Early in the “SaaSpocalypse” that roiled software companies this year, Manish Gupta, a portfolio manager at First Eagle Investments, added to Workday as he parsed through the wreckage for some bargains. In recent months, that bet has paid off for the value investor. After falling about 40% in the first quarter of 2026, during which the enterprise software stock at various points was trading at not even three times revenue, the stock went on to surge more than 55% in the third. It’s also outperforming this week, rising more than 2% amid a broader tech sell-off as investors exit for safer assets. Gupta, who attended the company’s analyst day this week, remains confident in Workday, which he said is trading at a discount to similar businesses in the private markets. Workday is a top 10 holding in the First Eagle Global Fund (SGENX) , which he co-manages and which had nearly $75 billion in total net assets as of Sept. 30. “The price has moved up, but the business has also become better,” Gupta said. “The ‘SaaSpocalypse’ has also turned around, but the stock’s still attractive.” Workday’s rally comes amid a broader rethink of software companies, which are starting to look more attractive as money rotates out of semiconductors and into parts of the AI trade that investors think will outperform going forward. Investors are growing wary that 2027 will be a challenging year for the stock market, due to fears of an impending deceleration in the AI capex cycle, an anticipated flood of new equity supply from major tech IPOs and a ballooning U.S. fiscal deficit . These factors, along with higher inflation and rising interest rates, could pressure the market and the economy. In that environment, investors are betting that mega-caps and infrastructure software companies are among the groups best positioned to take on the coming challenges. “[2027] could be a year where AI applications win over the AI build-out,” Ohsung Kwon, chief equity strategist at Wells Fargo, told CNBC’s ” Squawk on the Street ” earlier this month. “Meaning hyperscalers, infrastructure software, over semis.” The market is already moving to price that in. Over the past three months, the Roundhill Magnificent Seven ETF (MAGS) has gained more than 6%, while the iShares Expanded Tech-Software Sector ETF (IGV) has climbed more than 13%. The S & P 500 has advanced 2.7% during that same stretch. Meanwhile, the iShares Semiconductor ETF (SOXX) has fallen by more than 4% in the past three months. On Thursday, the ETF was down for an eighth straight session — notching its worst losing streak since a nine-day slide all the way back in May 2006. The case for AI applications In the first quarter, the software stocks ETF, IGV, tumbled 24%, as investors worried that AI agents could replace the software packages that are sold by many companies. But that narrative around obsolescence has largely receded in the intervening months, as strong earnings expectations and new AI products and partnerships out of the sector revived investor confidence. Big changes are still coming for software players, as they shift to AI chatbots from screens users previously clicked through. “AI is changing not only what software can do, but how humans interact with it,” Tal Liani, research analyst at Bank of America Securities, wrote. “If AI 1.0 was the model, AI 2.0 was the agent, then AI 3.0 is the interface, driving a meaningful shift in the relationship between users, applications and the software stack.” The analyst said that will help those companies that already own the platforms and daily productivity hubs that users work on. Microsoft , for example, reclaimed its $4 trillion market cap this week. Liani noted that the AI winner “already owns many of the surfaces where knowledge work occurs,” whether that be Word or Teams. ServiceNow is another winner in a different way. The software platform can manage workflows between human resources and IT, even if the tasks involve an outside chatbot such as Claude. Getting defensive Investors remain confident that the stock market could continue to climb through the rest of this year and next, given the strength of the underlying fundamentals that the current third-quarter earnings season is also expected to confirm. But the mounting challenges have investors getting more defensive as they worry about the end of the AI capex cycle that could bring some of the highest valuations in the market down a bit closer to earth. Wells Fargo’s Kwon said he expects that the AI spend will essentially peak in 2027, and then flatline in 2028. That deceleration could benefit the hyperscalers as they become more disciplined, but potentially make data center projects less attractive. The strategist said he turned cautious in September, anticipating a time when the market will have to reset expectations that could spark a 5% to 10% drawdown even by the end of the year. “There’s going to be an end to it, and people are starting to think about what that end looks like,” Kwon said.
