NEURA and SECO: Europe's Compute Sovereignty Play for Robotics

The Core Question: Why Compute Hardware Origin Just Became a Boardroom Issue

The partnership between NEURA Robotics and SECO marks a turning point: the compute modules powering cognitive robots — the chips and boards that translate perception into action — are being developed, engineered and manufactured entirely within Europe. For decision-makers planning physical AI investments over the next 12 to 24 months, this is more than a supplier announcement. It is a concrete building block of a European answer to a question that has so far gone unaddressed: where does the hardware making autonomous decisions on your factory floor actually come from — and what does that mean for compliance, dependency and support?

What NEURA and SECO Announced on 7 September 2026

On 7 September 2026, NEURA Robotics of Metzingen, Germany, and SECO of Arezzo, Italy, announced a strategic partnership. SECO will handle development, engineering and manufacturing of the electronic compute modules for NEURA's cognitive robots, including the humanoid 4NE1. NEURA frames physical AI as three interlocking technology layers: sensing, compute and cloud infrastructure. With this partnership, the compute layer is now covered end-to-end within Europe — a departure from most current robotics stacks, which rely on US or Asian chip suppliers.

Both companies also plan to use production data from real-world NEURA robot deployments to develop new physical AI automation solutions for industry — a sign that the partnership aims not only for hardware sovereignty but also for a distinct European data and learning loop.

Three Reasons Compute Hardware Provenance Is Turning Strategic

• Compliance: an increasing share of robotics applications fall into higher-risk regulatory categories in Europe; knowing the provenance and traceability of your compute hardware substantially reduces the effort required to demonstrate conformity.
• Geopolitical dependency: export controls and sanctions regimes increasingly touch semiconductors originating in the US and China; a European supply chain lowers this risk for long-term capital committed to robot fleets.
• Support paths: shorter physical and legal distances between manufacturer and operator shorten response times for maintenance, spare parts and software updates — critical for assets running at high utilisation.

This compliance dimension is not a footnote. Regulatory requirements for high-risk AI systems in the EU are tightening noticeably, and any organisation investing in physical AI today should factor governance obligations in from the outset rather than retrofitting them later.

Production-Scale Proof: BMW, Siemens and Otto Group

The NEURA-SECO partnership lands at a moment when physical AI is moving from pilot projects into production. In September 2026, BMW Group Plant Leipzig launched Europe's first pilot with a humanoid robot — the Hexagon AEON from Hexagon Robotics Zurich, introduced in June 2025 — in collaboration with Hexagon, BMW's long-standing partner for sensor technology and software.

Siemens supplies figures that underline the same maturity: in April 2026, its Erlangen plant tested the HMND 01 Alpha wheeled humanoid robot — built on the NVIDIA Physical AI stack — in logistics operations.

The Otto Group underscored how seriously large European corporates now treat physical AI with this investment into a proprietary 'Robotic Coordination Layer' designed to control all of the group's robot fleets in real time. Momentum is visible outside Europe too: on 3 September 2026, CJ Logistics in Korea deployed two humanoid dual-arm robots running its own robot foundation model in live packaging operations. And NEURA itself expanded its portfolio in August 2026 by acquiring Bosch's ACTIVE Shuttle business, an established autonomous transport platform with an existing installed base in factories and warehouses.

What This Means for Swiss Physical AI Investments

For Swiss companies planning robotics investments, the question is shifting. It is no longer only whether a robot can technically perform the required task, but who developed the compute hardware behind it, under which jurisdiction it was manufactured, and how that affects auditability, spare parts availability and regulatory documentation obligations. This supply chain governance perspective used to be a semiconductor procurement concern — it is now a matter for the executive board.

How seriously Europe actually intends to build an independent AI and robotics industry is being debated well beyond the trade press — worth a listen is the discussion in Where does Europe really stand on artificial intelligence?, which takes a sober look at the opportunities and limits of European sovereignty.

How fast this field is moving globally is also visible in Asia, where humanoid deployments are scaling at a pace that European and Swiss decision-makers cannot ignore — though it should not be copied uncritically either.

Conclusion: The Pragmatic Next Step

The NEURA-SECO partnership is not a reason to delay physical AI investments — it is a signal to approach them more deliberately. European compute supply chains reduce risk; they do not eliminate it. Organisations that now evaluate which robot platform, which compute architecture and which governance structure fit their operations gain an advantage that will be difficult to close within two to three years.