The past year has been one of exciting growth for EnCharge AI and the broader AI industry. The release of generative AI tools like ChatGPT fueled demand for AI applications with the potential to enhance high-value functions across a wide range of industries. However, progress has come at a price, with AI’s energy demands dwarfing conventional computing needs and projected to grow. The implications of AI’s processing requirements are broad, with the cost, size, cooling, and power requirements of the chips putting AI’s greatest potential out of reach for many critical use cases.
As companies spent 2023 evaluating how to harness AI’s benefits, its massive cost and energy demands became painfully apparent. Simply evolving today’s conventional technologies will not sustainably meet AI’s rapidly-increasing computing needs. Current solutions aren’t suitable for critical low-latency, low-cost, or energy-efficient applications, from industrial automation, defense, and connected transportation, to practical personal computing devices and sustainable servers. For example, a single cutting-edge NVIDIA server for enterprise AI can cost more than $400,000 and require significant cooling infrastructure for the 700W TDP rating of each card, and around 10kW needed to run it all. These servers are extremely supply constrained, putting AI out of reach for massive swaths of the market whether deployed locally or through cloud providers.
AI will positively and equitably transform businesses, communities, and individuals only when it can be leveraged for the needs of individuals and organizations of any size and across a wide range of constrained use cases.
EnCharge AI Meeting the AI Moment
At EnCharge AI, we’ve developed an AI processing solution that provides revolutionary improvements in compute power, energy efficiency, and cost, democratizing AI accessibility by removing bottlenecks for the AI applications of tomorrow. EnCharge AI’s full-stack AI processing solution is the first and only commercial technology based on peer-reviewed computing breakthroughs developed at Princeton backed by DARPA funding.
Industry has long struggled with tradeoffs between traditional digital accelerators and analog computing when trying to break through the energy, cost, and processing constraints of existing compute technology. Digital technology is limited in its efficiency and requires massive data movement costs due to the restrictive ways that the fundamental CMOS chip technology is operated. It has long been understood that analog in-memory computing (IMC) can yield 100X better efficiency by dramatically improving processing efficiency and reducing data movement costs, but it could not be used to build the scalable, full-stack architectures required for AI due to the inherent noisiness of traditional analog processing techniques.
EnCharge AI’s exclusively licensed technology, built on proven, peer-reviewed innovations from the founding team across five generations of chips, captures the scalability of digital solutions and the efficiency of analog IMC. The result is a full-stack AI solution that can be built with existing, standard fabrication methods while yielding 20x better energy efficiency and 12x better cost efficiency than the best of today’s digital processor technologies, claims that have been validated on chips fabricated at the TSMC foundry using 16nm technology.
An Exciting Year Ahead
Customers across a wide range of industries are taking note of the transformational potential of our technology, motivated by the costs and constraints associated with implementing AI using current computing solutions. This customer and partner demand for us to accelerate development was the motivating factor behind our most recent $22.6 million oversubscribed funding round. With a core technology that is fully developed and proven backed by over six years of research, rapidly bringing in additional funding at this stage will help us scale our internal capacity so that we can manage the extremely exciting customer interest and product roadmaps across multiple applications from edge to cloud.
As AI development moves at breakneck speed, EnCharge AI’s compute platform is tailored for today’s AI and adaptive and scalable for tomorrow’s models and applications. Our technology has been carefully engineered to enable AI to permeate through applications, businesses, and eventually benefitting our everyday lives without exhausting our world's digital and physical resources. We look forward to continued rapid growth and sharing additional information about our customers and product availability in 2024.
Sign Up for Updates
Join our mailing list for exclusive updates, releases, and exciting news from EnCharge AI.
Thank you for subscribing!
Oops! Something went wrong while submitting the form.