IBM
Century-old US tech company; its AI coding tool now saving .5bn in productivity.
IBM pledged on 1 July to triple its 2026 entry-level intake, a reversal prompted when its own recruitment AI stumbled on the toughest 6% of cases, even as its coding tool squeezed the consulting fees that had justified earlier cuts.
Last refreshed: 17 July 2026 · Appears in 1 active topic
If IBM's own AI saves .5bn, why are investors selling the consulting business?
Timeline for IBM
Mentioned in: A year-old reversal, laundered as news
AI: Jobs, Power & MoneyMentioned in: HCLTech's sharpest fall in five quarters
AI: Jobs, Power & MoneyPledged to triple US entry-level hiring in 2026
AI: Jobs, Power & Money: Ford and IBM start undoing AI cutsMentioned in: BLS still absent; NY Fed fills the gap
AI: Jobs, Power & MoneyReported Q1 2026 earnings with a consulting miss that triggered a 6–8% after-hours share decline
AI: Jobs, Power & Money: IBM beat sold on consulting missBackground
IBM (International Business Machines) was founded in 1911 and is headquartered in Armonk, New York. One of the oldest and largest technology companies in the world, it built the mainframe computing industry, pioneered enterprise software, and later refashioned itself around consulting and hybrid cloud. It operates through two main divisions, Software (including its Red Hat acquisition) and Consulting, and employs around 280,000 people globally. Chief executive Arvind Krishna has led the company since April 2020.
IBM's COBOL maintenance work, servicing the mainframe code still running global banking and government systems, functions as critical infrastructure and gives the company a revenue base that is harder for AI tools to disrupt than project-based consulting work.
That contrast, a durable infrastructure hedge sitting alongside a consulting Arm now exposed to its own AI's productivity gains, is the structural tension running through IBM's 2026 results and its parallel decision to protect entry-level hiring.
Its own AI is squeezing IBM's fees
IBM beat Wall Street's Q1 2026 revenue target on 22 April, reporting $15.92bn against a $15.62bn estimate, yet shares fell 6 to 8% after hours on a narrow consulting miss of $5.27bn against a $5.28bn estimate. Investors read the shortfall as the first hard evidence that AI productivity compresses billable hours rather than expanding IBM's revenue base.
On the same call, IBM disclosed that its internal coding tool, watsonx Code Assistant ('Bob'), delivers a 45% average developer productivity gain and had saved $4.5bn cumulatively since 2023, with generative AI now a third of the consulting backlog. A tool that makes engineers faster also means clients need fewer of IBM's hours, the mechanism the market priced into that after-hours fall.