From the CIO's Desk: Softness in Software

Software stocks have fallen on hard times to start the month of February. During the first week of the month, software stocks in the S&P 500 were down -8% versus a mere 0.1% dip in the index as a whole¹. The hit to software stemmed from news that Anthropic released plugins for its Cowork product that automate legal documentation and research for lawyers. The news sent shockwaves through the sector, notably lowering the stock prices of software companies. As the week progressed, a level of risk aversion set in amongst investors, with growing concerns around holding the debt or equity of a software company.

Artificial Intelligence (AI) continues to evolve, and Anthropic's announcement demonstrates notable progress in learning and inference. To us, this development was not overly surprising. Functions such as compiling research and drafting documents have been among those AI has been rumored to be capable of performing. To us, the reaction was investors recognizing that capabilities such as these are here today or on the near-term horizon. Therefore, Anthropic’s news led to a reassessment of risk and conviction among software investors.

It's important to call out that, as it stands today, software companies continue to deliver attractive financial results. Revenue, earnings growth, and margins are collectively strong across the industry. Coming into the year, software companies had annualized margins of 26%, a level nearly twice that of the S&P in aggregate². Looking forward, expectations are for software companies to grow earnings by 15% over the next two years.

If business conditions are strong, why is the market reacting so significantly? A key factor, we believe, is considerations around valuation. A common approach to valuing a stock is to consider its long-term expected earnings or cash flows and assign value to those. In a world where you are less certain about competitive dynamics or even the viability of an application over the next five or ten years, investors are going to be less confident in prescribing a higher valuation. Much of last week’s selloff was tied to investors reducing the price they were willing to pay for future growth amid the growing cloud of ambiguity stemming from AI applications.

The pace of the selloff was only exaggerated by investor positioning. Software companies have delivered outsized returns in recent years and have become favorites among investors. As conditions changed, these overowned stocks were swiftly sold, with hedge funds in particular reducing total exposure³.

Is now the time to pile into technology, particularly software? We don’t think so. Sentiment has changed. A broadening of the market beyond technology-centric businesses is now in full swing. We do, however, believe there will be opportunities to add value through stock selection. Within software, we see the greatest potential in companies that are deeply embedded in daily workflows, generate proprietary data, or provide security or compliance functions, as these are less likely to be displaced. The timing of a rebound, however, seems less urgent. The cloud of concerns about how AI could influence the long-term competitive dynamics in software, and how much one should pay for a stock given the uncertainty, is unlikely to lift in the coming weeks or months.

The growing capabilities, as evidenced by Anthropic’s Cowork, suggest that AI-driven efficiencies may be on the near-term horizon for companies beyond software. Business functions such as accounting, human resources, compliance, and customer support are likely to benefit from AI efficiencies. This suggests that higher profit margins could be on the horizon for businesses outside the technology sector that can effectively leverage AI.

One of the benefits we have at Composition Wealth is the depth and breadth of experience on our team. When we encounter episodes of disruption, we bring together these varied disciplines to broaden our perspective. Below, our equity and fixed income leaders summarize their perspectives on software companies.

The initial AI beneficiaries have been those companies that provide the hardware and infrastructure required to build out data center capacity. Nvidia, Advanced Micro Devices, and Broadcom provide the semiconductor chips that power much of the global compute expansion. Other companies that provide memory, storage, physical infrastructure, and power are also early winners as they have benefited from the extraordinary demand required to build data center capacity. Meanwhile, software and IT service companies are facing enormous pressure as investors question the durability of these business models that may be suddenly disrupted by AI.

While we acknowledge this risk, we believe that future economic value will accrue to software companies that successfully integrate AI into workflows that help enterprise customers improve productivity. Not all software is facing obsolescence risk, and in fact, incumbent software companies are in a unique position – they often have longstanding relationships with their customers, understand bottlenecks and opportunities, and have domain over data and workflows. However, to survive, incumbent software companies must embrace AI and fully incorporate it into their product and service offerings. The winners will be software companies that can help enterprise customers reduce expenses by lowering software coding costs, minimizing customer support staff, and shrinking research and development budgets. They must also accelerate revenue growth by leveraging AI agents to generate leads and support sales growth. – Rusty Hoss, CFA - Managing Director, Portfolio Manager (Equities)

Evolving trends in the technology sector are prompting fixed income investors to take a closer look at the credit quality of software companies. Large technology companies are significantly increasing spending on artificial intelligence infrastructure, which is expected to lead to higher debt levels. Several large technology companies have announced AI-related spending plans well above expectations, and we believe a significant portion of this investment is likely to be financed through the bond market. This increase in new debt can pressure bond prices and lead investors to demand higher yields.

At the same time, software companies are being reevaluated as investors consider whether AI will increase competition and reduce pricing power. A similar adjustment played out in healthcare when GLP-1 weight-loss drugs emerged: companies tied to diabetes-related medical device businesses faced significant uncertainty, and it took time for investors to regain confidence in which business models would remain durable. Software may face a similar period of uncertainty until the long-term impact of AI becomes clearer. For bond investors, this environment reinforces the importance of selectivity and credit discipline within the technology sector. – Brian Bunker, CFA – Director, Portfolio Manager (Fixed Income)

Conclusion

Episodes of volatility are expected in 2026 as evolving technology, expanding thematic trends, and global policy changes influence sectors of the economy. The spotlight has now shifted onto the software sector and the impact of AI disruption. While compelling investment opportunities are likely to present themselves, we continue to believe that broad and thoughtful diversification across portfolios will help support consistent outcomes for investors.

1. Source: Bloomberg. Software stocks proxied by the S&P 500 Software Industry
2. Goldman Sachs Investment Research, Factset. As of February 6, 2026.
3. Source: Goldman Sachs Investment Research, Goldman Sachs Prime Services. As of February 3, 2025