Enterprise

IBM Stops Reporting Cumulative AI Revenue Wins

The tech giant quietly retired its $7.5 billion generative AI bookings metric in favor of mix percentages and product-level disclosures.

Omega Editorial· September 4, 2026· 3 min read

IBM drops its signature AI metric

International Business Machines has stopped disclosing the cumulative dollar value of generative AI work it has secured, a metric the company had tracked publicly since 2024. The figure reached $7.5 billion inception-to-date as of July 2025, but disappeared from the company's July 2026 earnings call without explanation.

The shift came in the same quarter IBM reported performance below investor expectations, according to a Trefis analysis. Management had previously highlighted the metric as recently as the prior quarter, making the $3 billion figure reported in October 2024 and the $7.5 billion total in July 2025 visible benchmarks for the company's AI momentum.

New disclosure approach focuses on percentages

IBM now reports generative AI activity through business mix and product adoption rather than absolute dollars. In the June 2026 quarter, generative AI represented approximately half of Consulting signings and over 30 percent of the Consulting backlog. Unlike a cumulative total that can only increase, these percentage-based metrics can move in either direction.

The company's Consulting segment generates roughly $21.1 billion annually, representing about 30 percent of IBM's $69.1 billion revenue base. In the June 2026 quarter, Consulting signings grew 6 percent while revenue increased just 1 percent at constant currency, indicating AI work is accumulating in bookings before converting to recognized revenue.

Management now emphasizes specific products, particularly watsonx Orchestrate, which the company positions as a control plane for building and governing AI agents. IBM also highlights its recurring software revenue base, which accounts for approximately 80 percent of annual software revenue through subscriptions and consumption models across Red Hat, HashiCorp, and Confluent. That recurring base carried $24.6 billion in annual recurring revenue, up 8 percent year over year, while transactional software revenue declined in the high single digits during the June 2026 quarter.

Why it matters

The metric change makes it harder to track whether IBM's generative AI business is accelerating or plateauing. A cumulative total provides transparency on absolute growth; percentage-based disclosures can obscure underlying trends if the denominator shrinks. For investors evaluating IBM's AI positioning against competitors, the loss of a simple, comparable figure complicates cross-company analysis. The timing—coinciding with a performance miss—raises questions about whether the change reflects a strategic communications shift or an attempt to manage expectations around AI revenue conversion rates.

Where the AI work surfaces

IBM's Consulting revenue remains the primary place generative AI activity should materialize. The segment grew 2 percent over the trailing twelve months, and management projects low to mid-single-digit constant-currency growth for full-year 2026. If generative AI work continues to represent half of new signings, Consulting revenue trajectory becomes the key indicator of whether bookings are converting to recognized revenue at expected rates.

IBM shares have declined 1.3 percent over the past year, compared to a 21.1 percent return for the S&P 500. At $234.71, the stock trades approximately 28 percent below its 52-week high, though it remains up 24.5 percent over two years.

These details were first reported by Trefis.

#ibm#generative ai#enterprise ai#financial reporting#consulting services#watsonx

This is an original analysis by the Omega editorial team. Source reporting: AI Watch.

Want systems like this working for your business?

Book a Call

More in Enterprise

Enterprise· 3 min read

AI Inference Shifts Data Center Design from Compute to Memory

As enterprises deploy real-time AI workloads, data movement and storage architecture become the critical bottleneck—not raw processing power.

Via AI Watch · Sep 4, 2026
Enterprise· 3 min read

AI-Generated Code Requires Specialized Review Tools, Says Expert

Redesign Health's Argus system targets the predictable errors that AI coding assistants consistently make.

Via AI Watch · Sep 4, 2026
Enterprise· 3 min read

Zscaler's $246M Net New ARR Surge Tests AI Security Budget Reality

The cloud security vendor posted its tenth consecutive earnings beat, but rising capex and decelerating ARR growth will determine whether the AI narrative justifies the valuation.

Via AI Watch · Sep 4, 2026