ERICI Pushes RISC-V Into Factories to Cut Europe's Chip Dependence

What ERICI actually is

ERICI wants to give industrial RISC-V a common, tested baseline instead of a pile of vendor-specific boards.

The European RISC-V Industrial Computing Initiative, or ERICI, is not a chipmaker and not a standards body. It describes itself as a non-commercial effort to build a tested reference stack for RISC-V industrial computers, or RICs, running partitioned RVA23 systems. In plain terms: a known-good combination of hardware and software that a factory can deploy without reinventing the wheel.

That matters because industrial computing is a graveyard of good-enough boards with terrible software support. A PLC or vision controller might run for 15 years, and the company that made its chip may be gone in five. ERICI's answer is to anchor the stack on RISC-V, where the instruction set has no license fee, no single owner and no single country behind it.

The RVA23 profile is the modern application-class baseline for 64-bit RISC-V, and 'partitioned' suggests the design splits workloads across isolated domains, a sensible fit for mixed real-time and general-purpose factory tasks. ERICI has not published a full hardware bill of materials or a certification program yet, so treat the scope as a stated intent rather than a shipping product.

Why Europe is pushing now

Chip sovereignty has moved from a policy slogan to an engineering requirement.

The pitch lands at an awkward moment for Europe's industrial base. x86 is effectively an Intel and AMD duopoly, and Arm licenses its architecture under terms that can change with ownership and export rules. RISC-V is the only mainstream instruction set that none of them controls, which is why it keeps showing up in sovereignty conversations.

ERICI's framing is unusually blunt for a European initiative: a machine expected to run for 15 years should not depend on one company's roadmap or one government's decisions. That is a direct reference to the export-control and supply-chain shocks of the past few years, and it is the kind of argument that gets traction with plant managers who have been burned by long lead times.

The initiative is not alone in this lane. The EU has been funding RISC-V work through research programs and the EuroHPC ecosystem, and companies like SiFive and European IP firms are already selling RISC-V cores. What ERICI adds is the industrial angle: a reference stack aimed at the factory floor rather than a benchmark slide.

The hard part is software, not silicon

RISC-V can already run Linux. Industrial reliability is a different bar.

Getting RISC-V to boot Linux on a dev board is a solved problem. Getting it to run a 15-year production line with deterministic timing, certified toolchains and long-term security patches is not. Industrial buyers care about mean time between failures, not gigahertz, and they care about a vendor that will still answer the phone in 2038.

That is where a tested reference stack earns its keep. If ERICI can publish a validated combination of RVA23 silicon, a real-time-capable hypervisor or partitioning layer, and a supported Linux distribution, it lowers the risk for machine builders who would otherwise bet on a single board vendor. The open question is who maintains it and how conformance is verified, because 'non-commercial' and '15-year support' are not natural bedfellows.

There is also the performance gap to acknowledge. RISC-V application cores are improving fast, but x86 and Arm still lead in raw single-thread performance and in the breadth of industrial software certified for them. ERICI's bet is that factory workloads are modest enough, and longevity important enough, that the trade is worth it.

What would count as success

A reference stack only matters if machine builders actually ship it.

The clearest signal would be a European machine builder putting an ERICI-conformant RIC into a shipping product, with a published support window. The second would be a certification or test suite that other vendors can run against, turning 'tested' from a claim into a checkable property.

Until then, ERICI is a well-argued position paper with a reference architecture attached. That is not nothing. The industrial market is conservative, and a neutral, non-commercial baseline is exactly the kind of thing that can shift procurement conversations away from defaulting to x86.

The initiative's own framing is the one to watch: no license, no single owner, no single country. If it can deliver the tested part of that promise, European factories get a real third option.

Availability and pricing

ERICI is a non-commercial initiative, not a product you can buy. There is no price, no release date and no announced hardware partner list, so availability for US, UK or EU buyers is not yet defined.

For European manufacturers, the practical takeaway is to watch for a published reference stack and conformance criteria. Those are the artifacts that would let a machine builder evaluate RISC-V industrial computers against existing x86 and Arm options.

ERICI is early, and early initiatives deserve skepticism. But the problem it names is real: industrial machines outlive the companies that built them. A license-free, multi-vendor RISC-V baseline is a credible answer, provided someone actually maintains it for the next decade and a half.