Will RISC-V reduce auto MCU’s future risk? – Monthly Billet

By going RISC-V, suppliers of automotive MCUs based on proprietary processor cores hope to avoid getting designed out by OEMs and excluded from the huge Chinese market.

With billions already sunk into its proprietary TriCore MCUs, Infineon’s RISC-V shift reflects a huge change in the company’s strategy.

Companies such as STMicroelectronics, Renesas and NXP Semiconductors all say they are already using RISC-V cores. But those RISC-V cores are deeply buried in their architecture for specific embedded applications, not the kind that’s an open-source instruction architecture exposed to application developers.

Both NXP and STMicroelectronics have invested in Quintauris, a joint venture that includes Robert Bosch GmbH, Infineon, Nordic Semiconductor ASA, and Qualcomm Technologies, Inc. STMicroelectronics joined Quintauris, because “the RISC-V ISA has interesting features that have value in specific use cases.” But calling Arm cores and ecosystem as “STMicroelectronics’ choice for General Purpose (STM32) and automotive MCU roadmaps,” STMicroelectronics keeps mum on its plans to diversify its ISA options.

Renesas continues to develop, for automotive, MCUs with the company’s proprietary RH850 core, as well as MCUs and SoCs based on Arm cores. The company is still “evaluating automotive RISC-V cores and preparing our roadmap for RISC-V intersection as demand emerges.”

Infineon is preeminent in the real-time, secure and dependable MCU market. So, why risk a brand-new family of automotive RISC-V MCUs?

Thomas Schneid, Infineon’s senior director for SW & Ecosystem Development in Infineon’s Automotive Microcontroller Business Unit, acknowledged Infineon’s success and growth in the high-grade automotive market.

“But there is this big BUT,” he noted. “We are seeing that the automotive market is demanding more and more ‘something on top’ in the way of how we were successfully serving this market (with dependable MCU products).”

Schneid wakes up at night as he can’t stop thinking about OEMs who want “much more openness, flexibility and ‘pick-and-choose’ capability.”

“We understood they want to have their own way of orchestrating their software supply chain”, Schneid explained.

The problem is that such customers’ demands are “inverse” to what Infineon offers today with its TriCore MCU based on proprietary architecture, Schneider acknowledged. TriCore relies on a black box setup that ensures dependability and security for Infineon’s automotive MCUs.

In recent years. Infineon had to seriously ponder its options. One is an organic path, making TriCore even better and stronger.  Another is to adopt Arm, already a large presence in automotive.

In the end, Infineon chose RISC-V over TriCore or Arm because of “a specific footprint we have already established in the dependable MCU market segment.” Schneid saw the move to RISC-V “as expanding our flexibility moving from proprietary architectures to open standard.”

With RISC-V, “we can shape the future” and “influence the new way of doing business” in automotive.

Moreover, the RISC-V shift doesn’t just solve technological issues. “This is also about how we transform our business in the market,” Schneid said.

The automotive [MCU] business is “no longer focused on black box,” he stressed. “This will become a more open collaboration space where many players can contribute.”

Infineon envisioned “an opportunity to orchestrate and drive decision making,” he explained.

As it turns to RISC-V, what will Infineon then do with its own TriCore?

AURIX products based on TriCore and new derivatives which are in the pipeline will be still an Infineon offering for at least next 15 to 20 years, said Schneid. But it “will not continue to evolve afterwards”

Translation: Infineon will commit its resources for MCU development into RISC-V, not TriCore. Even then, TriCore MCUs “will still find their opportunities and projects for real time dependable control” in the embedded market, Schneid added.

Most MCU suppliers in the West prefer not to frame their MCU decisions in the context of growing RISC-V momentum in China. Infineon insists that it made the choice because RISC-V customers want it for their software defined vehicles, “regardless of their location.”

But here are some facts.

China already has many small RISC-V developers. Tom Hackenberg, principal analyst in computing and software for the Yole Group, explained that suppliers, even Western MCU leaders, can’t afford to be excluded from market demands for RISC-V ecosystem growth in China.

China’s RISC-V motivation is clear. It wants an ecosystem less beholden to Western technology providers.

RISC-V cores are already installed in Chinese vehicles.

Yu Yang, principal analyst for automotive semiconductors at Yole Group, offered examples.

Beyond these examples, “There are many, many smaller China companies,” noted Hackenberg. “Wuhan Binary Semiconductor, ChipEXT, CCore Technology, and Yuntu are examples of companies developing RISC-V,” he added.

Hackenberg also reminded that Qualcomm, NXP, Bosch, ST and Nordic Semiconductor “all have joint ventures active or in plans to serve mainland China.”

Local JVs and partnerships could prove crucial for the western companies, when they locally develop RISC-V IP.

Manuel Alves, senior vice president and general manager of Automotive Microcontroller at NXP, observed that Chinese want RISC-V cores and their IP, but they are even more concerned about where that IP was developed by which vendors.

The main RISC-V battleground in China won’t be in the classic MCU market.  Instead, Alves believes the demand in China lies in RISC-V-based solutions in high performance compute space.

That isn’t Quintauris’ initial focus, though.

A few days before Infineon’s RISC-V MCU announcement, Quintauris launched its first RISC-V profile, RT-Europa, tailored for real-time, safety-critical automotive applications.

Quintauris specified “real time” in the “automotive” segment to go first with its RISC-V profile. Automotive is the first market that “has already shown interest in transitioning into RISC-V,” said Pedro Lopez, business development head. In parallel, many vendors considering RISC-V are working on real-time processing.

Lopez said Quintauris also has the application processor market on its RISC-V roadmap, but later.

Why?

Barriers to the RISC-V market are much lower with real-time processing compared to application processors. The more mature ecosystem for embedded software development is primed to the RISC-V transition, compared to an application processor market that involves larger companies and groups like Google, and Linux, he said.

After the launch of the RT-Europa profile, Quintauris is developing a platform based on it. The JV will be unveiling the new platform at CES in 2026. It will be based on the first profile in the form of reference boards featuring IP and software from different vendors, said Lopez.

The reference boards based on the profile are a crucial step for RISC-V to transition to open ISA MCUs to which developers can write a variety of software.

At Quintauris, “we are creating a benchmark as an independent entity for software and IP vendors,” to ensure performance and proper validation of the software stack, Lopez explained.

By promising “software interoperability” and avoiding “vendor lock-ins”—two priorities for Quintauris — Lopez believes semiconductor companies will feel confident embracing the platform.

Despite Quintauris’ efforts, flipping the automotive market to RISC-V won’t be easy. Yole Group’s Hackenberg said, “Not likely in the next decade, if ever.”

“RISC-V has a lot of room for growth,” but “there will be huge competition and hurdles before trying to displace Arm.”

Hackenberg noted that RISC-V would benefit from synergy if Chinese auto sales go global. But this eventuality also faces many hurdles.

Source: www.yolegroup.com