Why US AI-Robotics Firms Are Scaling in the Gulf, Not From the US

A pure software company can internationalize its market without ever placing a team abroad: the product ships over the internet at near-zero marginal distribution cost. Yet that is not the path chosen by several AI-funded US industrial robotics firms, Gecko Robotics among them, which themselves describe a direct build-out in the Gulf — specifically the United Arab Emirates — for technology that combines a software layer (AI) with a hardware one (robots, drones).

Seen through the lens of international economics, that choice is not trivial. British economist John Dunning proposed a framework in 1977 — the eclectic, or OLI, paradigm — to explain when a firm chooses foreign direct investment (FDI) over exporting or licensing to a local partner. His answer: a firm opts for FDI only when three conditions hold simultaneously — an Ownership advantage it alone holds, a Location advantage that makes the host country more attractive than the home country for exploiting that advantage, and an Internalization advantage that makes managing the asset in-house preferable to handing it to a third party through a license.

The ownership advantage here is obvious: the robot-sensor-AI stack for inspecting critical infrastructure is proprietary technology the firm does not want leaking to a local competitor. It is the easiest condition to verify, and the least informative — it holds for almost any technology firm that internationalizes at all.

The location advantage is more instructive, because it reverses the textbook FDI-to-emerging-markets story. In the standard account, a Western firm relocates to exploit cheaper labor; here, the Gulf offers neither cheap labor nor a mass consumer market comparable to the US. It offers something else: access to patient, abundant capital via sovereign wealth funds (Mubadala, ADQ, PIF), markedly faster regulatory approval cycles for dual-use technology (civilian and defense), and — critically — real industrial assets, from refineries to power plants to mining and defense sites, offered up as full-scale testbeds. This is a capital- and speed-driven location advantage, not a labor-cost one.

The internalization advantage, finally, explains why these firms choose to scale directly rather than license their technology to a local distributor — a simpler, less capital-intensive option. The answer lies in the nature of the asset: inspection technology for sensitive sites (defense, energy) generates data on the integrity of critical infrastructure — data whose leakage or misuse by a third-party local partner would be a largely uninsurable risk. Keeping direct control of operations, via a subsidiary or majority joint venture, reduces that risk, at the cost of a heavier capital and organizational commitment than licensing would require.

This triptych carries a direct practical implication for any Western firm — French firms included — weighing a move into the Gulf, whether in aerospace, energy or industry: Dunning's framework is a ready-made decision checklist. Is the ownership advantage distinctive enough to justify direct investment over licensing? Does the Gulf offer a genuine location advantage for this specific activity — capital, regulatory speed, access to testbed assets — or is this simply a fashionable move? Does the sensitivity of the technology justify the cost of internalization over a cheaper local partnership?

One caveat is warranted: the source of this observation is itself a firm with a commercial stake in the trend it describes — Gecko Robotics has an obvious interest in framing the Gulf as the destination of choice for AI robotics. A single case is not proof of a systematic capital flow. It nonetheless sits within a broader, separately documented pattern (Gulf fund stakes in Western AI champions, large compute campuses announced in the UAE) that makes the underlying hypothesis plausible beyond this one example.

For Cardan-AI's target sectors — aerospace and defense, energy and oil & gas, industry, luxury — the lesson is not to rush to the Gulf, but to track the pace at which competitors are field-validating their own industrial AI there at scale: an accelerated iteration cycle on real foreign assets can, with a lag of a few quarters, become a domestic competitive gap that is hard to close.