Polar Capital's Software Exodus: A Strategic Pivot or a Market Misstep?

The scale of the software sell-off is historic. The iShares Expanded Tech-Software Sector ETF (IGV) has fallen for four straight weeks, declining roughly 22%-its worst four-week stretch since the pandemic. Relative to the broader market, the underperformance is staggering, marking the worst ever for software, surpassing even the dot-com bust and the 2022 rate-hike shock. This isn't a gradual correction but a rapid, sentiment-driven repricing that has compressed multiples indiscriminately.

Against this backdrop, the credibility of a strategic pivot becomes critical. Enter Nick Evans, a fund manager at Polar Capital, whose $12 billion global technology fund has delivered stellar returns, beating 99% of peers over one year and 97% over five. His thesis, therefore, carries weight. Evans argues that the rout is not a buying opportunity but a sign of deeper structural disruption. He believes application software faces an existential threat from AI, a view that has led to a decisive portfolio shift.

The fund's strategic pivot is now clear. It has sold all other holdings in the sector, exiting major application software names like SAP, ServiceNow, Adobe, and HubSpot. The portfolio is now majority invested in semiconductors and infrastructure, with seven of the top ten positions as of late January being semiconductor companies. This is a direct bet on the AI-driven hardware and foundational software stack, moving away from the very application layers Evans sees as vulnerable.

The Disruption Thesis and Cash Flow Reality

The core of Evans's thesis is a structural shift: AI tools are not just an incremental upgrade but a potential destroyer of the application software business model. He points to sophisticated platforms like Anthropic's Claude Cowork as evidence that AI can already replicate and modify much existing software. This creates a dual threat. First, it empowers clients to build their own internal tools, eroding the customer base of external vendors. Second, it opens the door for a new wave of AI-native startups to compete directly. The vulnerability is highest for horizontal platforms that provide basic automation-tools for writing, document management, and workflow. These are the easiest to replicate, making them the most exposed to customer churn and the resulting pressure on pricing and margins.

This disruption narrative is now being amplified by market mechanics. Hedge funds are acting as accelerants, increasing short bets against the sector. According to data, they have made a $24 billion windfall this year as the software industry's market value evaporated. The focus of these short sellers appears to be on the very companies Evans targets: those offering easily replicable services. This creates a feedback loop where fear-driven selling pressures valuations further, regardless of individual company fundamentals.

Yet, even if the long-term disruption is real, a more immediate operational risk looms. The market rout triggered by this threat could squeeze cash flow. Software companies often compensate employees with equity, and when stock prices fall sharply, the value of that compensation erodes. Managers may then need to pay out more cash to make up the difference, directly pressuring the balance sheet. Evans notes that any effort to buy AI startups to stay competitive could add to this financial strain. In his view, the current market prices do not reflect this heightened uncertainty about terminal value or the tangible pressure on free cash flow. The result is a sector caught between a structural overhang and a cash flow squeeze, where the path to survival may be far more expensive than anticipated.

Counter-Arguments and the Microsoft Dilemma

While the narrative of AI-driven destruction gains traction, a powerful counter-argument is emerging from the other side of Wall Street. JPMorgan Chase strategists recently argued that the recent price action in software stocks has been extreme, creating a potential buying window. They see the sector as oversold, with names like Microsoft and ServiceNow positioned for a rebound as AI adoption scales. This view directly challenges the notion that the entire application software category is toxic. For JPMorgan, the current panic is a sentiment-driven overreaction, not a fundamental re-rating of business models.

Polar Capital's position, however, reveals a nuanced and selective acceptance of this counter-argument. The fund manager does not buy the broad recovery thesis. Instead, he retains a small, defensive position in Microsoft, coupled with call options. This is not a bet on Microsoft's application software suite, but a recognition of its unique, unassailable moat. Microsoft's integration with AI infrastructure-through its Azure cloud platform and deep embedding of Copilot-creates a different kind of defensive advantage. Evans acknowledges that companies like Microsoft, which provide foundational tools for the AI stack, may be less vulnerable than pure-play application vendors. His call options, in particular, are a bet on a potential upside catalyst, not a full conviction in the sector's future.

This divergence in strategy highlights a stark market split. The software sector has plunged while semiconductor stocks have soared as the physical engines of AI demand. The money is flowing from the application layer to the hardware and infrastructure layer. This isn't just a sector rotation; it's a reallocation of capital based on perceived durability. The debate now centers on whether the software sell-off has gone too far, or whether it is a necessary, albeit violent, correction for a business model under siege. The market is effectively voting on that question with its capital.

Catalysts, Scenarios, and What to Watch

The immediate test for the sector is here. Over the coming days, a wave of earnings reports from major software firms will provide the first concrete data on how resilient their fundamentals are amid the storm. This is the catalyst that will either validate the disruption thesis or offer a reprieve for the bulls. The market is watching for signs of pricing power, margin stability, and customer retention-key metrics that could signal whether the existential threat is already materializing in the numbers.

The primary risk remains the acceleration of AI capabilities. If models like Anthropic's Claude prove to be more powerful and versatile than current estimates, the timeline for disruption shortens dramatically. This would validate the "toxic" thesis that Nick Evans and other skeptics have advanced, potentially prolonging the selloff even if some companies report decent quarterly results. The fear is that strong earnings could be overshadowed by a narrative shift, where investors price in a future of eroded moats and compressed margins. The hedge fund data shows the market is already primed for this, with short sellers having made a $24 billion windfall this year and the sector's ETF down 30% from its peak.

A secondary, and perhaps more persistent, risk is a broader market rotation that decouples software performance from its underlying business health. The capital flight has been decisive, moving from application software to semiconductors and infrastructure. Even if a company's fundamentals hold, its stock could remain under pressure if the sector-wide sentiment persists. This creates a scenario where the market's structural shift-betting on the AI stack rather than the applications built on it-could outlast the initial panic. The result would be a prolonged period of underperformance, regardless of individual company execution.

The bottom line is that the coming earnings season is a high-stakes calibration point. It will test the durability of the software business model against the accelerating force of AI. For now, the market's verdict is clear: it is pricing in significant uncertainty. The path forward will depend on whether the catalysts of earnings and AI progress confirm a deep structural change or merely a turbulent transition.