Artificial intelligence chip startup Etched has secured $300 million in a Series C funding round, pushing its valuation to $10.3 billion. The latest investment comes less than a month after the company emerged from stealth. Sequoia led the round, while SK Hynix, Andreessen Horowitz, Jane Street, and Diffusion joined as investors. The new valuation is more than double the company’s December valuation, when Etched completed a $500 million funding round. The rapid increase reflects growing investor confidence in specialised processors built for AI inference.
Investors Back the Next Phase of AI Computing
The latest funding highlights a broader shift across the AI hardware market. As generative AI adoption expands, inference workloads continue to grow faster than training workloads in many commercial deployments. That trend has increased interest in chips designed specifically for inference instead of general-purpose AI processing. Investors now see dedicated inference hardware as an important part of future AI infrastructure.
Speaking to TechCrunch, Etched Chief Operating Officer Robert Wachen said the latest funding values the company at $10.3 billion. The company achieved that milestone only months after its previous financing round. Strong customer demand and rapid product development helped drive the higher valuation.
Purpose-Built Hardware Targets AI Inference
Founded in 2022 by Harvard University dropouts Gavin Uberti and Chris Zhu, Etched has focused exclusively on inference computing. Last month, the company announced that it had developed a working inference chip and secured more than $1 billion in customer contracts. Those early commercial commitments suggest that enterprise customers are looking beyond traditional GPU architectures for production AI deployments.
The processor includes two custom-built components that manage different stages of AI inference. One engine handles the compute-intensive prefill stage, while the second focuses on the memory-intensive decode stage. By separating these workloads, the company aims to improve efficiency and maximise performance across large language model deployments.
Architecture Focuses on Performance and Efficiency
Etched says its processor operates at a “much lower voltage than any other AI chip,” allowing the hardware to produce less heat while supporting a higher transistor density. Lower operating temperatures can improve overall system efficiency and simplify cooling requirements inside AI clusters.
The company also says its architecture creates a shared low-latency memory pool across the entire scale-up domain. In addition, its proprietary ultra-low-latency, high-bandwidth interconnect enables “dramatically faster” memory access across multiple chips. These features target one of the biggest bottlenecks in modern AI inference systems: moving data quickly between processors.
Rack-Scale Products Move Toward Commercial Deployment
Etched manufactured its processor using TSMC’s N4P process technology. The company is currently validating complete rack-scale systems with customers before wider deployment. It plans to begin shipping its first production racks later this summer. The move reflects an important industry trend. AI hardware vendors are increasingly delivering complete infrastructure platforms instead of standalone chips. Customers now prefer integrated systems that reduce deployment complexity and improve performance across large AI clusters.
Etched has expanded its physical infrastructure alongside product development. The company employs around 400 people at its San Jose headquarters, where it operates a 2MW data centre for engineering and hardware validation. The startup has also opened a new 80,000-square-foot facility in Milpitas, north of San Jose. The site includes a 10MW data centre that provides additional capacity for product testing and customer validation. The expansion gives Etched more room to support commercial deployments as production ramps up.
Specialised Silicon Attracts Strategic Capital
Investment across the AI semiconductor industry continues to evolve. Investors are no longer focusing only on companies building general-purpose AI accelerators. Instead, they are also backing businesses that solve specific infrastructure challenges, including inference efficiency, energy consumption, and system scalability.
Etched’s latest funding reflects that change. The company combines specialised silicon, custom system architecture, expanding infrastructure, and growing customer demand. As enterprise AI moves from experimentation to production, developers with purpose-built inference platforms could play an increasingly important role in the next generation of AI infrastructure.
