MYBIGGAMING
LIVE
Articles 16 November 2026 • 6 min read •

Imec and ASRA Unveil ACIR: New Interoperability Standards for Automotive Chiplets

Imec and ASRA introduce the Automotive Chiplet Interoperability Requirement (ACIR) to foster a multi-vendor ecosystem for reliable, safety-critical automotive computing.
Author: Гика • PC •
Imec and ASRA Unveil ACIR: New Interoperability Standards for Automotive Chiplets

As the automotive industry demands increasingly powerful central computing for ADAS, autonomous driving, and digital cockpits, the limitations of traditional monolithic System-on-Chips (SoCs) are becoming impossible to ignore. Rising development costs, complex validation processes, and the sheer scale of modern dies are pushing manufacturers toward chiplet architectures. However, for chiplets to work in a vehicle, they need more than just physical connections; they need a shared language for reliability and safety. To address this, imec and the Japanese organization ASRA have introduced the Automotive Chiplet Interoperability Requirement (ACIR), an open architectural framework aimed at enabling a multi-vendor ecosystem for automotive applications.

The shift toward chiplets is driven by the escalating computational needs of modern vehicles. High-resolution sensor processing, AI-driven decision-making, and advanced infotainment systems require massive throughput that traditional monolithic designs struggle to provide efficiently. In a monolithic SoC, CPU cores, AI accelerators, and I/O logic are integrated onto a single die. While effective for simpler systems, this approach becomes prohibitively expensive as complexity grows, increasing the risk of costly design flaws and making it difficult to utilize the most advanced process nodes for every component. Not every function benefits equally from the latest manufacturing technology; for instance, analog controllers or I/O logic often perform better on older, more mature nodes than on cutting-edge ones.

Chiplet architectures offer a solution by combining specialized dies into a single package. This allows manufacturers to optimize different functional blocks independently and source them from various vendors, potentially reducing costs and increasing flexibility. However, for the automotive sector—characterized by long product lifecycles and extreme reliability requirements—the transition to chiplets introduces a significant hurdle: interoperability. It is not enough for two dies to simply exchange data; they must function as a cohesive, predictable system, especially when safety-critical decisions are at stake.

Beyond Connectivity: The Focus of ACIR

The ACIR framework, introduced by imec and ASRA on November 12, 2026, does not aim to invent a new physical interconnect standard. Instead, it focuses on the systemic requirements that chiplets must meet to ensure they work reliably within an automotive environment. According to the partners, the document covers several critical areas, including error and system management, boot sequences, power management, functional safety, and die-to-die interoperability.

A key distinction made in the ACIR documentation is the emphasis on functional safety. In a data center, a chip error might lead to a system crash or a re-run of a task. In a vehicle, an undefined error state in a chiplet can have life-threatening consequences. Therefore, ACIR specifies that chiplets must have a structured way to report and process errors so that the system can transition into a predictable, safe state. This requires coordinated behavior across different functional units, such as defining which chiplet initializes a subsystem first or how multiple chiplets collectively evaluate a partial failure.

Building a Multi-Vendor Ecosystem

Imec emphasizes that ACIR is intended as an open reference for a wide range of stakeholders, including OEMs, Tier-1 suppliers, semiconductor manufacturers, IP providers, and software developers. Crucially, the framework defines what capabilities chiplets must support without dictating how companies implement them internally. This distinction is vital for a global semiconductor market where a rigid, single-vendor implementation model would be unrealistic. By providing a common architectural framework, ACIR aims to reduce the risks associated with integrating components from different suppliers, thereby accelerating the adoption of chiplet technology in the automotive supply chain.

  • Error and System Management: Common mechanisms for error detection, reporting, and handling.
  • Boot Sequencing: Coordinated startup of multiple chiplets within a single system.
  • Power Management: Coordinated control of power states and energy consumption.
  • Functional Safety: Predictable behavior and safe decision-making during faults.
  • Die-to-Die Interoperability: Common requirements for communication and cooperation between chiplets.

The document was released in its first version to gather industry feedback, meaning ACIR is the beginning of a standardization process rather than its conclusion. The technical challenge lies in the fact that in a multi-vendor environment, different dies may bring different clock and voltage domains, firmware components, and error semantics. Determining which chiplet initializes a subsystem first and who decides on a safe state during a failure remain central challenges for future architectures.

The shift toward chiplets is driven by the escalating computational needs of modern vehicles. High-resolution sensor processing, AI-driven decision-making, and advanced infotainment systems require massive throughput that traditional monolithic designs struggle to provide efficiently. In a monolithic SoC, CPU cores, AI accelerators, and I/O logic are integrated onto a single die. While effective for simpler systems, this approach becomes prohibitively expensive as complexity grows, increasing the risk of costly design flaws and making it difficult to utilize the most advanced process nodes for every component. Not every function benefits equally from the latest manufacturing technology; for instance, analog controllers or I/O logic often perform better on older, more mature nodes than on cutting-edge ones.

Chiplet architectures offer a solution by combining specialized dies into a single package. This allows manufacturers to optimize different functional blocks independently and source them from various vendors, potentially reducing costs and increasing flexibility. However, for the automotive sector—characterized by long product lifecycles and extreme reliability requirements—the transition to chiplets introduces a significant hurdle: interoperability. It is not enough for two dies to simply exchange data; they must function as a cohesive, predictable system, especially when safety-critical decisions are at stake.

Beyond Connectivity: The Focus of ACIR

The ACIR framework, developed by imec and ASRA, does not aim to invent a new physical interconnect standard. Instead, it focuses on the systemic requirements that chiplets must meet to ensure they work reliably within an automotive environment. According to the partners, the document covers several critical areas, including error and system management, boot sequences, power management, functional safety, and die-to-die interoperability.

A key distinction made in the ACIR documentation is the emphasis on functional safety. In a data center, a chip error might lead to a system crash or a re-run of a task. In a vehicle, an undefined error state in a chiplet can have life-threatening consequences. Therefore, ACIR specifies that chiplets must have a structured way to report and process errors so that the system can transition into a predictable, safe state. This requires coordinated behavior across different functional units, such as defining which chiplet initializes a subsystem first or how multiple chiplets collectively evaluate a partial failure.

Building a Multi-Vendor Ecosystem

Imec emphasizes that ACIR is intended as an open reference for a wide range of stakeholders, including OEMs, Tier-1 suppliers, semiconductor manufacturers, IP providers, and software developers. Crucially, the framework defines *what* capabilities chiplets must support without dictating *how* companies implement them internally. This distinction is vital for a global semiconductor market where a rigid, single-vendor implementation model would be unrealistic. By providing a common architectural framework, ACIR aims to reduce the risks associated with integrating components from different suppliers, thereby accelerating the adoption of chiplet technology in the automotive supply chain.

The publication of ACIR follows the strategic partnership announced by imec and ASRA less than a year ago. The document was co-developed by ASRA and imec's expanded Autonomous Edge Chiplet Program. This collaboration brings together two major ecosystems: the European semiconductor research environment and the Japanese automotive and SoC development sector. Strategically, this is significant because the automotive industry is globally organized, while standardization often remains fragmented by region or specific corporations. A joint set of requirements from Europe and Japan increases the likelihood that chiplet architectures in vehicles will not split into isolated technological camps.

Ultimately, ACIR remains a set of requirements rather than a finished, guaranteed market standard. It serves as a foundational reference frame. While it does not guarantee immediate interoperability, it provides the necessary groundwork for a fragmented chiplet ecosystem to evolve into a truly interoperable automotive ecosystem.

Article author

Гика

Cookies

We use cookies

The site uses technically necessary cookies for sign-in, security, storing preferences, and correct page operation. Analytics and other non-essential technologies should only be enabled after your consent. You can read more about data categories, purposes, and retention in the privacy policy.

Cookie policy
Quick actions