The software sector has experienced a dramatic reversal over three months, shifting from a steep decline to a powerful rally as fears about AI's disruptive impact give way to a reassessment of its role as a growth driver.
Contrary to earlier doomsday predictions, artificial intelligence has not spelled the end for software companies. On June 1st, U.S. software stocks continued their strong rebound, with several firms reaching new all-time highs. MongoDB surged over 20%, Okta jumped 13.4%, while Snowflake, Oracle, and ServiceNow gained nearly 10%. IBM rose 7.6%, Adobe advanced 5.7%, and Palantir and Microsoft both climbed over 2%. The iShares Expanded Tech-Software Sector ETF (IGV) soared nearly 6%, capping a monthly gain of over 21%—its best monthly performance since October 2001.
This rally stands in stark contrast to the situation just three months prior, when the same group of software giants faced a panic over the notion that "AI will eat software," causing the IGV ETF to pull back 30% from its first-quarter high.
The rebound is underpinned by a confluence of factors: a series of earnings reports that exceeded expectations, confirmation of an industry trend where AI deployment is moving from "demonstration" to "production" scale, and a pivotal speech from Nvidia CEO Jensen Huang. In his address, Huang publicly refuted the "AI disruption theory," arguing that AI is leading to the hiring of more software engineers and that AI agents will create the largest opportunity ever for software companies.
AI is Revaluing the Software Sector
Throughout this year, the software industry has been at the epicenter of concerns about AI's impact. In February, the launch of Anthropic's AI coding tool, Claude Code, sent shockwaves through capital markets. Fears emerged that if AI could directly handle business tasks—coding, report writing, analysis, data organization—what value would traditional software hold? Would the pricing power of SaaS companies be eroded by AI?
The narrative that "AI is killing software" gained significant traction, triggering a broad sell-off in global software stocks. IBM's stock, for instance, suffered its worst single-day drop in 25 years.
The turning point arrived with the latest earnings season. The market began to re-examine software companies' actual revenues, order books, and the commercialization of AI.
Snowflake served as a major catalyst for the recent software rally. In late May, the data cloud giant reported first-quarter revenue growth of 33% to $1.39 billion, surpassing market expectations. Its remaining performance obligation (RPO), a key indicator of future revenue, reached $9.21 billion, a 38% year-over-year increase. Furthermore, Snowflake announced a five-year, $6 billion infrastructure cooperation agreement with Amazon AWS, sending its stock up nearly 50% over two trading sessions.
The deployment of AI agents is, in fact, creating stronger demand for data infrastructure like Snowflake's, as companies require massive storage, querying, and management of internal data. Snowflake disclosed that over 13,600 accounts are using its AI capabilities, with its Cortex AI service adopted in more than 7,100 accounts.
Salesforce Inc (NYSE: CRM)
Another closely watched company is Salesforce. Its latest earnings showed pressure from revenue guidance that was slightly below expectations, but its AI agent business delivered impressive figures. The annual recurring revenue (ARR) for its Agentforce platform has surpassed $1.2 billion, growing 205% year-over-year. Combined with its Data 360 business, the total ARR for AI and data products reached approximately $3.4 billion, representing over 200% growth. The company also bolstered market confidence by initiating a share buyback program.
Usage metrics further underscore the momentum: Agentforce and the Slack platform have cumulatively delivered 3.8 billion Agentic Work Units (AWU), a 111% sequential increase, and have processed over 28.6 trillion tokens to date.
Okta Inc (NASDAQ: OKTA)
Identity technology provider Okta is another hot software stock. While identity verification was a relatively stable software niche, the advent of AI has heightened the importance of identity security. Okta's latest quarterly report showed first-quarter revenue grew 11% year-over-year, with RPO reaching $4.719 billion, a 16% increase, signaling robust future revenue. Company executives noted that AI agents are rapidly becoming a new workforce within organizations, creating a wave of identities that must be protected and governed alongside human users.
Adobe Inc (NASDAQ: ADBE)
Adobe represents the application layer considered most vulnerable to AI disruption. The emergence of tools like Midjourney, Runway, and various generative image platforms initially raised fears of a fundamental challenge to Adobe's creative software suite.
However, Adobe's first-quarter results told a different story. Driven by AI, its core subscription revenue reached $6.17 billion, a 13% year-over-year increase. AI-related businesses showed strong growth momentum, with AI-related ARR more than tripling year-over-year. The ARR for its AI image generation tool, Firefly, exceeded $250 million. Adobe also raised its full-year revenue and profit outlook.
For enterprise clients, generative capability itself is becoming less scarce. The true differentiators are copyright safety, workflow integration, and enterprise-grade application capabilities, which is why Adobe consistently emphasizes "commercially safe AI."
Other notable performers include Oracle, whose backlog (RPO) skyrocketed to $553 billion, a 325% surge largely attributed to orders from OpenAI. Application software firm Datadog saw its first-quarter revenue surpass $1 billion for the first time and raised its annual revenue guidance, driven by demand for AI observability and security solutions. MongoDB reported a 25% year-over-year revenue increase to $687 million for its latest fiscal quarter, with its Atlas cloud database service performing strongly, leading the company to raise its full-year guidance; its RPO grew a substantial 88%.
Collectively, these earnings reports suggest that AI has not, as previously feared, rapidly eroded the business foundations of software companies. For many industry leaders, they are experiencing demand expansion driven by AI needs. In other words, AI is becoming a new revenue stream and growth narrative for the software sector. The highest certainty currently lies in AI's role within underlying infrastructure like data platforms, security governance, and observability, while validation within traditional application software layers is still ongoing.
Industry-Wide Transformation and Challenges
While software companies' earnings have countered the "AI吞噬论" (AI consumption theory) from an industry perspective, Nvidia CEO Jensen Huang's recent public comments directly addressed market anxieties about the software sector. Huang argued that the advent of generative AI will not cause software companies to fail but will instead lead to a vast proliferation of AI agents. These agents will require the use of more software tools than ever before, representing an incredible era for software companies. The key, he noted, is that software must be presented in a way that agents can utilize.
He cited data showing GitHub code commits grew from 300 million in 2023 to 500 million in 2025, and have already doubled in the first few months of 2026. He dismissed claims that AI reduces jobs as "complete胡说八道" (complete nonsense), stating that the number of software engineers is actually increasing.
In reality, the AI transformation is forcing a re-evaluation of the software industry's business models, value distribution, and development direction. Initial market panic stemmed from two premises: that AI would drastically increase individual productivity, reducing headcount and thus software license purchases, and that powerful AI tools could allow small teams to accomplish work previously done by large ones, lowering industry barriers.
However, large software companies, including IBM, are among the earliest and largest users and beneficiaries of these advanced AI tools. IBM itself has tens of thousands of developers using internal and partner AI tools across the entire development lifecycle.
Looking ahead, the software industry may undergo a "differentiation" in the AI era. Firms with "irreplaceable" capabilities in data infrastructure, security governance, and workflow collaboration are best positioned to benefit directly from the explosion in AI token consumption and achieve above-expectation growth. Conversely, thin application layers focused on standardized operations and human labor substitution find themselves in a more vulnerable position.
Analysis suggests the shift of AI from "demo-grade" to "production-grade" essentially places all IT infrastructure, once seen as "targets for AI consumption," back onto the critical path for large-scale AI deployment. The concurrent operation of massive numbers of AI agents requires compliant, secure, and high-frequency access to internal corporate data. As system complexity rises, corporate demand for data governance, real-time monitoring, permission systems, and workflow orchestration is being structurally amplified.
Persisting Concerns and Unresolved Questions
Despite the growth, market concerns are not fully alleviated, and industry challenges remain. For instance, while Salesforce and Adobe have rebounded recently, their year-to-date stock performance is still down around 20%. Salesforce's Q2 revenue guidance fell short of analyst expectations of $11.4 billion. The coming quarters are seen as a critical phase for validating the value of platforms like Agentforce.
Furthermore, as AI drives revenue growth, costs are also rising significantly. Questions remain about whether the software industry can maintain a high-margin growth path if AI expansion requires higher capital intensity, longer depreciation cycles, and greater energy investment.
Large language model (LLM) providers are also encroaching on traditional software company territory. The risk of direct competition is increasing as firms like OpenAI and Anthropic expand into downstream enterprise application layers, potentially challenging companies like Salesforce. The moats software companies built on product functionality are being rapidly eroded by AI. Future differentiation may hinge less on features themselves and more on data, workflows, and customer relationships.
The most significant challenge may come from the business model itself. Historically, enterprise software has been priced per user; more employees meant more software seats and higher revenue. As AI agents begin to take on knowledge work, investors worry this logic may change. This is a key reason behind the software sector's valuation re-rating this year. Analysts are increasingly debating whether the industry needs to shift from per-user pricing to models based on usage, tasks, or even outcomes. While this shift is not yet widely reflected in financial statements, the era of AI agents is forcing the entire software industry to reconsider its business models and value measurement methods.
Comments