What is MemryX?

Artificial intelligence (AI) technologies continuously evolve and power various processes around us. Similarly, AI is one of the significant growth drivers in the semiconductor industry. It is because AI generates large amounts of data much faster and makes predictions more accurately than humanly possible. 

MemryX is a startup focusing on AI processing for edge devices and is creating new technologies to address this emerging market. Let’s take a closer look at MemryX and its future directions. 

Company Background

MemryX Inc was established in 2019 by Dr. Wei Lu to create an AI accelerator suitable for Edge devices. Dr. Lu has been an Electrical Engineering and Computer Science professor at the University of Michigan since 2005. He specializes in memory systems, neuromorphic computing, and in-memory computing system. Dr. Lu and his team spent almost three years developing and proving MemryX’s unique approach.

In December 2021, MemryX appointed Keith Kressin as President and CEO to commercialize MemryX’s technology. Kressin forms a dynamic leadership team in Engineering and Sales/Marketing bringing decades of executive experience from AMD, Intel, Micron, and Qualcomm. 

MemryX recently announced customer sampling of its pre-production MX3 AI Accelerator hardware and software. And Series B round may happen soon, given the needed financial resources to enter the increasingly strategic semiconductor market.

The Target and Flag Plant

Unlike most AI startups that assume to solve all things for all people, MemryX is focused on delivering an Edge AI inference technology for vision and sensing solutions. It complements a central processor for new or legacy systems rather than replacing them. Generally, MemryX is planting the flag on delivering high efficiency, low power solutions.

The Challenge

The increasing diversity of smart devices makes employing traditional computing architecture to run AI models at the Edge challenge. In addition, AI systems are data-centric, while classic processors are designed to execute instructions. 

Developing AI accelerators for edge devices are expensive due to the inconsistency of software implementation. It takes a significant engineering effort to create and update AI algorithms across a widening array of hardware solutions. On-going changes to algorithms and operating systems continuously increase the need for additional funding. 

As AI capabilities evolve, it becomes obvious why dedicated AI accelerators have limited market demand. A new approach to Edge AI that addresses this concern could add significant value to the project. 

A solution with ‘One Click Optimization.’

MemryX should address the bottlenecks and complexity of software implementation and meet the needs of Edge AI on a broader scope. 

The Neural Processing Unit (NPU) of MemryX has lower latency than other AI accelerators. It is because of its inherent ability to store AI models on the chip. The proprietary dataflow architecture could also minimize data movement within the chip since moving data requires more energy than AI computations. 

Conclusion

While some might argue the AI accelerator space is crowded, it’s apparent to me that MemryX is different. Patrick Moorhead, a Senior contributor at Forbes.com, says it uniquely solves well-known problems in Edge AI. MemryX’s focus on unique technology to meet real market needs is half the battle. MemryX is showing a pretty good start.