Goldman Sachs' technical research team has identified that the core challenges currently facing software stocks are concentrated in the new competitive landscape at the application software layer and the return on capital expenditure for infrastructure companies.
This week, a routine update from Anthropic significantly impacted software stocks. Although a rebound occurred on Friday, the sector remains in a state of low-level volatility.
The team, led by Gabriela Borges at Goldman Sachs, anticipates that an improvement in investor sentiment will require 2-3 quarters of stable fundamental performance. Consequently, the software sector is expected to continue facing valuation pressure in the near term. Goldman Sachs emphasizes that the onset of a new technology cycle, the large number of competitors nurtured by the 2020-2021 funding boom, and the necessity to redesign system architectures are all factors intensifying competition within the application software layer. Borges has outlined an observation checklist for investors, detailing seven key indicators that could signal a move towards stabilization in the AI-SaaS industry. This framework is designed to help the market assess whether application software companies can withstand the impact of new competitors and when the massive capital expenditures by infrastructure providers might translate into substantial returns.
Two Core Debates in the AI-SaaS Realm Goldman Sachs' technical research team highlights two primary debates within the AI-SaaS domain. The first debate centers on whether established software companies will be replaced. Goldman Sachs believes new technology cycles invariably bring disruption. Currently, traditional SaaS giants, bespoke software providers like Palantir, and startups with genuinely unique datasets and products—not merely an AI veneer—are all competing for the same customers on the same battlefield. This intense competition stems from three reasons. Firstly, whenever new technology emerges, enterprises often want to experiment in-house, creating a window of opportunity for new entrants. Secondly, companies that flourished during the funding surge around 2021 may have moats that are less defensible than previously thought. Most critically, whereas past software was primarily designed for human use, future software must simultaneously serve both humans and AI agents. This implies that the entire software technology stack requires reconstruction. The burden of overhauling legacy systems for established players represents a significant opportunity for new entrants. Therefore, selectivity becomes paramount. Investors should critically favor application software companies that are progressing more rapidly in redesigning their technology to adapt to a "human-AI collaboration" paradigm.
The second focal point is the market's concern about where the money will ultimately flow. Goldman Sachs notes a consensus in client discussions that every software vendor will offer AI agents in the future. The question then arises: where will differentiation and excess profits come from? The core lies in "orchestration" capability—the ability to efficiently and reliably integrate underlying compute power, models, enterprise data, business processes, and security rules to deliver a functional agent. This task spans the infrastructure, platform, and application layers, requiring both high technical complexity and deep industry knowledge. This makes vertically integrated capabilities more crucial than in the cloud era. For giants like Microsoft, the ability to dominate across all layers becomes a greater advantage. At the company level, Goldman Sachs analysts believe that for application software vendors, value will aggregate at the "orchestration layer." The transition from SaaS to "SaaS+AI" is a brutal race, with participants starting from different positions. For infrastructure companies providing compute power, the key to success lies in diversifying chip sourcing to improve margins and adding valuable platform services on top of basic compute capabilities.
Seven Key Signals for Recovery The aforementioned changes will unfold gradually. Goldman Sachs analysts have listed seven more specific, quarterly-trackable signals. Their sequential emergence would indicate that industry fundamentals are beginning to stabilize. The first is a quiet shift in revenue structure. If overall corporate budgets for traditional software do not grow this year, but the total revenue of leading software companies stops declining and begins to recover, it would signal that AI-related orders are substantively compensating for, or even surpassing, weakness in traditional businesses. Secondly, the return from "in-house" projects. When companies like ServiceNow begin reporting that clients are abandoning their own AI projects in favor of established products, it is a positive signal. This indicates that packaged software is building sufficient advantages in functionality, security, and compliance controls, with enterprise market rules taking effect. The third signal is the confidence to raise prices. Currently, most AI features are in a promotional phase. Future attempts by vendors to charge separately for agent functionalities or implement price increases—that are accepted by customers—would prove that AI is creating recognized commercial value. This path resembles the strategy employed by Palantir a year or two ago: "get the client using it first, then discuss pricing." Next, monitor whether industry knowledge becomes more valuable than the model itself. The market needs clearer examples to understand why, in specific scenarios, agents from companies with deep industry expertise like Salesforce or HubSpot might outperform generic Copilots. This will determine the pricing power and market scope of different companies' agent products. Fifth, watch for unexpected mergers and acquisitions. It should not be surprising if an AI platform or large language model company suddenly acquires a vertical SaaS firm. This is the most direct method for AI players to rapidly gain industry knowledge, sales channels, and "enterprise-grade" credibility. Sixth, observe how companies use capital to address talent challenges. AI has intensified competition for expensive technical and sales talent. Investors should monitor how companies manage the associated risks of equity dilution and employee attrition. Whether they engage in more aggressive stock buybacks or decisively adjust compensation structures to attract new-generation talent reflects management's resolve. Finally, clarity on the cloud giants' strategies is needed. For giants like Microsoft and Oracle that provide AI compute, the market requires clearer insight into when their new chip production capacity will come online and how much of this capacity will be allocated to their own businesses versus sold to external customers. This directly impacts the growth and profit margins of their cloud businesses. Goldman Sachs ultimately concludes that it expects 2-3 quarters of stable fundamentals will be necessary for investor sentiment to improve. Even then, the possibility remains that bearish scenarios could be delayed to future years.
Comments