JPMorgan joins the ranks of supporting US software stocks: Concerns over AI impact are exaggerated; a rebound is expected after a historic decline

JPMorgan strategists say that as the market has priced in unrealistic expectations of AI disrupting the software industry in the short term, software stocks are poised to rebound from their historic decline. Led by Dubravko Lakos-Bujas, the team of strategists pointed out that given the current “extreme price volatility,” investors should increase exposure to high-quality, AI-resilient software companies, at least in the near term, as there is a possibility of funds rotating back into this sector. In a report, they wrote: “Given that market positioning has been fully de-risked, overly pessimistic about AI’s potential to disrupt the software industry, and fundamentals remain solid, we believe the risk-reward balance is increasingly skewed toward a rebound.”

Recently, US software stocks have been under pressure due to concerns that new AI tools could impact traditional SaaS business models. This sell-off did not distinguish between companies that have established partnerships with AI firms or those with proprietary data assets, treating almost all related software companies equally.

The software sector has fallen to its lowest level since the “Doomsday” market turmoil in April last year

Microsoft (MSFT.US) and CrowdStrike (CRWD.US) are among the representative companies mentioned by JPMorgan strategists that demonstrate AI resilience, benefiting from AI-driven workflow efficiencies. The team noted that high switching costs and multi-year contracts in enterprise software provide a buffer against short-term shocks.

JPMorgan also pointed out that, in the long run, whether traditional software companies will be replaced by AI remains uncertain, but the current market’s pessimism about AI disrupting the industry is an “overreaction.” They added that overall, Q4 earnings reports from the software industry have been positive, and analysts expect profit growth of 16.8% in 2026.

This bullish view aligns with the assessment by Morgan Stanley strategists led by Michael Wilson. The team stated this week that US tech stocks still have room to rise, and the decline in software stocks has created an “attractive entry point.” Wilson wrote in the report: “Such volatility last week is not uncommon during major investment cycles. Nevertheless, the fundamentals of AI-enabled companies remain positive, and we believe the trading value of AI adopters is still undervalued.”

Last week, Dan Ives, a Wedbush analyst known as a “tech bull,” also said that although AI could pose some short-term pressure on traditional software business models, the market’s reaction to this risk is clearly excessive. The current sell-off in software stocks has already implied an extreme assumption of “industry-wide AI disruption,” which is not feasible in reality.

Ives pointed out that enterprise clients are more cautious about AI migration than the market thinks. Many companies are reluctant to expose their core data to immature new platforms just to chase AI benefits, nor will they easily abandon software infrastructure built over decades with hundreds of billions of dollars. He said, “AI is a headwind in the short term, no doubt, but the current market pricing, as if the software industry is facing ‘doomsday,’ is completely unrealistic.”

Wedbush emphasized that the current large enterprise software ecosystem contains trillions of data points, and emerging AI companies like OpenAI and Anthropic, in terms of data capacity and enterprise-grade security, are unlikely to fully take over these complex systems in the short term. This suggests that AI is more likely to be integrated as “embedded tools” within existing software platforms rather than replacing them entirely. Wedbush also pointed out that Microsoft, Palantir Technologies (PLTR.US), CrowdStrike, Snowflake (SNOW.US), and Salesforce (CRM.US) are the five most promising software stocks to hold during this “software winter.”

Nvidia (NVDA.US) CEO Jensen Huang also dismissed concerns last week that AI would replace software and related tools, calling such ideas “illogical.” Speaking at an AI conference in San Francisco hosted by Cisco, Huang said that the idea that AI would diminish the importance of software is misleading. He believes AI will continue to rely on existing software rather than rebuild foundational tools from scratch.

Huang stated, “There is a view that tools in the software industry are declining and will be replaced by AI… this is the most illogical thing in the world, and time will prove it.” He added, “Whether you are human or a robot, whether artificial or general-purpose, will you use tools or reinvent them? The answer is obvious: use tools… That’s why the latest breakthroughs in AI are about tool usage, because tools are designed to serve specific purposes.”

AI reshaping the value chain and future outlook

From the perspective of software engineering realities and SaaS industry structure, the narrative that “AI will replace the entire enterprise software stack” is an oversimplification that the market tends to extrapolate linearly. The “value density” of enterprise software is not only in interfaces and features but also in proprietary data, permissions/audit chains, compliance and accountability boundaries, system integration, SLAs and availability, change management, and organizational processes. These factors mean that even the most powerful large language models (LLMs) often require high-quality proprietary corpora, structured knowledge bases, controllable tool calls, and traceable outputs to operate effectively in production environments.

Looking at the underlying technological logic of AI tools and SaaS software, panic selling of software stocks does not mean “software is no longer needed.” Instead, the value chain is being redistributed by AI: more advanced general-purpose models and agentic AI workflows threaten to internalize many point-function SaaS features (feature absorption) or bypass UI and seats through “dialogue-based entry + automation,” impacting traditional seat-based pricing and renewal models. As a result, the market is more eager to categorize software stocks into “AI winners” and “AI losers.”

Conversely, leading enterprise SaaS providers’ “system record layers” (ERP/CRM/ITSM/databases/security/compliance) often have data sovereignty, governance, permissions, audit, and migration barriers. The reality is more likely that AI will transform these long-standing software giants into distribution channels for delivering AI capabilities, rather than replacing their entire existing infrastructure overnight.

According to Morgan Stanley’s strategists, in the short term, the sharp decline in the software sector has indeed sparked discussions about a “near-term bottom” in technical terms, with some funds slightly increasing positions. However, most investors are still waiting for “hard catalysts” that can turn AI narratives into actual revenue growth—such as software companies reporting AI-related product revenues or penetration, enterprise clients announcing large-scale deployments, or renewal metrics (net retention, expansion rates) improving significantly after integrating large AI models or agentic AI agents. Without such evidence, the rebound is more likely a “oversold correction” rather than a new trend.

This wave of selling in software stocks is more like the market answering a new question with extreme measures: to what extent will profit pools for SaaS vendors be redistributed to “model factories + agents”? In the short term, the answer can only be verified through two “hard indicators”: (1) the actual deployment and paying adoption rate among enterprise clients; (2) the elasticity of SaaS vendors’ AI-related product revenues and renewal/net retention—similar to what some buy-side representatives in the Thomson Reuters Breakingviews report are waiting for: actual revenue growth data from AI-related products or more enterprise deployment announcements as catalysts for increased positions.

Prior to this, the volatility in software stocks is likely to continue: on one hand, technical oversold rebounds may occur; on the other, capital will continue to shift structurally—favoring vertical software/data asset companies with strong data and workflow stickiness related to AI training/inference systems, and platforms capable of deploying AI in a “controllable, auditable, integrable” manner. Companies with weaker moats, higher homogeneity, and more expensive valuations in the application layer will continue to demand higher risk premiums. Therefore, software giants like Microsoft, MongoDB, Snowflake, Palantir, and SAP—who possess substantial data assets and solid fundamentals—are more likely to rebound strongly after panic-driven declines.