JPMorgan drops blunt take on the AI ​​threat to software stocks

Software stocks have had a rough few weeks. The sector has lost roughly $2 trillion in market value since the start of 2026, dragging down some of the most recognized names in enterprise technology.

Sales accelerated after Anthropic launched new industry-specific plug-ins for its Claude agent platform, sparking fears that autonomous artificial intelligence tools could erode the core businesses of traditional software companies.

Salesforce (CRM) and Adobe (ADBE) bore the brunt, with the iShares software ETF (IGV) trading more than 20% below its 200-day moving average at its worst selling level.

JPMorgan is pushing back hard. The bank’s strategists say the market has overreacted and that investors willing to look past the noise may be staring at one of the better entry points in years for quality software names.

JPMorgan’s latest private note says investors are considering worst-case scenarios for AI disruption that are unlikely to materialize anytime soon.

The damage is historic. The software sector’s weight in the S&P 500 has fallen from 12% to 8.4%, and the relative strength index of the S&P 500 Software Index recently reached 18, a figure last seen at the end of the dot-com crash.

According to JPMorgan, the complete replacement of business software by AI will only happen after 2028 at the earliest. Current tools enhance workflows. They don’t replace them. The fear is real. The timeline investors attach to it is not.

One factor that investors seem to be ignoring is how deeply embedded enterprise software really is. Multi-year contracts, high switching costs and mission-critical workflows aren’t going away just because a new AI tool was launched last month.

Wedbush Securities analyst Dan Ives called this the most structurally baffling sell-off he’s seen on Wall Street in 25 years. In a post on

The irony, JPMorgan notes, is that the software companies that investors are fleeing from are themselves becoming beneficiaries of AI. Business AI users are already seeing the positive impact in their margins, with net margins for companies adopting the S&P 500 AI reaching 16.4%, compared to 13% for companies outside that group as of February 2026.

The bank lays out a clear framework for why the risk-reward in quality software names has shifted in favor of buyers:

• Timeline mismatch. The complete replacement of Software-as-a-Service (SaaS) workflows with AI agents is a post-2028 story. Copilot-style features dominate enterprise AI today, not large-scale replacements.

• Sticky earnings. Leading SaaS companies have a net retention rate of almost 90%. Customers don’t just stay, they increase their spending year after year.

• AI as a tailwind. Software companies are integrating AI to reduce their own costs and increase margins, making the technology a growth engine rather than just a threat.

• Rating reset. Forward earnings multiples have compressed from about 40 times to about 25 times. The free cash flow yield on quality names has risen above 4%.

The strongest counterargument to the doomsday scenario may be the figures from Salesforce itself. The company’s fourth quarter results show that Agentforce’s annual recurring revenue was $800 million, up 169% year over year, with a total of 29,000 deals closed.

More AI stocks:

That doesn’t look like a company is going bankrupt. Salesforce ended fiscal 2026 with total revenue of $41.5 billion, up 10% year over year, and expected fiscal 2027 revenue of $45.8 billion to $46.2 billion.

Meanwhile, Anthropic itself is deepening its software partnerships rather than dismantling them. The company’s Claude and Slack integration frames its AI as a productivity layer built on top of existing tools, not as a replacement for them.

JPMorgan does not dismiss the risks. Agentic AI could mature faster than expected. Startups without overhead can erode the margins of established players. A broader economic slowdown would put pressure on information technology budgets across the board.

However, the bank’s basic scenario is that fundamental factors will reassert themselves. Short interest in software is at record levels and sentiment has swung to an extreme level that has historically preceded a recovery, according to the JPMorgan analysis.

For investors with the patience to look past the fear, JPMorgan’s message is clear. In addition to the speculation, the market also threw away quality names. Picking up what was left behind could be the business of 2026.

Related: JPMorgan Refreshes AI List of ‘Stocks to Buy’ Ahead of Earnings Reports

This story was originally published by TheStreet on February 26, 2026, where it first appeared in the Investing section. Add TheStreet as a preferred source by clicking here.