Which software giants in the US stock market will be disrupted by AI? A chart to see the distance to the "cliff"

Author: Dong Jing

Source: Hard AI

JP Morgan's latest research report reveals that AI is reshaping the competitive landscape of the software industry, with many traditional software giants facing the risk of being disrupted by AI. This Wall Street investment bank has pioneered the "AI Cliff" assessment framework, which quantifies the vulnerability of software companies through nine dimensions, providing investors with a clear risk map of the industry.

At the beginning of 2025, in a market atmosphere characterized by tariff concerns and corporate budget tightening, the software sector had remained relatively calm. This industry, traditionally viewed as a defensive sub-sector in technology, is favored for its minimal direct impact from tariffs and its reliance on stable, sticky products to generate recurring contract revenue.

However, on October 21, according to Hard AI, JP Morgan's North American credit research team pointed out in their latest report that this calm is being disrupted. Investors and corporate clients are beginning to ask the same unsettling question: If AI is good at writing the code that constitutes software, why can't it directly write code to improve and compete with existing software products?

The report states that this concern is not unfounded. Software is essentially made up of code, and AI has demonstrated strong capabilities in the field of code writing. The bank pioneered the "AI Cliff" assessment framework, which quantifies the vulnerability of software companies through nine dimensions such as replacement cost, criticality, and level of automation, finding that companies with strong ecosystems and high user visibility are more defensive, while traditional software systems with weaker adaptability face greater risks.

JP Morgan emphasizes that although analysts believe AI could ultimately impact almost all software companies, the timing of technological change is often difficult to predict, and there are significant differences in how close different software companies are to the disruptive "cliff."

Nine-Dimensional Framework: Deconstructing Key Variables of AI Disruption

JP Morgan's assessment framework evaluates the AI disruption risk of software companies through nine key factors, employing a qualitative analysis method that rates each factor as "high" (strong defensiveness), "medium," or "low" (vulnerable to impact).

In terms of replacement cost, it considers the time, financial investment, and level of disruption to customers; for example, Microsoft Windows has a high replacement cost due to learning curves and inertia, while Alteryx is relatively easy to replace.

The criticality dimension distinguishes between mission-critical software (like CDK) and auxiliary tools (like Alteryx).

Level of automation is another important indicator. Highly automated billing systems are less affected by AI, while software that relies on manual processes (like Microsoft Excel) is more vulnerable.

User visibility is another aspect, where software that users interact with daily (like Microsoft Windows) is stickier than backend middleware (like TIBCO).

Ecosystem scale determines the difficulty of replacement. Bloomberg has a large user ecosystem and vendor support network, while niche market software like PTC has a relatively limited ecosystem.

Data resource dimension focuses on the value of proprietary datasets, with Experian possessing rich proprietary data, while CoreLogic's data is more non-proprietary Scale and resources affect a company's resilience. Large companies like Google have the ability to weather adjustment periods, while smaller firms like ZipRecruiter, although more agile, have limited resources.

In terms of adaptability, modern API-based software (like Elastic) is easier to integrate AI features compared to traditional legacy systems (like Unisys).

Regulatory requirements provide additional protection for existing software in sectors like finance and healthcare.

Heatmap Insight: Who is Closest to the Edge of the Cliff

JP Morgan applied this framework to the software companies it covers and created a resilience heatmap.

The assessment results show significant differences in performance across various dimensions, with "high" (green) indicating greater defensiveness and "low" (red) indicating greater susceptibility to disruption. Specifically:

For example, CrowdStrike excels in criticality and adaptability but is rated low in automation levels.

GoDaddy received a moderate rating in data resources but scored low in scale and resources dimensions.

Open Text performs poorly in adaptability but has advantages in criticality.

Twilio and Elastic are rated low in user visibility and regulatory requirements, indicating relatively weak defensive barriers in these areas.

JP Morgan stated:

“We want to emphasize that our analysis is subjective... but we found this framework helpful for our own assessment of potential AI disruption and believe analysts can combine it with their personal knowledge and experience to better evaluate a company's software relative vulnerability to AI.”

The bank also noted that in addition to the nine major factors, there are other issues that may affect the impact of AI on existing software companies. For instance, the legality of using existing software code to train AI models to develop similar software remains unanswered. The end market for software will also affect adoption rates, as decision cycles in consumer markets are often shorter than in enterprise markets.

Finally, JP Morgan pointed out that technological change is an inevitable trend, but the time spans vary greatly. Some changes are completed within a few years (like paper maps and film), while others take decades (like copper telephone lines and electric vehicles).

For AI-related software disruption, the timing of change may not be imminent. However, the distance of specific software from the "cliff" of disruption will vary significantly due to multiple factors. This framework helps assess the vulnerability of software facing AI challenges