S05.3 · Industrials & Advanced Manufacturing
A $40-50B, high-margin instrument segment where AI is starting to generate test protocols and detect anomalies once done by hand.
Test, measurement and instrumentation — the equipment that measures, monitors and calibrates physical processes — runs $40-50B in 2025 (Fortune Business Insights / Grand View Research). It carries the highest gross margins in industrials, built on IP-heavy instrument design, and that is exactly where the pressure is arriving: AI is starting to write the test protocols and detect the anomalies that test engineers have historically produced by hand, which puts the labor line and the IP moat in play at the same time.
Growth of 5-6% historically is forecast at 6-8%, carried by semiconductor test demand, EV and battery testing, and aerospace and defense programs. Revenue spreads across North America, Europe and Asia; semiconductor test specifically concentrates in Taiwan, Korea and Japan; R&D and production center on the US, Germany and Japan.
Keysight, National Instruments (now part of Emerson), Fortive's Fluke business, Rohde & Schwarz and Yokogawa hold roughly 30-35% combined, with a long tail of specialty sensor and calibration houses filling out the rest. Upstream inputs are semiconductors and precision optics and sensors; downstream, instruments reach semiconductor, automotive, aerospace and pharmaceutical manufacturers through calibration labs and integrators.
Gross margins of 50-60% lead the industrials sector, a direct read on IP intensity. Aftermarket calibration and service contracts contribute 20-30% of revenue and recur, which cushions the segment against the capex cycle in a way pure capital-equipment categories cannot match — though demand still swings moderately with semiconductor capex specifically. NIST and ISO 17025 accreditation gate calibration services, a durable barrier around that slice of the business.
Three adjacencies matter: industrial automation, which consumes the sensor feeds directly; testing services, principally calibration labs; and data and analytics software, increasingly built on the test data these instruments throw off. Calibration service contracts defend recurring revenue that accreditation already protects. Test-data analytics platforms are the expansion — the move up the stack from measurement into interpretation and insight, where the incremental dollar is software-priced rather than instrument-priced.
Agentic AI bites first at test-sequence generation and anomaly detection. Generating a protocol straight from a product specification, with no test engineer writing it by hand, compresses a labor line that is a meaningful share of program cost in semiconductor and aerospace test programs specifically. It also drains a moat that has protected established players for decades: the proprietary test script. When effective test sequences no longer require years of accumulated in-house scripting expertise, one source of switching cost holding customers to the incumbent instrument vendor goes with it.
The boundary now forming separates hardware instrument makers from a newer set of AI-native test-data-analytics platforms that compete on what can be inferred from the data rather than on measurement precision. Automated test-script generation is commercially deployed today. Closed-loop AI test optimization — the system adjusting the test process from its own results, without an engineer in the loop — is a 2-5 year horizon.