Artificial intelligence (AI) technology is progressing at a rapid pace, as is the application of the technology to solve real-world problems. While the market for chipsets to address deep learning training and inference workloads is still a new one, the landscape is changing quickly – in the past year, more than 60 companies of all sizes have announced some sort of deep learning chipset or intellectual property (IP) design. A new report from Tractica finds that virtually every prominent name in the technology industry has acknowledged the need for hardware acceleration of AI algorithms and the semiconductor industry has responded by offering a wide variety of solutions. Tractica forecasts that the market for deep learning chipsets will increase from $1.6 billion in 2017 to $66.3 billion by 2025. System-on-a-chip (SoC) accelerators such as those found in mobile devices will lead the market in terms of sheer volumes by the end of the forecast period, followed by application-specific integrated circuits (ASICs) and graphics processing units (GPUs).
Definition: The A-List in AI Chipset Index includes companies providing software and hardware components of AI chipsets. AI chipset products include central processing unit (CPU), graphic processing unit (GPU), neural network processor (NNP), application specific integrated circuit (ASIC), field programmable gate array (FPGA), reduced instruction set computer (RISC) processor, accelerators and more. Some of the chipsets are directed at edge processing or devices, some are for servers used in cloud computing and others are directed at machine vision and autonomous vehicle platforms. Some of the products are computational framework for AI and others are for AI training platforms. The edge and server chipsets are optimized for high performance and ultra-low power running sub one trillion operations per seconds (TOPs) to 30 TOPs.
Oh snap, Samsung is about to devour Apple when it finally releases its Galaxy S8 flagship smartphone early next year. Based on the AnTuTu benchmark, the Snapdragon 835 chipset that is arriving with the S7 successor easily trumps the A10 processor of the iPhone 7 and iPhone 7 Plus when it comes to overall performance. The AnTuTu result for the benchmark test of Qualcomm's Snapdragon 835 processor is already in, and the upcoming chipset that is being produced using the 10nm processor scored a whopping 181,434. The chipset's score is quite high, it toppled the previous record of Apple's A10 chipset at 172,644, as per PhoneArena. The scores translate to performance, so without a doubt, the Galaxy S8 could likely reign supreme in the market and snatch the crown from Apple's latest iPhones.
The collaboration portends interesting possibilities, particularly for gamers. Intel approached longtime rival AMD for the collaboration and both companies came to an agreement to design the chip, which is expected to be available by the start of next year. The collaboration is unique since AMD's graphics cards are generally used on high-end PCs and Intel makes processor for both affordable and premium PCs. Building a graphics card into the chipset is not easy. As a result, the company was able to design an advanced module which can tie both the Intel chipset and the Radeon chip together.
Qualcomm on Friday is introducing a new Cellular Vehicle-to-Everything (C-V2X) chipset and reference design that should bring automakers one step closer to deploying the communications systems needed for fully autonomous vehicles. The Qualcomm 9150 C-V2X chipset, expected to be available for commercial sampling in the second half of 2018, is based on specs from the 3rd Generation Partnership Project (3GPP), a collaboration between groups of telecommunications associations. Meanwhile, Qualcomm's C-V2X reference design will feature the 9150 C-V2X chipset, an application processor running the Intelligent Transportation Systems (ITS) V2X stack, as well as a Hardware Security Module (HSM). Qualcomm already has mulitple automotive partners endorsing the new chipset, including Ford, Audi, the PSA Group and SAIC. "We welcome Qualcomm Technologies' cellular-V2X product announcement, as the automotive industry and ecosystem work towards C-V2X implementation, and pave the path to 5G broadband and future operating services," Don Butler, executive director of Connected Vehicle and Services at Ford, said in a statement.