"Selling software, flooding into semiconductors" is this really the version answer? Salesforce, Inc. (CRM.US) performance outlook lower than expected! "AI disrupt everything" narrative targeting CRM leader.

Focus on customer relationship management (CRM) software field of the American cloud software giant Salesforce, Inc. (CRM.US) gives current quarter performance outlook slightly below the consensus expectations of Wall Street analysts, undoubtedly causing more anxiety to investors who have been worried about the potential disruption of the entire software industry by artificial intelligence since February. Goldman Sachs Group, Inc.'s tracking of 13F data shows that under the catalyzing theme of the pessimistic narrative of "AI disrupting everything," hedge funds significantly sold off software stocks in the first quarter while dramatically increasing their exposure to semiconductors. Hedge funds' long positions in semiconductors have reached the highest level on record at Goldman Sachs Group, Inc., while their software exposure has dropped to the lowest level since 2019. With the lackluster performance outlook of CRM software leader Salesforce, Inc., software stocks may face a new round of intense selling. Salesforce, Inc.'s latest performance outlook falling below expectations, with RPO data falling short of Wall Street expectations, triggered not just concerns about the performance of a single software company, but rather a collective skepticism among investors about the entire application software sector: if AI agents like Claude Cowork and OpenClaw can directly complete traditional software workflows such as customer service, sales, coding, data analysis, and backend analytics, then traditional SaaS software vendors' profit growth models reliant on seat-based charging, modular expansion, and manual workflows may be completely disrupted. Therefore, the data flow from Goldman Sachs Group, Inc. shows that global hedge funds and mutual funds have been synchronously selling off software stocks and moving towards semiconductors since the beginning of this year, expressing this disruption logic through their investment actions. Salesforce, Inc. in a performance outlook statement on Wednesday Eastern Time stated that in the second quarter ending in July, the midpoint of the expected total revenue is expected to be approximately $11.3 billion (the specific revenue outlook range provided by the company is $11.27 billion to $11.35 billion), compared to the Wall Street analysts' average expectation of $11.4 billion. This expectation has even been revised down since February. In terms of first-quarter performance indicators, Salesforce, Inc.'s overall performance is still on a growth trajectory, but the forward-looking indicators and second-quarter guidance are not enough to alleviate the valuation fears of "AI disrupting software stocks." In the first quarter of the fiscal year 2027, the company's total revenue is $11.133 billion, a 13% year-on-year growth, higher than market expectations of approximately $11.05 billion; adjusted EPS is $3.88, significantly higher than the market expectation of approximately $3.13. The company also reported that the first-quarter remaining performance obligations (RPO), an important indicator measuring future sales, is approximately $67.9 billion, an 11% year-on-year growth, compared to Wall Street analysts' average expectation of $68.9 billion. The outlook for the next quarter and the quarterly RPO indicators falling short of market expectations highlight the AI intelligent agent tool, Agentforce, that the company is actively promoting, which has not received positive recognition from major customers compared to the global spread of AI agents like Claude Cowork and OpenClaw (formerly known as Clawdbot, Moltbot). From being a CRM leader to facing the pressure of the AI era, software stock valuation logic is reshaped by AI intelligent agents Salesforce, Inc. is the absolute leading software manufacturer in the customer relationship management software field, and is currently under tremendous pressure from new AI forces like Anthropic, needing to prove its ability to flourish in the AI era. The company is promoting Agentforce, its proprietary AI agent tool focused on agent-based AI workflows, aimed at handling core business tasks such as customer service without the need for human supervision. However, there were media reports last week indicating that the capabilities of this AI agent product do not always match the company's advertising and marketing promises. With performance outlook and RPO falling short of expectations, the company's stock price fell almost 4% in after-hours trading, closing at $177.51 at the end of Wednesday's trading day. The company's stock price has fallen by about 33% since the beginning of the year, a decline similar to other traditional software giants such as ServiceNow Inc. and Adobe Inc., significantly underperforming the S&P 500 index and the Nasdaq 100 index. The over 30% decline in Salesforce, Inc.'s stock price since the beginning of the year reflects the "higher growth proof threshold faced by traditional software/internet operating companies in the AI era." The pessimistic tone of the "AI disrupting everything" since February is mainly due to the market's increasing concerns that AI agent workstreams like Claude Cowork and OpenClaw, which are viral and rapidly spreading, may weaken the entire software empire based on SaaS seat subscription revenue models, module expansion, and manual workflow dependence, ultimately disrupting the entire business model. Therefore, the data flow from Goldman Sachs Group, Inc. indicates that global hedge funds and mutual funds have been synchronously selling off software stocks and moving towards semiconductors since the beginning of this year, expressing this disruptive logic through their investment actions. Salesforce, Inc.'s management currently expects Agentforce to contribute $1.2 billion in annual revenue, slightly higher than the $800 million estimated in February. The company noted that the overall use of AI large models within the Salesforce, Inc. software platform has more than doubled compared to the previous quarter. The performance report also shows that the combined ARR of Agentforce and Data 360 is close to $3.4 billion, a year-on-year growth of over 200%; the combined delivery of 3.8 billion Agentic Work Units by Agentforce and Slack is up 111% month-on-month; the cumulative processing of Tokens has exceeded 28.6 trillion, an increase of 152% month-on-month; Data 360's single-quarter ingestion of records is 520 trillion, a year-on-year growth of 136%, of which Zero Copy ingestion was 350 trillion, a year-on-year growth of 277%. As AI agents like Claude Cowork and OpenClaw increasingly penetrate the enterprise operations end, the market demands Salesforce, Inc. to prove faster that Agentforce can become a new entry point in the AI era, rather than being bypassed by native AI agents. It is clear that Salesforce, Inc.'s AI-related business is indeed a highlight, but the current performance outlook and the highly anticipated RPO indicators failing to meet market expectations highlight that Salesforce, Inc.'s AI-driven growth trajectory is not yet strong enough to reverse the pessimistic narrative of "AI disrupting everything." Agentforce is not your ordinary chatbot like Siasun Robot & Automation, but rather Salesforce, Inc.'s attempt to upgrade CRM, Data 360, Slack, and enterprise workflows into a product called "AI Agent Execution Layer." Its goal is to enable AI agents to handle tasks such as customer service, sales follow-ups, data queries, and work order circulation instead of humans. Therefore, Agentforce is essentially Salesforce, Inc.'s AI agent platform/tool focused on agent-based AI workflows for enterprises: it features customizable autonomous AI agents that connect enterprise data, automatically answer questions, execute actions, and drive tasks in scenarios such as sales, customer service, marketing, and business. Salesforce, Inc. also describes Agentforce as a "proactive, autonomous AI application," stating that it allows companies to build autonomous AI agents that can take action across various business processes. Mike Spencer, Executive Vice President of Salesforce, Inc., stated in an interview, "Some customers are increasingly confident in AI, which helps boost the company's performance. We are seeing a certain level of adoption and usage, which is anchoring acceleration." For the fiscal quarter ending April 30, Salesforce, Inc. saw total revenue grow by 13% to $11.3 billion. This result was driven by $444 million in revenue brought in by the acquisition of data software company Informatica. According to compiled data from institutions, excluding certain items, adjusted earnings per share (EPS) is $3.88, higher than the average analyst expectation of around $3.13. Analyst Raimo Lenschow from Barclays PLC Sponsored ADR (Barclays) wrote in a report that Agentforce's performance did not significantly boost overall data. "We are not sure if this is enough to drive a meaningful market reaction," Lenschow said. Robin Washington, Chief Financial and Operations Officer of Salesforce, Inc., stated in a declaration that the company still expects revenue to accelerate in the second half of this year driven by AI agents. Earlier this month, the company announced that it would provide fewer detailed revenue specifics related to specific products. Now, the company will report performance growth data in two main categories: applications, and infrastructure and data. Salesforce, Inc.'s management stated that, excluding the impact of currency fluctuations, application type revenue grew by 7%, while infrastructure and data revenue increased by 23%. "Panic of AI disrupting everything" sweeps through the software industry! Anthropic's soaring valuation behind the intelligent agent era's software application layer repricing The reason why the narrative of the "AI disrupting everything" led by Anthropic has severely impacted Salesforce is that the market is concerned that the traditional SaaS model of "seat-based charging + manual workflow + CRM system entry" is being revalued by AI agents. When AI agents like Claude Code, Claude Cowork, OpenAI Agent etc. start directly completing tasks such as sales, customer service, coding, document processing, data analysis, investors are very worried that businesses will no longer pay for SaaS seats in large numbers, but turn to the AI-native work layer where they pay based on tasks/tokens/results. From a software engineering perspective, the threat of agents like Claude Code/Claude Cowork is not about "replacing a CRM database", but rather replacing the operational layer between users and software: salespeople, customer service representatives, and operations personnel who used to click, input, query, build reports, and trigger processes in Salesforce may in the future be replaced by AI agents who directly call APIs, access enterprise data, generate action suggestions, and automatically execute tasks. In other words, Salesforce, Inc. is not at risk of its database being replaced, but rather the "interaction interface, automation logic, and workflow control" being shifted by AI agents. Some academic research also shows that agents like Claude Code possess architecture characteristics such as running commands, editing files, calling external services, and executing tasks in loops, which are the technical basis for Agentic AI to infiltrate traditional enterprise software workflows. Enterprises' urgent need to improve efficiency and reduce operating costs has recently vigorously promoted two core categories of AI applications - generative AI applications and AI agents. AI agents (i.e., AI agents) that independently perform a variety of tedious and complex tasks are highly likely to be the ultimate trend in AI applications for the next decade. The emergence of AI agents means that artificial intelligence is evolving from an information assistance tool to a highly intelligent productivity tool, which is why Anthropic's valuation has skyrocketed to $1 trillion, surpassing OpenAI. The AI agent tool launched by Anthropic has been an important trigger for the panic selling of software stocks since February, but more accurately, it is not that Anthropic "alone disrupts the entire software stocks," but that its Claude agent AI tool has made the market suddenly realize that large model companies are completely disrupting the enterprise application layer from the "bottom-up model supplier," potentially eroding the profit pools of traditional SaaS software companies focusing on legal, sales, marketing, data analysis, and low-code editing. Fear of AI disrupting the narrative has torn apart tech stocks, and funds vote with their feet: hedge funds massively sell software stocks and flood into semiconductors The current market is reconsidering AI from being a "software efficiency tool" to a "software business model disruptor + amplifier of semiconductor capital expenditure." This is why Goldman Sachs Group, Inc.'s data tracking of 13F shows a rare consensus among hedge funds and large mutual funds in the first quarter, wherein they significantly sold off software stocks and flooded into semiconductors, pushing semiconductor long positions to historic highs. Global funds are actively betting on the logic more solid, scarce, and difficult to replace links in the global AI value chain, rather than the traditional application software that still needs to prove the resilience of the Agent-era business model. This "selling software, flooding into semiconductors" investment strategy can be considered the most perfect answer in the current stock market investment arena. Goldman Sachs Group, Inc.'s tracking data from 13F shows that institutions are significantly reducing software exposure and increasing semiconductor exposure: hedge funds' semiconductor long positions have reached the highest level on record, while their software exposure has dropped to the lowest level since 2019; mutual funds' software positions have dropped to the lowest level since 2012, and after excluding Microsoft Corporation, mutual funds' relative over-allocation to semiconductors has reached the largest level since 2012. In other words, funds are moving from the "application layer that may be restructured by AI agents" to the "AI agent's reliance on computing power." From an engineering perspective, this round of portfolio shifts has a clear technical foundation. The stronger the capabilities of AI agents, the greater the demand for AI inference computing power dominated by AI GPU/AI ASIC, high-performance CPUs in data centers, DRAM/NAND/HBM storage, AI PCBs, liquid cooling systems, advanced packaging equipment, gas turbines, data center fiber optic interconnection systems, etc.; while software companies need to prove that they are not replaced by agents, they will increasingly need to purchase more computing power, access stronger models, or use AI computing infrastructure to restructure product architecture. This means that the other side of "AI disrupting software" is, in fact, "AI strengthening chip demand." Software stocks are facing pressure from rewritten valuation models, while semiconductor stocks enjoy support from visible capital expenditures, supply constraints, and profit upgrades. Morgan Stanley indicates that the AI computing arms race has entered a phase of system-level expansion, with the institution revising upward its expected capital expenditures of U.S. tech giants to $805 billion for 2026, up from $433 billion just a year ago, and is expected to reach $1.1 trillion for 2027, surpassing the previously predicted $950 billion, with a forecast that by 2028, nearly $3 trillion in AI-related infrastructure investments will flow through the global economy, with over 80% of spending still ahead. Morgan Stanley's latest expectations highlight that the supply chain bottleneck in AI computing infrastructure has expanded from "large-scale GPU/ASIC purchases" to "simultaneously solving the complete chain of AI data center delivery processes, from data center power equipment, liquid cooling, data center CPUs, DRAM/NAND/HBM, optical communication/optical interconnection, high-performance network interconnection, transformers, gas turbines, etc.".