The Announcement
Ainekko, a startup focused on open AI infrastructure, has contributed its CORE-ET Silicon Platform (ETSP) to the OpenHW Foundation, a RISC-V-centric open hardware initiative hosted under the Eclipse Foundation. The platform combines a many-core 64-bit RISC-V processor architecture, MRAM-based intelligent memory, and open developer tooling designed specifically for low-power AI inference at the edge. The move places silicon-level AI hardware IP under neutral governance with permissive Solderpad 2.1 licensing, and the underlying design is already progressing through tape-out on a 16nm process node. For the edge AI market, this is a meaningful signal: open-source principles are now extending below the software stack all the way to transistors.
The Bigger Picture
Open Silicon Follows the Open Software Playbook
Ainekko’s CTO frames this as history repeating itself, and he’s not wrong. The open-source transitions in operating systems and cloud platforms both followed a pattern: proprietary stacks dominated early markets, then open alternatives commoditized the baseline and shifted competition to higher layers. We’re watching the same dynamic begin to play out in AI silicon, particularly for edge deployments where power budgets are tight and deployment environments are wildly heterogeneous.
The RISC-V ecosystem has been building toward this moment for years. What makes the CORE-ETSP announcement more than a press release is the combination of factors it brings together: a concrete architecture with vector and SIMD extensions tuned for inference, an architecture emulator for pre-silicon experimentation, and the institutional weight of OpenHW Foundation governance. The last point matters more than it sounds. Open hardware projects frequently stall not because the technology is weak but because IP ownership ambiguity and governance gaps make enterprises reluctant to build on them. Solderpad 2.1 with Eclipse Foundation stewardship aims to remove that friction.
The tape-out confirmation on 16nm is also notable. Many open silicon projects remain perpetually in the RTL stage. The fact that CORE-ETSP is moving toward physical silicon accelerates the credibility timeline considerably.
What This Means for ITDMs
For IT and operations leaders evaluating edge AI infrastructure, the framing question is usually: what happens to our hardware dependency when a single vendor controls the roadmap? That question becomes acute as AI inference moves into factories, logistics systems, and autonomous equipment. The ability to inspect, modify, and contribute to the hardware architecture is not an academic concern when a production line depends on a particular inference profile.
The open model Ainekko is proposing shifts the economics of edge AI silicon in a meaningful way. Procurement teams will recognize this pattern from the software world: open platforms tend to reduce per-unit licensing costs while expanding the ecosystem of compatible tooling and support partners. The governance structure under OpenHW Foundation provides the neutral IP framework that enterprise legal teams require before they will sanction deep architectural dependency.
There is also a supply chain angle. Proprietary AI accelerators from established vendors carry concentration risk. An openly governed silicon architecture, especially one that can be licensed and manufactured by multiple foundries, distributes that risk across a broader supplier base.
What This Means for Developers
For engineers building edge AI systems, the CORE-ETSP architecture emulator is the most immediately useful artifact in this announcement. Hardware bring-up is typically one of the most costly and time-consuming phases of embedded AI development. An architecture emulator that allows AI inference frameworks to be ported and tested before physical silicon is available compresses that timeline substantially.
The many-core RISC-V architecture with vector and SIMD extensions is well-suited to the parallelism that neural network inference demands. The inclusion of a scalable network-on-chip with integrated MRAM-based memory is technically interesting because MRAM offers non-volatility with near-SRAM latency, which has direct implications for power-efficient inference at the edge where battery or harvested energy budgets are often the binding constraint.
The co-design model Ainekko is promoting, where RTL and tooling evolve alongside the AI software stack, represents a genuine architectural philosophy shift. Developers who want to optimize an inference workload all the way down to the memory topology or the instruction scheduling layer can do that here. That’s a capability essentially unavailable in proprietary accelerator ecosystems.
ECI Research data reinforces the demand context here. According to ECI Research’s 2025 AI Builder Summit survey, 59% of organizations are investing in Agentic AI for IT Operations today, and many of those deployments will ultimately require inference at the edge rather than round-tripping to a centralized cloud. Separately, an ECI Research analysis found that over 80% of mid-market and enterprise organizations have launched or plan to launch AI/ML initiatives in the next 12–18 months, with 62% citing AI as a strategic priority. That scale of deployment ambition will not be served entirely by cloud-resident inference; silicon-level efficiency at the edge will be a prerequisite for the latency and cost profiles these initiatives require.
Competitive Positioning
What Ainekko is attempting is vertically integrated: open architecture, open tooling, open developer community, all organized around a specific inference performance profile. The AI Foundry community platform they previously launched suggests they understand that the value in an open ecosystem accrues to whoever organizes the developer community around it, not just whoever designs the chip.
The primary competitive risk is speed to viable ecosystem. Open hardware projects can take years to accumulate the breadth of toolchain support, reference designs, and production deployments that make them genuinely compelling to risk-averse enterprise buyers. The 16nm tape-out timeline and the OpenHW Foundation’s existing community infrastructure both help, but Ainekko will need to demonstrate working silicon in production-representative deployments to move beyond early adopters.
Looking Ahead
Near-Term: Ecosystem Formation and Developer Adoption
The next 12–18 months are essentially a proof-of-ecosystem phase for Ainekko and CORE-ETSP. Success here will be measured by the number of AI inference frameworks ported to the platform using the architecture emulator, the quality of the open toolchain, and the diversity of contributors in the OpenHW repository. The RISC-V Summit Europe appearance in Bologna this June is a logical first organizing moment for attracting the European industrial AI community, which has been among the most active in exploring RISC-V for embedded applications.
ITDMs evaluating edge AI hardware platforms should begin tracking this project now, even if procurement decisions are 18–24 months away. The governance model and licensing terms are already in place, which means diligence work done today will not be wasted.
Longer Term: The Open Silicon Stack Thesis
The deeper thesis here is that the AI infrastructure stack, currently proprietary from silicon to framework, will ultimately follow the same open-source trajectory as the web server stack or the container orchestration stack. If that thesis is correct, the organizations that build early relationships with open silicon ecosystems will have structural cost and flexibility advantages over those locked into proprietary hardware by the time these platforms mature.
ECI Research projects the AI-native development platform market will reach $9.8 billion by 2026, growing at a compound annual rate of approximately 38%. A meaningful fraction of that growth will be driven by edge deployments where open, efficient silicon architectures like CORE-ETSP are directly competitive. Ainekko is early, but the structural tailwinds are real, and the governance infrastructure is now in place to support serious enterprise engagement.
Stay Ahead of Application Development Trends
Get weekly analyst insights, research notes, event coverage, and AppDevANGLE updates delivered directly to your inbox.
Subscribe for Weekly Insights
Join technology leaders, practitioners, and GTM teams following the trends shaping modern software delivery.
Looking for deeper research access?
Explore ECI Research reports, survey insights, and market analysis through the ECI Research Portal.
