IBM Falls 25% as Customers Shift AI Budgets to Hardware

IBM shares plummeted 25% on July 14 following the release of preliminary second-quarter results that revealed a significant shift in how enterprise customers are allocating their AI and technology budgets. CEO Arvind Krishna attributed the decline to a structural reorientation: as hardware prices have continued to rise, customers are prioritizing spending on servers, storage, and memory infrastructure rather than software solutions. This shift is causing material delays across numerous large software deals, particularly in IBM's core mainframe and transaction-processing business segments.

The strategic implication, according to Krishna on the company's earnings call, is that "value will increasingly shift toward the orchestration and data layers"—a recognition that while enterprises must invest in foundational hardware, they ultimately need software that integrates and governs that infrastructure. IBM is attempting to position itself in this emerging layer through several initiatives: Sovereign Core, which provides auditable runtime control; Confluent, recently acquired for data streaming capabilities; and Lightwell, representing a $5 billion（約8000億円） commitment to help clients address open-source vulnerabilities. However, IBM's position as a multisegment business—with separate software, consulting, and infrastructure divisions each operating on different performance cycles—complicates its ability to fully capitalize on a single market trend.

In contrast, Palantir Technologies is structured as a pure-play provider of data integration and AI-powered enterprise applications. The company's platforms, Foundry and the Artificial Intelligence Platform, are designed to integrate an organization's data and run AI applications. In May, Palantir reported 85% revenue growth, marking its 11th consecutive quarter of acceleration. The growth has driven strong operating leverage: net margins expanded from 44% to 60% year over year. Customer metrics are particularly telling: net dollar retention reached 150%, and total remaining deal value climbed 98% to nearly $12 billion（約1.9兆円）, indicating that existing customers are significantly increasing their spending even in an environment where hardware is consuming a large share of AI budgets.

The contrast between IBM's struggle and Palantir's momentum highlights the market's current preference structure. The trade-off for Palantir, however, is valuation: the stock trades at roughly 39 times this year's sales and 67 times free-cash-flow guidance. IBM, by comparison, is cheaper at 17 times forward earnings, but its multisegment structure means each segment will perform differently in an AI-enabled future. While Palantir's shares have declined 30% year to date and are beginning to look more reasonable relative to their prior highs, the company's concentrated exposure to the orchestration-and-data-layers thesis remains a substantial valuation bet.

IBM's July earnings stumble reflects a structural shift in how enterprises are allocating AI budgets in 2024. As CEO Arvind Krishna explained on the earnings call, customers are prioritizing the acquisition of hardware infrastructure—servers, storage, memory—at a time when hardware prices remain elevated. This infrastructure spend is crowding out investment in software, particularly in IBM's traditional strongholds of mainframe and transaction-processing software, causing large deals to slip. This dynamic does not suggest a decline in enterprise AI demand overall; rather, it indicates that the near-term capital allocation cycle is weighted heavily toward the foundational layer.

Krishna's statement that "value will increasingly shift toward the orchestration and data layers" signals IBM's recognition of this moment. The company is investing in software that sits above raw infrastructure—governance (Sovereign Core), data streaming (Confluent), and vulnerability management (Lightwell)—but IBM's position as a multisegment incumbent means its software, consulting, and infrastructure divisions each follow their own performance cycles, making it harder to capitalize on a single macro trend.

Palantir Technologies, positioned as a pure-play provider of enterprise data integration and AI-powered applications through products like Foundry and its Artificial Intelligence Platform, is seeing different dynamics. The company's 85% revenue growth in May, combined with net dollar retention of 150% and nearly $12 billion（約1.9兆円） in total remaining deal value, suggests that existing customers are continuing to increase their spending—implying that at least some enterprises are funding the software orchestration layer even amid infrastructure budget pressures. However, Palantir's valuation at roughly 39 times this year's sales reflects the market's high expectations for continued growth.

No comments yet. Be the first to share your thoughts!