Financing the Body, Forgetting the Process: A Demand-Side Pillar for Physical AI and Humanoid Robotics Incentives in Italy

In autumn 2026 the Italian government announced a package of measures for advanced robotics and Physical AI: a dedicated fund of 200-300 million euro in the 2027 budget law, the use of Innovation Agreements (Accordi per l'innovazione) as the delivery vehicle, a publicly controlled "chain leader" to aggregate the supply side, an inter-ministerial committee on labour impacts, and a joint position with France and Germany at European level. This paper argues that the package is necessary but incomplete. It is built on a supply-side logic that finances the production of robotic bodies and the consolidation of a national champion, while leaving unaddressed the problem that has determined the outcome of every Italian technology incentive since 2017: the conversion of purchased capital into productive capacity. Drawing on the final evaluation of the Transizione 4.0 plan, on the evidence on the productivity J-curve, and on the technical literature on humanoid reliability, data scarcity and deployment, the paper identifies the specific features that make Physical AI adoption riskier and more complement-intensive than previous waves of automation. It then proposes a demand-side pillar built around five instruments: a pilot voucher delivered through the existing network of Competence Centres and European Digital Innovation Hubs; a transformation tax credit within the Nuova Transizione 5.0 regime, recognising integration, data and labour costs against a measured baseline; a network of mission testbeds operating under a common evaluation protocol and feeding a national evidence registry; performance-based public first-market contracts; and binding conditions on interoperability, data portability and workforce transition for any platform receiving public money. The paper quantifies the order of magnitude of each instrument, explains how the pillar can be integrated with the announced fund and with the Innovation Agreements without new legislation of general scope, and proposes a set of indicators, including stop rules, to govern the programme. The central claim is that for an immature, learning technology, reducing the risk of adoption is at least as important as reducing the price of the asset, and that the Italian institutional system already possesses most of the components needed to do so. JEL codes: O25, O33, O38, L52, H25, J24. The views expressed are personal and do not represent the position of any institution with which the author is affiliated. Version 1.1 (October 2026). This version integrates the evidence presented by the Italian Parliamentary Budget Office (Ufficio parlamentare di bilancio, UPB) at the hearing of 29 September 2026 before the Tenth Committee of the Chamber of Deputies and in chapter 5 of the UPB Report on budgetary policy 2025 (sections 3.2, 3.4, 3.5 and 4.3); instrument B now takes the form of a tax credit on complementary costs (section 6.2, Table 2, abstract); a comparison with the productivity tax rebate proposal (Baumol, Blackman and Wolff, 1989; Tria, 2026) is added to design principle 3 (section 5). Six references added (53-58). Orders of magnitude unchanged. A colophon page is added. Also available on SSRN, Abstract ID 7554638.