The AI Manufacturing Panel Nobody Asked For — And What It Reveals About CNH, IBM, and ITW

A manufacturing trade show in Illinois is hosting a keynote panel this October called "AI in Manufacturing." It features executives from three public companies — IBM, Illinois Tool Works, and CNH Industrial — moderated by the editor of ASSEMBLY magazine. The press release, distributed Tuesday through Accesswire, describes AI in manufacturing as moving from "experimentation to real-world applications".

That framing matters because investors have been paying for it. CNH Industrial, the maker of Case and New Holland farm and construction equipment, has surged 46% this year. Its panelist is a senior manager in continuous improvement. The stock's run appears to reward a narrative the company itself promotes: autonomous combines, AI-guided farming, smart factories.

The problem is the gap between what gets staged and what gets deployed.

The US Census Bureau's Business Technology and Operations Survey, released in May 2026, found roughly 18% of US firms use AI in any business function — and manufacturing adoption trails even that average. Under the Census's earlier question wording ("in producing goods or services"), the national rate was closer to 10%. The headline figures you see elsewhere — 77%, 95% — count investment intent or vendor-commissioned surveys of large firms, not production deployment.

Among very large manufacturers ($500M+ revenue), Deloitte's 2025 survey found 29% have AI deployed at facility or network level and 24% have deployed generative AI. Meanwhile, 38% are still piloting gen AI. A Q2 2026 survey of 500 manufacturing executives found that 47% are stuck in continuous pilot mode — running experiments indefinitely without advancing to scaled deployment. Only 31% report measurable P&L impact.

MIT's NANDA initiative, based on 153 senior leaders across 52 organizations, found something starker: only about 5% of integrated AI pilots reach measurable financial return. The technology works. The economics don't follow because manufacturers evaluate pilots on whether the model performed, not whether the investment justified itself against a capital budget.

The panel tells you who wants to be associated with the trend. The data tells you who is actually earning from it.

IBM is the company on the panel with a real AI revenue stream. Its software segment grew 11% in Q1 2026, driven by watsonx, and the company reported $15.9 billion in quarterly revenue. Year-to-date, IBM's revenue growth sits at nearly 8%, with free cash flow growing 13%. The company has 58% gross margins and operates with a 17.5% operating margin.

Yet IBM's stock is down 21% year-to-date, trading around $234 — well below its 52-week high of $332. The market has stripped the AI premium from a company that is, by the panel's own framing, executing on it. The disconnect between IBM's software performance and its stock suggests investors are no longer willing to pay a narrative multiple for enterprise AI that runs on consulting cycles and deployment timelines rather than product revenue and margin expansion.

Illinois Tool Works offers the most honest version of AI in manufacturing. ITW — the diversified industrial behind fasteners, welding systems, and food processing equipment — runs an 80/20 operating model across 80+ business units that has delivered 26.4% operating margins and 28% returns on invested capital. Its panelist, Roy Devadas, heads global IT strategy. AI at ITW means predictive maintenance and incremental efficiency gains in factories that are already highly optimized.

ITW's stock is up 7.7% year-to-date, a reasonable return for a company growing revenue at 4% and growing free cash flow at 7%. It's not an AI story. It's a capital-allocation story that uses AI the way it uses everything else — quietly, to protect margins. The company generates $409 million in annual capex against $1.8 billion in inventory. Every dollar is accounted for.

CNH Industrial is where the narrative and the economics diverge.

CNH makes tractors, combines, and construction equipment under the Case and New Holland brands. Its panelist, Girish Gopalakrishnan, is a senior manager in continuous improvement — the kind of role that pilots AI for ergonomics analysis and production optimization. The company's own press materials promote autonomous combines and AI-guided farming. At its 2025 Tech Day, CNH showcased "customer-centric innovations across AI, Autonomy, Robotics and Automation".

The stock has rallied from around $9 to its current level near $13.50 — up 46% year-to-date and 34% over the last four months. The move has been fueled by the autonomy narrative: CNH unveiled an R4 autonomous robot for specialty crops and has promoted AI-driven farming through partnerships and product showcases.

The company's operating economics don't support the premium. CNH runs a 10.8% operating margin — less than half ITW's and two-thirds of IBM's. Returns on invested capital sit at 5.1%, meaning the company earns barely more on its capital than a savings account yields. Free cash flow growth is down 71% year-over-year. The company carries $5.2 billion in inventory against $1.2 billion in annual capex, a ratio that suggests either a cyclical buildup or working-capital friction.

Revenue growth is 0.6% year-over-year — effectively flat. The 25% quarter-over-quarter revenue jump reflects seasonality in agricultural equipment, not structural acceleration. CNH's Q2 2026 earnings showed $4.8 billion in revenue, up 2% from a year ago, with adjusted net income of $161 million.

The stock's trajectory looks like what happens when a cyclical equipment company rides an innovation narrative while its underlying margins and returns stay where they were five years ago. Investors reward the autonomy roadmap and the AI farming vision, but those investments — R4 robots, sensor systems, autonomous platforms — are capital expenditures that reduce free cash flow before they generate revenue. CNH's 71% decline in FCF growth is the arithmetic of innovation spending that hasn't yet produced returns.

What to separate when evaluating the three.

IBM is executing on AI revenue but is being treated as if the execution has a ceiling. The company's problem isn't deployment — it's that enterprise AI sells in consulting cycles, and consulting margins don't carry the same premium as software licensing. The stock's decline despite double-digit software growth suggests the market has recalibrated its view of what IBM's AI contribution is worth.

ITW is the baseline case for how industrial companies actually use AI — incrementally, to protect an already superior operating model. There's no stock pop to be had here because the market already prices ITW correctly as a high-return industrial. The company's 28% ROIC tells the whole story.

CNH is the risk. The 46% rally rewards a roadmap that exists as capital spending, not as revenue. The company's 5% ROIC and collapsing free cash flow are the kind of economics that don't change because an executive sits on a panel about AI in manufacturing. The autonomy investments are real, but they are also expensive, and the agricultural equipment cycle is flat. A cyclical company trading at a growth premium is a valuation mismatch that resolves when the cycle turns — or when the narrative does.

The ASSEMBLY Show's press release describes AI moving from experimentation to real-world applications. For most manufacturers, it hasn't. The Census says under 10% use AI in production. Deloitte says under 30% have deployed it at facility level. MIT says 95% of pilots don't reach P&L impact.

Investors who buy the stock moves implied by those panels are paying for the stage, not the deployment.