MIPS is celebrating its 40th anniversary as a compute IP company this year. Known for a RISC instruction set, the company has risen and fallen, was private, then owned by SGI, was private again for a while, then Imagination Technologies owned it for several years, sold to a VC, Tallwood, in 2017, and then owned by Wave Technologies briefly before returning to the same VC hands.
In March 2021, MIPS announced that it would cease development of the MIPS architecture as was and transitioning to RISC-V designs. This move marked the end of an era for the MIPS architecture and signaled the company’s adaptation to changing market demands. Most notably, they delivered a high-performance compute with simultaneous multi-threading (SMT) licensed to Mobileye. JPR’s David Harold met with MIPS CEO Sameer Wasson to discuss the company’s progress and strategy.

Various companies have taken different approaches with RISC-V, from GPUs to AI architectures. As MIPS focuses on its new strategy, I’d like to understand where you’ve gotten. What’s the outlook moving forward? And what are your thoughts on the overall market?
Sameer Wasson: From a RISC-V perspective, the past year has felt like 10 years of maturation compressed into one. It’s been a critical year for MIPS as well. I’ll highlight two major accomplishments.
First, we delivered our P8700 core to Mobileye, a demanding automotive customer. They’re about to tape it out for production in their vehicle system. This is a significant technical and financial milestone for us. Strategically, it sets us apart from other RISC-V entities. We’ve moved into a bracket where only a handful of players operate—perhaps just SiFive and us. Completing the last 20% of a CPU, which often requires 80% of the effort, is where we’ve demonstrated maturity.
Second, we’ve refined our focus. Last year, we were exploring what MIPS could be. I was clear about what we didn’t want to be—another SiFive or Arm clone. Instead, we’ve identified application-specific compute as our path forward. The world increasingly needs compute solutions tailored to specific applications. We’re positioning MIPS as the go-to for these cases, whether it’s RISC-V or not. For example, automotive remains a major focus, but we’re expanding into autonomous machines and other areas requiring constrained, application-specific compute.
Is this primarily about automotive and drones?
Wasson: Autonomous systems, whether they move or not, often require complex data processing in constrained environments. They need safety, application-oriented designs, and efficient compute. This spans drones, robotics, and other applications requiring tailored compute solutions. That’s where we’re focusing.

Are you pursuing safety certifications for your cores?
Wasson: Absolutely. Every core we develop is safety-certified. Mobileye’s application, for instance, requires ASIL B certification, and we’ll announce products with even higher safety levels in the future. Safety is non-negotiable for us.
You mentioned refining your focus. How does that translate into 2025 and beyond?
Wasson: In 2025, we’ll continue building on our current vectors, while expanding our portfolio. This includes entering the AI market with new offerings. We’ll announce more specifics at Embedded World. We’re targeting areas where we can differentiate and add significant value, focusing on compute systems that solve real problems—whether technical, commercial, or operational.

What about tools? Are you developing new ones to support your IP?
Wasson: Yes, tools are critical. While custom instructions seemed exciting initially, we’ve realized they’re not a sustainable approach. Instead, we’re developing the Atlas platform, which includes tools like Atlas Explorer. This enables customers to create digital twins of our IP, model performance, and evaluate workloads without revealing sensitive information. It’s cloud-based, cycle-accurate, and will expand to include virtual FPGA platforms.
Editor’s Note: Atlas Explorer is an application that lets you fine-tune software performance using the MIPS Core Performance Simulator without requiring a hardware target or emulator.
How do chiplets fit into your strategy?
Wasson: MIPS isn’t a chiplet company—we’re a compute company. Chiplets are a means to an end, a way to transact business. If chiplets ceased to exist tomorrow, we’d pivot to another model while continuing to solve compute problems. Chiplets help bridge gaps, but our focus is on delivering solutions beyond just CPU cores. We aim to solve subsystem and system-level problems, which requires a different approach to team structure and R&D investment.
How do you see the overall RISC-V IP market evolving?
Wasson: The RISC-V IP market is challenging. You can’t succeed by constantly worrying if your customer will build their own core. That’s not sustainable. Instead, we focus on areas where we can add value beyond just cores. Many companies are cash-rich but engineering-poor after years of relying on Arm. We help bridge that gap by providing complete solutions, not just IP.
You’ve grown significantly. How has your team grown?
Wasson: When we last spoke, MIPS had around 40 people. Now we’re over 200, almost all engineers. They’re not just designing CPUs but also subsystems, low-level software, and AI capabilities. We’re expanding into areas like control and machine learning, aligning with our strategy of driving intelligence into action.
How much of the original MIPS architecture remains?
Wasson: It’s like RoboCop—the heart remains, but everything else has changed. MIPS’ legacy was in solving system-level problems, and we’re continuing that under the RISC-V banner. By upstreaming and collaborating with the community, we leverage the essence of MIPS while adapting to modern challenges like AI and autonomous systems.
How are you navigating geopolitics, particularly in China?
Wasson: We’re selective in our engagements and focus on working with larger OEMs that have multinational ambitions. IP security is always a concern, so we’re emphasizing hardened IP to protect our technology. We’re not chasing every opportunity in China or anywhere else—our strategy is focused and deliberate.
Have I missed anything obvious about MIPS that I should’ve asked?
Wasson: Not really. The key takeaway is that MIPS isn’t just an IP company. We’re a compute company solving system-level problems, with a focus on application-specific solutions. The era of pure-play IP companies is over. Everything’s blurred now, and we’re thriving in that space.
LIKE WHAT YOU’RE READING? INTRODUCE US TO YOUR FRIENDS AND COLLEAGUES.