Autopsy: International Business Machines Corp (IBM)

Sector: Computer and office equipment, enterprise software and services (SIC 3570)
Revenue band: ~$62B (FY2025)
Previously covered: No

The announcement side

April 9, 2026: Rob Thomas, SVP IBM Software and Chief Commercial Officer, published a piece introducing "IBM Enterprise Advantage" for building internal AI platforms (Source: newsroom.ibm.com press release index; finance.yahoo.com summary).
March 25, 2026: IBM announced an ElevenLabs integration with watsonx Orchestrate to deliver multilingual conversational experiences at scale (Source: newsroom.ibm.com/press-releases-artificial-intelligence).
March 17, 2026: IBM completed its acquisition of Confluent, a data streaming platform reported to serve more than 6,500 enterprises (Source: newsroom.ibm.com, March 17, 2026).
March 5, 2026: IBM announced a partnership with Deepgram to support its generative AI solutions (Source: finance.yahoo.com, March 5, 2026).
October 8, 2025: IBM and S&P Global jointly announced an agentic AI deployment for supply chain and vendor risk, combining watsonx Orchestrate with S&P Market Intelligence data (Source: press.spglobal.com, Oct 8, 2025).
February 12, 2026: IBM said it would triple entry-level US hiring in 2026, framing the move around AI reshaping early-career roles (Source: Bloomberg; hrexecutive.com).

The reality side

The watsonx.ai product page documents "AI Agent Development" with deploy paths from IDE, low-code Agent Lab, and AI service notebooks for single-agent and multi-agent applications (Source: ibm.com/products/watsonx-ai/ai-agent-development; ibm.com/docs/en/watsonx/saas).
One named, publicly disclosed customer-facing agentic deployment surfaced in this corpus: S&P Global (Source: press.spglobal.com, Oct 8, 2025). No additional named enterprise customer agent deployments appeared.
The hiring increase referenced January 2026 industry data from ManpowerGroup/Experis, which is sector-wide, not IBM-specific. The IBM-specific commitment is a volume increase, not a disclosed agent engineering headcount (Source: hrexecutive.com).
February 23, 2026: IBM stock posted its steepest daily drop since 2000 after Anthropic published claims that its models can modernize COBOL, a workload tied to IBM's mainframe and consulting business (Source: reuters.com/company/international-business-machines-corp).
The FY2025 10-K filed February 24, 2026 was retrieved as XBRL metadata only in this corpus. No narrative MD&A or risk-factor text on AI investment, capex, or governance was accessible from the snippet provided, so no inference is drawn about 10-K disclosure in either direction (Source: sec.gov filing 000005114326000010).

Keep reading these weekly.

One research note per Thursday morning. An autopsy of a large-firm AI announcement paired with the shipping-evidence read of a quiet mover in the same sector.

One email per week. Unsubscribe at any time.

The gap

IBM's announcement cadence is steady: roughly weekly press activity tied to AI, a large acquisition (Confluent), three named partnerships in a six-month window (S&P Global, Deepgram, ElevenLabs), and a hiring narrative that frames AI as expansionary. The reality side surfaces one named enterprise agentic deployment in production (S&P Global), a documented but generic agent-building product on watsonx.ai, and a market reaction on February 23 suggesting investors see IBM's modernization and mainframe revenue as exposed to third-party AI tools. The delta is between IBM's positioning as, in one analyst's headline, an "AI growth machine" (Source: seekingalpha.com), and the relatively thin set of named, in-production agentic customer deployments visible from public sources.

One caveat worth naming: IBM's largest agentic work likely sits inside Consulting engagements that are covered by NDAs and rarely become press releases or product pages. The S&P Global disclosure is closer to an exception than a template for how that revenue is reported. The absence of more named deployments in this corpus is not, by itself, evidence that more do not exist.

Why this matters for the enterprise reader

Two patterns repeat for the CFO or COO reader. First, the announcement cadence keeps the investor narrative active, but the count of named, in-production customer deployments is much smaller than the count of press releases. Second, when a competitor (Anthropic) made a single targeted claim against a core IBM workload, the market repriced IBM by the steepest single-day drop in 26 years. Together, the two patterns suggest the question worth putting to any AI vendor is not what they have announced this quarter, but which named enterprise customer is currently running their agents in production, and which systems those agents are allowed to touch.