Mitesh AgrawalChief Executive Officer

I knew the founders and leadership team well and admired their vision for designing AI inference hardware from first principles. I also saw inference becoming a much larger market than training, creating a significant opportunity.

Inference workloads are increasingly limited by memory bandwidth and capacity rather than raw compute. Solving those problems creates meaningful performance improvements for customers.

We are not trying to replace NVIDIA. Instead, we focus on workloads where our architecture delivers superior memory capacity and efficiency, allowing customers to achieve better performance for specific inference applications.

AI models continue growing dramatically in size. Larger models require significantly more memory, making memory capacity one of the biggest performance bottlenecks in modern inference.

AI infrastructure represents one of the fastest-growing technology markets in the world. Even serving a specialized segment provides enormous room for growth.

We launched our first product within two years of founding the company, generated commercial revenue quickly, and continue accelerating our product roadmap.

Trust comes from consistently delivering products. Customers gain confidence when they see predictable product releases and continuous improvements over multiple generations.

Fast product cycles keep customers confident while allowing engineers to see their ideas become real products much faster, creating momentum across the company.

Supply chain strategy matters just as much. We intentionally designed products that avoid many of the same manufacturing and infrastructure bottlenecks facing competitors.

Most investment goes into research and development, silicon design, and recruiting exceptional engineering talent. Maintaining a strong balance sheet also strengthens supply chain execution.

Building semiconductor products requires outstanding engineers. A small, highly capable team can accomplish remarkable things when everyone executes at a high level.

Execution defines our culture. We avoid unnecessary bureaucracy and focus on delivering products, solving customer problems, and continuously improving.

Everyone thinks like part of the customer-facing team. Every engineer considers how their work improves customer outcomes and strengthens the product.

We want to continue shipping increasingly capable AI inference hardware while expanding into larger portions of the AI infrastructure market through continuous innovation.