the S&P 500 rising, but SaaS stocks are crashing?

Capital is flowing massively out of traditional software segments into areas related to AI infrastructure. The 2026 Philadelphia Semiconductor Index is up more than 60%, while the SaaS stock index is down 27%. The market believes AI hardware is a deterministic beneficiary, while software companies face the risk that their business models are being disrupted by AI Agents.

Q2: How exactly do AI Agents disrupt the SaaS business model?

Traditional SaaS charges by user seats, and more users means higher revenue. AI Agents can automate tasks across systems, reducing enterprises’ need for multiple independent SaaS products. Gartner expects that by 2030, about 20% of enterprise application SaaS spending will be impacted.

Q3: If ServiceNow’s performance is still growing, why is the stock down 51%?

The market is trading expectations rather than historical data. ServiceNow’s Q1 subscription revenue grew 22%, but investors worry that AI Agents will automate enterprise workflows in the future, eroding demand for its core products. The stock fell from its 52-week high of $210.20 to about $103.

Q4: When does the valuation reset in the SaaS industry end?

It depends on two key variables: first, how quickly and effectively AI Agents penetrate enterprises; second, whether traditional software companies can successfully transition to an “outcome-oriented” pricing model. Gartner believes this is a “metamorphosis” process that continues through 2030.