AI Isn't Hurting IBM—Its Expensive Hardware Is

For years, IBM has positioned itself as one of the biggest beneficiaries of the enterprise AI revolution. From watsonx and Red Hat to consulting and quantum computing, the company has repeatedly argued that AI would fuel its next phase of growth. Ironically, it was the same AI spending boom that delivered IBM one of its most painful quarters in recent memory.

The technology giant reported second-quarter revenue of $17.16 billion, up just 1% year over year but below Wall Street expectations, while adjusted earnings of $2.93 per share also missed estimates. More importantly, IBM cut its 2026 revenue growth forecast to 4-5%, down from its earlier expectation of more than 5%.

The disappointing results came just a week after CEO Arvind Krishna took the unusual step of issuing a preliminary earnings warning—a move that wiped nearly $70 billion or 25% off IBM's market value. It also resulted in the company's steepest single-day stock decline in more than a century.

Why IBM is Hurting?

Unlike many technology companies hurt by AI disruption, IBM's challenge isn't that customers are abandoning its software or replacing workers with generative AI.

Instead, enterprises are redirecting IT budgets toward expensive AI infrastructure. Demand for GPUs, high-speed servers, networking equipment, memory, and storage has surged as companies race to build AI data centres.

"In the last few weeks of June, we saw clients shift their quarterly capex spend toward servers, storage, and memory purchases to secure supply-constrained infrastructure ahead of expected price increases. This dynamic impacted client buying patterns," Krishna said in a letter to investors.

Those purchases, made more urgent by rising hardware prices and supply constraints, have squeezed budgets for traditional enterprise systems, including IBM's flagship mainframes.

IBM executives said "tens" of customers delayed planned mainframe purchases during the quarter after prioritising AI infrastructure investments.

"While we anticipated some supply chain-related impact in our expectations, we did not anticipate the magnitude of the capex reprioritisation," Krishna added.

Revenue from IBM's Z mainframe business plunged 42%, dragging overall infrastructure revenue down 7%. Since IBM typically earns roughly three dollars in software revenue for every dollar of mainframe hardware sold, the slowdown rippled across multiple business lines.

Yet the picture wasn't uniformly bleak. Software revenue increased 5% to $7.76 billion, driven by hybrid cloud offerings and Red Hat, while consulting remained broadly stable. IBM's distributed infrastructure business—including Power servers and storage—grew 37%, building a record backlog approaching $500 million.

Temporary Pause or Structural Shift?

IBM executives insist the weakness reflects timing rather than declining demand. Krishna said many delayed transactions have already begun closing during the current quarter, describing the missed deals as deferred rather than cancelled. CFO James Kavanaugh similarly argued that customers continue expanding mainframe workloads for AI, analytics, Linux and mission-critical applications rather than moving away from the platform altogether.

Industry analysts largely agree that IBM's problems appear tied to changing enterprise spending priorities rather than a collapse of its core business. The mainframe has repeatedly survived predictions of its demise over several decades and remains central to banking, government and large enterprise computing.

"IBM is not broken, but it needs to prove that its reinvention can deliver consistent growth. Until then, the market may treat IBM as a mature enterprise technology company with strong long-term options rather than a pure AI growth winner," Will Conaway, Tuxedo Cat Consulting President, said.

The broader lesson extends beyond IBM. The AI boom is reshaping enterprise technology budgets in unexpected ways. Rather than simply creating new winners and losers, it is forcing companies to rethink where capital is allocated. Spending that once flowed toward software upgrades and infrastructure refreshes is increasingly being redirected toward AI chips, servers and networking.

"For years, IBM made its money by being the middleman - helping companies set up and manage complex software But today, companies want raw AI hardware, and they want it immediately. This crash is a warning sign for the entire tech sector," Anushikha Dwivedi, a banking sector executive, said.

For IBM, the challenge now is navigating this transition without losing momentum in its traditional businesses. Whether the current slowdown proves to be a short-term pause or an early signal of a lasting shift in enterprise IT spending may become clearer over the coming quarters.