Software stocks just recorded the worst relative performance against the S&P 500 in the sector’s entire recorded history. The iShares Expanded Tech-Software Sector ETF (IGV) cratered more than 24% in Q1 2026, its steepest quarterly plunge since Q4 2008. Salesforce, Adobe, Oracle, and even Microsoft are deep in correction territory right now.
The fear gripping investors centers on a single threat: artificial intelligence agents could hollow out traditional software platforms entirely. Short-selling volume across single stocks hit the highest level Goldman Sachs has recorded since 2016, a sign of genuine capitulation.
But one of Wall Street’s most powerful research desks is pushing back hard against the panic driving this historic selloff. Goldman Sachs has released a framework, a custom stock basket, and four specific buy-rated names that it says are being unfairly punished.
Goldman’s six-factor AI framework separates winners from casualties
Goldman Sachs Research analyst Matthew Martino published a report in February 2026 that entirely reframes the AI-versus-software debate. The selloff reflects a rapid shift in investor sentiment rather than a sudden deterioration in fundamentals, Goldman Sachs Research reports.
The team created a repeatable “AI Impact Framework” that evaluates software companies across six dimensions that determine AI resilience.
“We recognize that rapid AI innovation creates legitimate uncertainty and warrants a higher risk premium…Even so, we believe the repricing has been applied broadly rather than selectively.”— Matthew Martino, (Goldman Sachs Research analyst.)
Those dimensions include orchestration risk, monetization exposure, system-of-record ownership, data integration moat, AI execution, and budget alignment. The goal is to help you distinguish between stocks that face genuine displacement and those that got sold off indiscriminately.
Software valuations collapsed from 15% implied growth to just 5%
The scale of repricing becomes clear when you examine what the market now prices into software revenue growth. At their recent peak, software valuations implied a 15% to 20% medium-term revenue growth rate through 2028, Martino noted in his research.
Current multiples now correspond to an expected growth rate of only 5% to 10%, a dramatic downshift in investor confidence.
Relative to the SPDR S&P 500 ETF Trust (SPY), the software sector’s drawdown represents a 21% underperformance this calendar year. That figure exceeds what software experienced during the dot-com bust, the global financial crisis, and the 2022 rate-hike shock, individually making it the worst relative drawdown ever recorded for the software sector.
More AI Stocks:
- Morgan Stanley sets jaw-dropping Micron price target after event
- Bank of America updates Palantir stock forecast after private meeting
- Morgan Stanley drops eye-popping Broadcom price target
The software ETF’s forward price-to-earnings multiples have dropped below the S&P 500’s for the first time in recent memory. Goldman acknowledges that rapid AI innovation creates legitimate uncertainty and warrants a higher risk premium on software stocks right now.
The repricing has been applied broadly rather than selectively, creating potential opportunities where fundamentals remain intact, Martino wrote in his report. If you are a long-term investor, the distinction between broad fear and selective risk is the most critical takeaway from this analysis.
Oscar Wong/Getty Images
Four buy-rated software stocks Goldman says you should watch closely
Goldman applied its AI Impact Framework to four specific mid-cap software companies and concluded that each offers meaningful insulation from disruption. The bank maintained buy ratings on MongoDB (MDB), Rubrik (RBRK), Procore Technologies (PCOR), and Nutanix (NTNX) with revised price targets.
Each company scored favorably across multiple dimensions of Goldman’s six-factor assessment, suggesting their core businesses remain more durable than sentiment reflects. MongoDB’s consumption-based pricing model aligns with how AI workloads scale, supporting an upside case for Atlas revenue growth in the high 20s.
Goldman reiterated a buy rating with a $475 price target, citing the company’s database infrastructure as critical to enterprise AI deployment pipelines. Rubrik, down roughly 33% year to date, received an $80 price target as its data security platform benefits directly from AI-driven data proliferation.
Related: Goldman has very good news for beaten-down Microsoft investors
Procore Technologies, which has fallen about 30% this year, earned a $75 price target based on its system-of-record position in construction management. Goldman highlighted that Procore’s new CEO emphasized AI as a transformative force for the under-digitized construction sector during an investor meeting.
The firm expects steady revenue growth in the low-to-mid-teens and continued expansion of free cash flow margins toward 30%. Nutanix (NTNX) has suffered the steepest decline in the group, falling approximately 47% over the past six months as a mix of concerns weighed on the stock.
Goldman views Nutanix’s hybrid multi-cloud infrastructure positioning as strategic, with demand for hybrid environments likely rising alongside AI complexity. The bank reaffirmed a buy rating with a $60 12-month price target, using an enterprise value-to-free-cash-flow valuation approach, Benzinga reports.
How AI agents could reinforce platforms instead of replacing them
The dominant fear driving this selloff is that AI agents will become the primary interface for enterprise work, bypassing software platforms entirely. Goldman’s research team acknowledges this concern but argues that it applies unevenly across different layers of the software stack.
At the application layer, agent orchestration could shift engagement and value capture, especially for products monetized through seat-based licensing.
At the platform and infrastructure layers, however, the dynamics are fundamentally different because agents increase the demand for core data services. Data management, workload orchestration, security, and recovery capabilities sit beneath the user interface and cannot be easily bypassed by AI tools.
Key differences Goldman identified in AI vulnerability across software:
- Software monetized per user or seat faces a higher risk of displacement from AI agents that automate individual workflows at lower cost.
- Platforms tied to data assets, compliance, and execution serve as systems of record that AI agents still depend on to operate effectively.
- Companies with strong data integration moats and active AI product execution are positioned to benefit from broader enterprise AI adoption.
- Budget alignment with enterprise priorities around security and hybrid cloud provides additional insulation against AI-driven cost-cutting decisions.
Goldman’s AI-proof stock basket bets against the most vulnerable software names
Beyond individual stock picks, Goldman’s trading desk launched a custom pair-trade basket in February 2026, designed for this AI-disruption theme. The basket goes to long companies whose businesses require physical execution, regulatory entrenchment, or human accountability that AI cannot replicate easily.
It simultaneously shorts firms whose core workflows face the greatest risk of internal automation or AI-driven replication, Bloomberg reported. Goldman CEO David Solomon reinforced this view at a UBS conference, telling attendees that the sell-off narrative had been too broad.
There will be clear winners and losers among software companies rather than a wholesale collapse of the entire sector, Solomon emphasized publicly. For your portfolio, blanket selling of software positions may lock in losses on companies that Goldman expects to recover meaningfully.
The $780 billion software market projection changes the long-term calculus
Goldman Sachs Research estimates the application software market could grow to $780 billion by 2030, reflecting a 13% compound annual growth rate. The agent-driven portion of the software market may account for more than 60% of the total addressable market by that point,according to Goldman Sachs Research.
The profit pool is shifting toward agents, but the overall software market is expanding rather than shrinking under AI’s growing influence. For you as an investor, this means the total opportunity in software could be materially larger in five years, even amid ongoing disruption.
What you should do with your software positions right now
If you’re holding software stocks that have been punished this year, Goldman’s report provides a structured way to evaluate which positions deserve patience. The six-factor framework gives you a repeatable method to assess individual holdings beyond just looking at headline price action alone.
Steps to consider for your software stock positions going forward:
- Review each software holding against Goldman’s six dimensions to identify your portfolio’s specific exposure to AI-driven displacement risks.
- Determine whether your holdings generate revenue from seat-based licensing or from data and infrastructure, since that distinction drives vulnerability.
- Evaluate whether names like MongoDB or Nutanix fit your risk tolerance and investment timeline before adding any new exposure to the sector.
- Monitor upcoming earnings reports for signals about AI-related revenue contributions, since full AI agent replacement is a post-2028 event at the earliest.
No single analyst report guarantees a bottom has been reached, so sizing positions appropriately remains critical in this environment. Goldman’s framework replaces panic with structured analysis, but you should still verify each company’s fundamentals before making decisions.
The AI disruption story is real, but the market may be pricing in worst-case scenarios for companies built to benefit from the transition.
Related: Goldman Sachs spots a buying opportunity in bruised Big Tech
