IBM (IBM.US) performance fell more than 20% before the explosion! AI computing power and hardware siphon corporate budgets “software stock apocalypse theory” resurface

The Zhitong Finance App learned that the preliminary quarterly sales results announced by the established US tech giant IBM (IBM.US), which has the two most cutting-edge technology horizons of artificial intelligence and quantum computing on Tuesday, fell short of Wall Street analysts' unanimous expectations. The company's CEO Arwin Krishner said that customers are drastically tightening their spending and attributed the poor performance to chips and servers on a large scale. After the announcement of preliminary results that fell short of expectations, the tech giant's stock price, with a market capitalization of nearly 300 billion US dollars, fell sharply by more than 20% before the US stock market, and disrupted the technology sector of the US stock market. In particular, technology stocks focusing on AI application software and AI computing power infrastructure themes fell collectively.

IBM said in a preliminary results statement on Tuesday that it expects the company's second-quarter revenue to be around US$17.2 billion, compared to the average Wall Street analysts' estimate compiled by the agency of about US$18 billion. According to the statement, sales of IBM's infrastructure business unit were particularly hard hit and unexpectedly fell by about 7%. The company said it is still reviewing the accounts and the final results may be slightly different.

Krishna said that in order to cope with supply shortages across the industry, customers generally shift capital expenditure to AI servers, data center enterprise-grade NAND memory devices, and high-performance DRAM memory components, thereby weakening their spending and budget on the company's software products.

“These environments require perfect execution by our team, and we made mistakes this quarter,” Krishna said. “Our adjustments and actions were not quick enough, and a number of large transactions were not completed in time as we expected, which contributed to most of this performance gap.”

IBM's performance that fell short of expectations highlights that AI computing power hardware is gradually “siphoning” enterprise IT budgets, thus reigniting the “software stock apocalypse theory.” It also dragged down software stocks such as ServiceNow, Oracle, and Adobe, proving that corporate budgets are indeed shifting from the “application layer” to the “computing power base” front. However, IBM's mistakes also include poor sales execution and failure to execute large orders on schedule; AI replacement cannot all be blamed; the more likely pattern is low barriers, per-seat charges, and software that can easily be copied by agents, while software that controls core data, workflow, compliance, security, and infrastructure management capabilities gets stronger pricing power.

IBM may hit the biggest one-day decline in decades, and Workday and ServiceNow are under pressure at the same time

Before the opening of the New York Stock Exchange, the company's stock price fell sharply by more than 21% in pre-market trading. If this decline persists, the stock will record its biggest single-day intraday decline since the 1980s. This performance also dragged down other popular technology companies. Workday's stock price fell by more than 8% before the market, and ServiceNow's stock price fell by about 7.7%.

The blow to IBM's hardware sales may hinder its efforts to reinvent itself as a high-growth software company through large-scale acquisitions of Red Hat, HashiCorp, and Confluent. Even the company's new strategic focus has made it a source of concern for investors, who fear that artificial intelligence tools will replace many existing software products. In February of this year, artificial intelligence startup Anthropic PBC released a tool that may help modernize an old programming language on IBM mainframes. Since then, IBM's stock price has been drastically sold off.

IBM CEO Arvind Krishna said that the company had previously anticipated supply chain issues to drag down performance. The shortage of memory chips caused by the boom in artificial intelligence has had a ripple effect in the global manufacturing industry. However, he said the company did not anticipate that customers would eventually shift spending from IBM products to large-scale purchases of servers, storage equipment and memory to hedge against the risk of further price increases.

“The actual situation was worse than we expected, and we miscalculated,” Krishna said in a letter to investors, adding that the company's Z series mainframes and related popular software made up most of the performance gap. “These environments required perfect execution from our team, and we made mistakes this quarter. Our adjustments and actions were not quick enough, and many large transactions were not completed in time as we expected.”

The company also initially reported that after dilution, earnings per share could fall by about 2% to $2.27.

Like most software vendors, IBM has fully integrated cutting-edge artificial intelligence technology into its products and touted its ability to provide customers with the latest technology. The company has been trying to convince investors that artificial intelligence will enhance its business rather than replace it. IBM executives have stated many times that work related to artificial intelligence will increase the demand for infrastructure software related to IBM's underlying AI applications. Such software allows customers to use leading artificial intelligence models.

A major shift in AI capital expenditure has begun. AI computing power hardware eats meat, and software is being judged by valuation! Global tech stocks face a “profit must be perfect” stress test

The core significance of IBM's warning is that its AI and quantum computing technical narratives have yet to be fully translated into actual revenue and order execution that can withstand quarterly fluctuations. Initial sales for the second quarter were only 17.2 billion US dollars, lower than the market's growth forecast of about 18 billion US dollars, while IBM's revenue in the first quarter still increased 9% to 15.92 billion US dollars, and software sales increased 11.3%. The contrast between before and after shows that the market previously gave it a “high-growth software company,” and there is not much room for fault tolerance.

What is more noteworthy is that management also acknowledged that customers prioritize their budgets for servers, storage, and memory, as well as IBM's own failure to adjust sales strategies in a timely manner and delay multiple large orders; this means that the performance gap is due to both external budget migration and internal execution errors, and may not be simply due to the macro environment.

For IBM, quantum computing is still a forward option with extremely high strategic value and a long commercial return period. It cannot replace the growing responsibilities of Red Hat, automation, data infrastructure, consulting, and mainframe software in the current income statement. IBM has announced that it will invest more than $10 billion over the next five years to advance quantum computing, with the goal of building large-scale fault-tolerant quantum computers by 2029.

However, what the capital market currently needs to verify is not the technology roadmap itself, but whether companies can use hybrid cloud, artificial intelligence governance, data orchestration, and mission-critical infrastructure to transform enterprise AI investment into continuous software subscriptions, consulting sales, and free cash flow. Artificial intelligence is also a double-edged sword for IBM: it can increase enterprises' demand for infrastructure software, security, and model management, but it also reduces manual labor hours for traditional code maintenance, consulting, and system modernization. Old code modernization tools introduced by companies such as AnthroPic have prompted investors to re-evaluate the economic moat of IBM's mainframe ecosystem.

IBM's initial performance fell short of expectations, which could lead to budget reallocation and further fragmentation of valuations within the technology sector: customers snapping up servers, memory and storage, validating demand for semiconductors, server manufacturers, and data center supply chains, but temporarily squeezing traditional software, consulting, and digitalization projects that can be delayed. As a result, IBM is more like an alarm for the enterprise software and information technology services sector, rather than evidence of the collapse of demand in the entire AI industry; the real danger signal will be that cloud vendors such as Microsoft, Amazon, and Google cut capital expenses simultaneously, or that orders for servers, HBM, and network equipment weaken simultaneously. This combination has not yet appeared. The market expects S&P 500 earnings to increase by more than 23% in the second quarter, and profits in the technology sector to increase by more than 65%. Therefore, the biggest risk now is not that there has been no growth, but that growth must continue to exceed already extremely high expectations.

IBM's stall is more likely to be a catalyst for the phased pullback of overvalued technology stocks and the rotation of the “sell software, buy hardware” style, rather than triggering the systemic collapse of global technology stocks alone. According to the Bank of America's latest survey, fund managers' cash positions have dropped to 3.6%, and the bulls and bears have reached an extremely optimistic area of 9.4, indicating that market positions are already crowded and the buffer against negative surprises is very thin; in this environment, as long as a company “performs well but falls short of perfect expectations,” the decline may be significantly amplified.