EnCharge is betting on analog chips for the next-gen AI | Computer hardware parts and functions | 3 components of computer hardware | Computer hardware | Turtles AI

EnCharge AI, a startup born out of research at Princeton University, has closed a Series B funding round raising more than $100 million. The goal is to accelerate the deployment of its innovative analog chips, designed to reduce AI power consumption by up to 20 times compared to traditional solutions. The company aims to move AI inference workloads from data centers to on-premises devices, improving security, latency and cost efficiency. The round, led by Tiger Global, strengthens EnCharge’s position in a strategic market, aligning with government initiatives for hardware innovation.

Key Points:

• Strategic Funding: Raised over $100 million to bring analog AI accelerators to market.
• Energy Efficiency: Chips up to 20 times more efficient than current solutions.
• Institutional Support: Collaborations with Princeton University, DARPA, and key industry investors.
• Technology Innovation: Noise-Resistant In-Memory Architecture for Advanced AI Processing.

EnCharge AI stands out in the semiconductor landscape by adopting an innovative approach based on analog chips, which overcome the traditional model of separate processing between memory and compute. The company aims to improve the AI ​​sector with a technology capable of processing data directly in memory, drastically reducing energy consumption and optimizing device performance. This methodology responds to one of the main challenges of the sector: the high computational cost of AI, which today weighs on data centers powered by energy-intensive GPUs.

The financing, led by Tiger Global and supported by prominent investors such as Samsung Ventures, HH-CTBC (a joint venture between Foxconn and CTBC VC), Maverick Silicon and SIP Global Partners, brings the total capital raised by EnCharge to over $144 million. The startup, based in Santa Clara, plans to launch its first commercial solutions by the end of the year, consolidating its role in an increasingly strategic sector for the global economy.

Unlike conventional chips used for AI training and inference in data centers, EnCharge’s technology is designed to run AI models directly on edge devices, such as laptops, smartphones, and wearables. This approach reduces cloud dependency, improves security and latency, and offers a more sustainable and cost-effective alternative to current GPU and TPU-based solutions.

The company is full-stack, having developed not only the hardware but also an entire software platform to optimize the integration and use of its AI accelerators. The key to EnCharge’s technological success lies in its noise-resistant in-memory computing architecture, which sets it apart from its competitors. CEO and co-founder Naveen Verma highlighted how the great innovation lies in the ability to leverage a highly controllable geometric device of metal wires, already present in the standard supply chain, to achieve more efficient processing.

The involvement of strategic partners such as TSMC, which produces EnCharge’s chips, demonstrates the strength of its technology roadmap. The startup has been working for years before emerging from the shadows, preferring to refine its technology before accessing venture capital funding. In addition to the support of Princeton and DARPA, EnCharge has an advisory board with experts such as Donald Rosenberg (formerly of Qualcomm), Sam Heidari (CEO of Lumotive) and Andrea Goldsmith (Intel and Princeton).

The company enters a highly competitive market, challenging giants such as Nvidia and AMD and emerging startups such as Mythic and Sagence. However, the analog in-memory model could represent a crucial turning point, especially given the growing interest of the United States government in strengthening domestic semiconductor production.

EnCharge has already shown strong potential, with major investors believing in its vision of a more efficient, scalable and sustainable AI.