Teece and complementary assets: why Emerson is building a marketplace around its AI

Emerson announced on August 11, 2026 the launch of PACEdge 3.0, a new version of its industrial IoT (IIoT) software platform that aggregates data from disparate technologies and makes it available to AI analytics, machine-vision, and visualization applications running directly at the edge of oil, gas, and industrial sites.

The detail worth dwelling on is not the AI analytics function itself — by now a category expectation in IIoT — but the architecture Emerson wraps around it: containerized applications, a group manager to control entire fleets of edge devices from a single interface, and above all a third-party application marketplace designed as a central repository for discovery and deployment.

That last piece is what clarifies the strategy. In 1986, economist David Teece published "Profiting from Technological Innovation," now a classic article addressing an empirical paradox: the firms that innovate are not always the ones that capture the value of their innovation. Teece showed that capture depends on two variables: the appropriability regime (how easily the innovation can be imitated) and ownership of complementary assets — distribution, service, manufacturing, software platforms — without which an innovation cannot reach the market.

AI applied to industrial analytics is, by itself, a weak appropriability regime: machine-learning models, computer-vision techniques, and predictive-analytics tools are widely taught, published, and increasingly available through open-source libraries. Any sensor vendor or systems integrator can, in principle, plug an AI model into an industrial data stream. The risk for a player like Emerson is that the value of industrial AI slides toward pure-AI specialists who need not own any physical infrastructure to sell their software layer.

Emerson's response, read through Teece, is to build the complementary assets that AI alone cannot supply: physical access to sensors and controllers already installed at the customer site, the ability to manage a fleet of edge devices remotely and securely, and above all a marketplace that positions Emerson as the mandatory gateway between third-party AI application developers and industrial operators. Rather than defending a proprietary AI model — a risky bet in a field where open source moves fast — Emerson is betting on control of distribution.

That choice carries strategic weight beyond Emerson's own case. It illustrates a pattern spreading across heavy industry as it confronts AI: holders of physical complementary assets (installed base, customer relationships, process know-how, regulatory certifications) can capture a disproportionate share of the value created by a technology they did not invent, provided they build the distribution platform before someone else does. The multi-device group manager is not a minor technical detail: it is the complementary asset that makes the marketplace useful, since it solves the operational problem — managing a heterogeneous, dispersed fleet — that AI alone does not solve.

For industrial leadership teams weighing their own edge-AI investments today, Teece's question remains the right one: who in the value chain already owns the complementary assets needed to deliver this AI — and how urgent is it to build them in-house before a platform vendor claims that position instead?

One limit of the exercise: the source article discloses no pricing and no named customer for PACEdge 3.0, which makes it impossible to quantify how far this platform strategy has actually been adopted at this stage.