With artificial intelligence (AI) chipmaker Nvidia hovering around a mind-blowing $3 trillion market capitalization, while Apple generates massive publicity as it finally jumps into the generative AI game, it’s no wonder that an ambitious startup would highlight any AI aspect of its product.
The startup 5G chip designer EdgeQ has already raised a total of $126 million over 2 rounds of venture funding since November 2020 by delivering a 5G chip “with AI.” The basestation-on-a-chip has been in production since last year and has already been adopted by customers, Edge Q said.
Indeed, principal analyst at J.Gold Associates, Jack Gold, noted that ActionTec — a longtime Bluetooth, Wi-Fi and modem maker — is a customer.
“[EdgeQ’s] target market is really the compact base station,” Gold told us in an email. “They want to enable companies to build out small cells, which is a significant market and growing.”
“Qualcomm is a competitor in the space and a bigger player for sure, although they [EdgeQ] probably also have an edge when it comes to price. Qualcomm is not known for bargain pricing.” Gold noted. Some of the top small vendors in 2024 include Nokia, Qualcomm, Huawei, Ericsson and Cisco, according to a Mordor Intelligence report.
As EdgeQ is led by former Qualcomm, Intel and Broadcom employees, the startup will be aware of where its small cell rivals stand. It should be noted that Qualcomm started looking at the role of AI in 5G in September 2021.
AI in 5G?
“[Part] of EdgeQ's uniqueness is realizing that the mathematic operations between baseband processing (ie. 5G) and AI are rather similiar,” EdgeQ head of marketing Edward Wu told Fierce in an email.
“So we have constructed our Tensor Execution Unit (TXU) cores to flexibly process both 5G and AI in tandem," said Wu. "The cores are super elastic and customizable, giving customers the ability to portionate how much 5G or how much AI inference they would like to split across the cores."
Inference is the process of running live data through a trained AI model to make a prediction or solve a task, IBM Research noted.
“5G’s connectivity will be enhanced by on-device AI training so that factors like network traffic, beam pattern optimization, radio frequency spectrum monitoring, and resource allocation are managed better,” Adil Kidwai, VP and head of product management at EdgeQ said in a 2023 company blog. “AI algorithms can analyze vast amounts of data generated by the network and provide insights into network usage patterns, traffic congestion and performance issues.”
Analyst Gold is a little more circumspect about the AI aspect of the basestation-on-a-chip. “If they are using the AI to manage the cell, the needs are fairly modest and doesn’t require high end processing. That’s not to say it’s not important, just not a big technology stretch,” he said.
Nonetheless, no doubt 5G plus AI inference will be commonplace soon enough.