AheadComputing Develops RISC-V CPUs for AI with Ex-Intel Engineers | Hardware near me | Gpu benchmark | Intel hardware | Turtles AI

AheadComputing, a Portland-based startup founded by former Intel executives, has raised $21.5 million to develop CPUs based on RISC-V, an open source architecture that offers greater flexibility and efficiency for AI computing. The funding is led by Eclipse Ventures and involves high-profile investors.

Key Points:

• Innovative technology: Development of RISC-V CPUs to improve performance of AI workloads.
• Experienced leadership: Founders with decades of experience in semiconductor design.
• Solid financial backing: $21.5 million raised from prominent investors, including Jim Keller.
• Strategic goal: Overcoming the limitations of traditional CPUs for AI and data centers.

AheadComputing secured initial funding of $21.5 million to develop CPUs based on RISC-V architecture, with the goal of optimizing performance in AI-related workloads. The startup was founded in 2024 by a group of top engineers from Intel, including Debbie Marr, Jonathon Pearce, Mark Dechene and Srikanth Srinivasan, who have a combined experience of more than 100 years in the semiconductor industry. Their mission is to address the limitations of traditional CPUs in AI systems, particularly in bandwidth, power efficiency and scalability. The RISC-V architecture was chosen for its open and independent nature, allowing greater scope for customization and innovation than proprietary x86 and Arm solutions.

The financing was led by Eclipse Ventures, with participation from Maverick Capital, Fundomo and EPIQ Capital Group, as well as the involvement of Jim Keller, renowned chip designer with experience at Apple, Tesla and Intel. Keller, now CEO of Tenstorrent, sees AheadComputing as a potential player in CPU design for AI. According to Greg Reichow, partner at Eclipse, the AheadComputing team has the expertise to tackle the most critical challenges in advanced computing.

The company is working on 64-bit RISC-V CPUs with the goal of improving performance per core and reducing bottlenecks related to AI workloads. The design focuses on thermal density optimization and multi-core scalability, which are key elements for greater efficiency and high performance in server and data center environments. The company says its first chips will offer cutting-edge performance in both single-thread and multi-core environments, setting new benchmarks in the industry.

A recent study by Flexential found that 82 percent of companies experienced performance problems in AI workloads due to limitations of general-purpose CPUs. Although GPUs are essential for AI acceleration, CPUs remain critical for handling general-purpose operations in AI workflows, which is why AheadComputing aims to improve their capabilities.

Funds raised will be used to refine the technology and demonstrate improvements in CPU performance, as well as expand the engineering team.

With the advanced computing market set to exceed $100 billion by 2030, AheadComputing’s approach could have a significant impact on strategic sectors such as mobile, industry and networking.