"At the time when the 'software stock doomsday' discourse is sweeping Wall Street, the growth narrative of 'AI reshaping software profits' is quietly spreading."

The pessimistic narrative of "Software-mageddon" regarding the global stock market is gaining momentum. Currently, global investors, including large Wall Street institutions, are fiercely debating whether it is time to call for a bullish outlook on software stocks that have experienced significant losses recently. Based on the recent capital outflows from global large software stocks, some institutional funds are beginning to enter the market to take advantage of the recent plunges in software stocks, and they agree with Huijiong Huang's positive view on software stocksthat the market has mistakenly killed off strong software giants focused on "AI + core operations processes." However, some investors remain cautious. At the same time, this also indicates that a new narrative of using AI to empower profits is quietly spreading, with some investors already incorporating future growth expectations into the evaluation of software stocks. AI Agent Tool Sparks Concerns Earlier in January this year, Anthropic, known as a rival to OpenAI, launched the Claude Cowork tool, a collaborative AI programming tool with significant engineering innovation. This tool aims to extend the functionality of AI agents from programming terminals to general office scenarios, such as document management and software interactions. This move intensified market fears of AI intelligence completely disrupting the SaaS software industry. Starting this Tuesday, the primary culprit behind the collective collapse of global software stocks has been the new AI tool introduced by Anthropica highly efficient AI agent capable of performing various document tasks, including tracking compliance issues and reviewing legal documents. It is important to note that reviewing legal documents, financial analysis capabilities, and exclusive data service technologies are strong long-term competitive advantages for most SaaS software companies. Concerns over the impact of Anthropic's AI tool have triggered the latest round of volatility, which has been compounded by disappointing earnings guidance from companies like Microsoft Corporation (MSFT.US) and expectations of larger-scale AI infrastructure expenditures. It was reported that the American Software, Inc. Class A sector experienced its most severe sell-off since 2022, but the "buying the dip" funds that typically enter quickly during tech stock crashes did not fully participate in the continuous collapse of the software sector. After the S&P 500 Software and Services Index plummeted nearly 4% on Tuesday, it fell another 1% on Wednesday, marking six consecutive days of decline. Since its recent high in late October, the S&P 500 Software and Services Index has dropped by about 25%while the overall S&P 500 has remained relatively unchanged. Some traders are not particularly interested in "buying the dip" with these struggling software stocks at the current levels, and some options traders are choosing to wait and see. Art Hogan, Chief Market Strategist at B Riley Wealth, said, "This is 'Software-mageddon.'" On Thursday, the impact of this AI tool on global software stocks continues to weigh on them to some extent. This is particularly evident in the continued plunge of Asian software stocks and technology stocks more broadly on Thursday. The Korean stock market, known as a gauge for Asian technology stocks, experienced a significant drop on Thursday, with foreign investors selling a record 4.99 trillion Korean won (US$3.4 billion) worth of stocks in the Kospi Index. Large institutional investors also sold a net amount of up to 2.07 trillion Korean won worth of Kospi Index components. The S&P 500 Software and Services Index experienced a sharp decline of 13% over the past week, resulting in a market capitalization loss of over $800 billion. This decline has been driven by the significant drop in stock prices of long-standing software giants in the industry, such as Intuit (INTU.US), ServiceNow (NOW.US), and Oracle Corporation (ORCL.US). Stock strategists at Evercore ISI noted that, as of Tuesday, the U.S. software sector recorded its worst three-month performance since May 2002, following the aftermath of the bursting of the dot-com bubble. The chart above shows the performance of software stocks relative to the broader U.S. market benchmark, the S&P 500 Index, since late October. It has significantly underperformed the S&P 500 Index. Institutional Responses and Disagreements In the face of the market turmoil, institutional investors are showing clear differences in their reactions. Some are making small-scale investments, while others remain cautious. The steep decline in prices has triggered some technical signals that suggest the sector may have reached a temporary bottom. Some portfolio managers have started to make small purchases of these software stocks that have been affected by panic selling. However, investors are still hesitant to declare that the "all-clear signal" has been given. Jake Seltz, a portfolio manager at Allspring Global Investments in Minneapolis, stated, "There is indeed some long-term value in these names, and they are reaching price levels that I find more attractive for buying on the dips." He has been gradually increasing his holdings of some software stocks over the past few months, including ServiceNow and Monday.com. Seltz emphasized that he is waiting for catalysts that could lead to more aggressive buying on the dips, such as major software companies reporting strong revenue data related to AI products or more corporate clients announcing large-scale deployments of these companies' AI tools or AI agent-based updates for "AI+SaaS software product series." Walter Todd, Chief Investment Officer at Greenwood Capital, revealed that his investment firm has recently made small purchases of stocks in ServiceNow and Microsoft Corporation. He pointed out that while he is not looking to "bet the farm" on a rebound in software stocks, he does believe they are starting to show buying opportunities. "I'm not convinced that a complete replacement of existing software infrastructure with AI solutions is entirely realistic in these circumstances," he added. Brad Conger, Chief Investment Officer at Hirtle, Callaghan & Co., stated that he has started to weigh potential buying opportunities, including software giants like SAP (SAP.US), Adobe (ADBE.US), and Intuit, which have been hit hard in the software stock sell-off. He said, "You can argue that they are due for a rebound." However, he added that he is not ready to become a final buyer at current price levels because he is not "sure they have reached a price level that fully accounts for the worst-case scenario threat." Rene Reyna, Head of Thematic and Specialized Product Strategies at global asset management giant Invesco, said, "We are beginning to understand the capabilities of AI more clearly, and the market is re-pricing to reflect reduced confidence in the future sales growth of software driven by AI." However, he also added, "Whether they have been oversold? We cannot tell at the moment. There may be a rebound, but selling pressure could also trigger more selling." AI Reshaping Value Chains and Future Observations From the perspective of software engineering reality and the SaaS industry structure, the narrative of "AI replacing the entire enterprise software stack" is an easy one for the market to linearly extrapolate. The "value density" of enterprise software lies not only in interfaces and functionalities but also in proprietary data, permissions/audit chains, compliance and responsibility boundaries, system integration, SLAs and availability, change management, and organizational processes; these factors determine that even with strong large language models (LLM), a high-quality proprietary corpus + structured knowledge base + controllable tool invocation + traceable outputs are often required for production environments. An example cited in a study by Thomson Reuters Corporation Breakingviews argues that the market may have overpriced data-centric companies like RELX/Reuters, stating that most of RELX's revenue expectations face low LLM substitution risks, and the business segments that are truly "at risk" form a limited proportion. These companies are also introducing AI tools trained/enhanced based on proprietary databases. Likewise, the viewpoint cited by Thomson Reuters Corporation Breakingviews emphasizes that equating an AI plugin with replacing the entire critical enterprise software layer is a leap in illogical inference. From the underlying technological logic of AI tools and the SaaS software field, the panic selling of software stocks does not equate to "software becoming obsolete," but rather a redistribution of value chains by AI: stronger general large models and agent-based AI workflows are posing risks to a multitude of "point-functional SaaS" through internalization of functions at the model/platform level or bypassing UI and seats with "conversational entrances + automated executions," which could challenge traditional seat-based pricing and renewal logic; therefore, the market is keen to categorize software stocks as "AI winners/AI losers." Conversely, leaders in enterprise SaaS software often have data sovereignty, governance, permissions, auditability, and migration cost barriers. In reality, AI is more likely to turn these longstanding software giants into distribution channels for delivering AI capabilities, rather than overnight replacements for entire existing software infrastructure. According to the Morgan Stanley strategy team, at the short-term level, the sharp decline in the software sector has indeed triggered discussions about "technical signs approaching a temporary bottom," with some funds making small-scale additions. However, more funds are still awaiting hard catalysts that can translate the AI application narrative into actual revenue curvessuch as software companies disclosing revenue/penetration rates related to AI products, corporate clients announcing significant deployments, or renewal indicators (net retention, expansion rates) showing significant strengthening after the introduction of large models or agent-based AI entities; without this evidence, the rebound is more likely to be a "cheap repair" rather than a new trend. The recent software stock sell-off wave seems more like the market's extreme response to a new question: to what extent will the profit pool of SaaS software vendors be redistributed by "model factories + agents"? In the short term, the answer can only be verified through two "hard indicators": (1) the speed of actual deployment and payment diffusion on the enterprise side; (2) the elasticity of SaaS vendors' revenue from AI-related products and renewals/net retentionjust as some buyers described in a study by Thomson Reuters Corporation Breakingviews, they are waiting for "actual revenue growth data from AI-related products" or more corporate deployment announcements as triggers for further buying. Before that, the volatility in software stocks is likely to continue: on the one hand, technical signs may indicate an "oversold rebound," while on the other hand, funds will continue to make structural shiftspreferring vertically integrated software/data asset companies closely tied to AI training/reasoning systems, as well as platforms that can deliver AI in a "controllable, auditable, integrable" manner. Application layer names with weaker moats, higher homogeneity, and higher valuations will continue to demand higher risk premiums. Therefore, software giants like Microsoft Corporation, MongoDB, Snowflake, Palantir, and SAP, which aggregate data assets and have strong fundamentals, may find it easier to rebound strongly after the panic.