Beaten-Down Software Stocks Are Still Good Buys, Despite Investors’ AI Fears

Software Stocks & the AI Disconnect: An Analysis

Key Concepts:

• SaaS (Software as a Service): A software distribution model where applications are hosted by a provider and made available to customers over the internet.
• DCF (Discounted Cash Flow): A valuation method used to estimate the value of an investment based on its expected future cash flows.
• ARR (Annual Recurring Revenue): A key metric for SaaS companies, representing the value of recurring revenue normalized to a one-year period.
• Free Cash Flow Margin: A financial ratio that measures the percentage of revenue a company generates as free cash flow.
• Sentiment-Driven Market: A market where investor behavior is heavily influenced by emotions and perceptions rather than fundamental analysis.
• Token/Conversation Basis (AI Pricing): A new pricing model for AI services based on usage, such as the number of tokens processed or conversations generated.

1. The Performance Divergence & Initial Fears (July 2025 - Present)

Software stocks experienced a significant underperformance relative to artificial intelligence (AI) stocks starting in July 2025. Prior to this period, software performance was aligned with the broader Morningstar US Equity Index. However, as AI gained prominence, software stocks “fell off a cliff,” experiencing a “bloodbath” in performance. This decline stemmed from several fears: initially, the belief that AI would eliminate the need for software altogether. This evolved into concerns that AI would reduce the demand for software licenses within the SaaS model. A further concern was the lack of immediate revenue generation from AI initiatives. The market was caught in a contradictory position, fearing both a lack of AI revenue and the potential for AI to overtake the software industry. Dan Romangh, Senior Equity Research Analyst at Morningstar, emphasizes this was largely “sentiment driven.”

2. Fundamental Performance Remains Strong

Despite the negative sentiment and stock performance, the fundamental performance of software companies remained remarkably robust. For the third quarter (reported in October), Romangh notes that all 20 companies under his coverage beat both revenue and profitability (EPS) expectations. Guidance provided by these companies was generally in line or slightly better than anticipated. This positive trend continued throughout 2025, with consistent beats and raises in performance. Romangh anticipates a similar pattern in 2026. This creates a stark contrast: strong fundamentals coupled with aggressive selling pressure. He highlights that even with software’s underperformance, technology as a whole remained one of the best-performing sectors, demonstrating the overall strength of the tech landscape.

3. AI Revenue: Small but Rapidly Growing

The actual revenue contribution from AI within the software industry remains relatively small. Romangh estimates that, excluding Microsoft, the total revenue of his covered companies was approximately $100 billion. A few quarters ago, AI-related revenue accounted for roughly $1 billion (1% of total revenue). Currently, this figure has risen to around 1.5%. In contrast, companies like Microsoft, Amazon, and Google are already generating billions in AI revenue, primarily through cloud computing services (renting out NVIDIA GPUs). Salesforce recently reported $500 million in ARR from AI, growing at over 300%, but this still represents a small percentage of their overall revenue. Romangh predicts AI revenue could reach 2% within a quarter and potentially 5% in a couple of years.

4. Margin Implications & the Shift to Consumption-Based Pricing

A key concern is the potential impact of AI integration on software margins. However, Romangh notes that management teams generally believe AI margins are more attractive than their existing business. This is because companies are charging a premium for AI-enhanced services, often as an add-on to existing SaaS licenses. A significant shift is occurring towards consumption-based pricing (charging per token or conversation), which introduces uncertainty but is likely to be margin-accretive. Romangh argues that even if a customer stops paying for a traditional software seat, the revenue from an equivalent AI instance, priced appropriately, could remain neutral. He observes that, overall, margins have been expanding even as AI products are developed and deployed.

5. Valuation & Top Picks (Undervaluation & Opportunity)

Romangh believes software stocks are currently undervalued, offering approximately 25% upside based on Morningstar’s fair value estimates. He contrasts the current situation with mid-July 2025, when software stocks were considered fairly valued. To justify current market prices, DCF models would require significant reductions in revenue forecasts and margin assumptions – a scenario he deems unrealistic. His top picks are Microsoft and ServiceNow.

• Microsoft: Romangh highlights Microsoft’s strong free cash flow margin (35%), monopolies in key markets (Windows, Office, Gaming), and leadership position in public cloud computing and AI. He notes that AI revenue is largely driven by Microsoft’s cloud offerings.
• ServiceNow: ServiceNow is identified as a top growth story in the large-cap enterprise software space, with a 20% growth rate exceeding the industry median. They were early to disclose AI revenue and boast superior profitability and margins compared to peers like Salesforce. Recent acquisitions, while viewed with some skepticism, are considered immaterial to the overall growth trajectory.

6. DCF Modeling & Margin Analysis

Romangh’s valuation relies on DCF modeling, involving 10-year forecasts of revenue, expenses, and profitability. He points out that for some companies, current market prices imply a need to halve revenue projections and reduce margins by 1000 basis points (10%) to justify the valuation. He finds this scenario implausible, particularly for companies like Microsoft with historically stable margins.

7. Notable Quote:

“If I’m Warren Buffett, I say look at Microsoft’s free cash flow margin and Service Now’s free cash flow margin of 35%. And say that is a winning formula, you know, for basically low multiples and, you know, the divergence of our fair value from where the stock's trading now. It's just intrinsically undervalued.” – Dan Romangh

8. Logical Connections & Synthesis

The analysis demonstrates a clear disconnect between market sentiment and fundamental performance. The initial fears surrounding AI’s impact on software have driven down stock prices despite consistently strong earnings and guidance. While AI revenue is currently small, it is growing rapidly and is expected to contribute more significantly in the coming years. The shift to consumption-based pricing introduces uncertainty but is likely to be margin-accretive. Romangh concludes that software stocks are currently undervalued, presenting a compelling investment opportunity for those who can look beyond the prevailing negative sentiment. The key takeaway is that the market is overreacting to perceived threats from AI, creating a buying opportunity in fundamentally sound software companies.