The AI Chip Market Continues to Heat Up
DECEMBER 18H, 2019
Over the past several years, artificial intelligence (AI) has quietly become part of nearly every tech product we use on a daily basis. The targeted ads on Facebook, the song recommendations on Spotify, product recommendations on Amazon, the directions provided by Google maps, and the video recommendations on streaming services all use the predictive power of AI.
The principal tasks of artificial intelligence (AI) are training and inferencing. The former is a data-intensive process to prepare AI models for production applications. Training an AI model ensures that it can perform its designated inferencing task—such as recognizing faces or understanding human speech—accurately and in an automated fashion.
Inferencing is big business and is set to become the biggest driver of growth in AI. McKinsey has predicted that the opportunity for AI inferencing hardware in the data center will be twice that of AI training hardware by 2025 ($9 billion to 10 billion vs. $4 billion to $5 billion today). In edge device deployments, the market for inferencing will be three times as large as for training by that same year. For the overall AI market, the market for deep-learning chipsets is predicted to increase from $6.6 billion in 2018 to $91.2 billion by 2025.
NVIDIA was the early beneficiary of the rise of AI, primarily due to its graphics processing units (GPUs). The massive parallel processing capability of GPUs to render images also turned out the be the most computationally powerful available solution for AI systems. The company's rivals have been scrambling to build a better mousetrap in an effort to seize control of the lucrative AI chipset market ever since.
Intel, having missed out on the initial AI boom, has been scrambling to play catch-up via acquisitions. Intel acquired Israeli AI chipmaker Habana Labs for $2 billion last week. Habana develops processors that are optimized for AI applications. This isn't the first such acquisition. Intel paid $16.7b for Altera in 2015 to gain control of its field-programmable gate arrays (FPGAs) for use in AI. In 2016, the company bought AI chipmaking start-up Nervana for $400 million, and in 2017 acquired MobileEye for $15.3b to gain a foothold in self-driving car processors.
Other AI chip acquisition market participants include hyperscale cloud providers Amazon, Microsoft, Google, Alibaba, and IBM; consumer cloud providers Apple, Facebook and Baidu; semiconductor manufacturers AMD, Arm, Samsung, Qualcomm, Xilinx and LG; and a staggering number of China-based burgeoning tech giants such as Huawei and Bitmain.
The intensity of the AI chip acquisition market has led to AI chip startup success stories like that of Graphcore, which was recently valued at $1.6b. Brightstone has an investment in an AI cipmaker called Vathys. Its architecture is a clean rethink of the processor, unconstrained by legacy limitations. Vathys was born and bred for deep learning. Vathys aims to achieve one petaflop per second performance on a single chip. Current deep learning processors (NVIDIA GPU) achieve 20 teraflops per second, and ‘next gen’ processors (those listed above) achieve 100-200 teraflops per second. Vathys is 50x faster than currently available GPUs and 5-10x faster than ‘next gen’ processors.
We believe the Vathys team is well suited to deliver on these performance specs, and if they do the intense competition we're seeing in the AI chip acquisition market should result in a very successful exit opportunity.
