Competition in AI platform market intensifying

Major chip players, including Intel, Nvidia, Qualcomm and AMD, have all stepped up their efforts to develop related artificial intelligence (AI) platforms, making the AI market a fierce battlefield in 2017.

Intel, which already quit the smartphone market, has begun re-organizing its businesses and integrated all its AI-related businesses under the newly established AI Products Group and AI laboratories. The company has also been acquiring companies and forming strategic alliances with partners to strengthen its AI platforms for applications including drone, autonomous driving, medical care and smart factory.

Intel has recently launched its Nervana platform, designed specifically for AI application and the solutions that are based on the platform use products including Intel’s Xeon and Xeon Phi processors and adopt technology such as FPGA and from Intel’s acquired Nervana

In the first half of 2017 Intel will test its new Lake Crest chip that has been optimized for neural network. Intel also has a chip codenamed Knights Crest, which combines Intel’s Xeon processor with Nervana’s technology.

Intel’s Saffron platform will provide transparent analysis on heterogeneous material using a memory-based reasoning method, while the company has also been offering smart APIs, expanding its distributed computing processor series, releasing RealSense camera and Movidius VPU to assist the development of related technologies.

Intel has recently acquired Mobileye for US$15.3 billion to push its effort in the AI market, while its autonomous driving car fleet is expected to hit the road some time in 2017. Intel also acquired a 15% stake in map content developer HERE to ensure the autonomous driving cars have the latest map information.

Unlike other chipmakers, Nvidia is using its GPU technology to push AI applications. The company launched Tesla P100 chip and Tesla P4P40 chip for image, text and sound recognition. Nvidia has also cooperated with electric car vendor Tesla and developed its car-use AI platform Xavier. Nvidia also has partnership with Audi to release autonomous driving cars.

For the home-use Internet of Things (IoT) area, Nvidia has Shield+Spot platform, while its Jetson TX2 can be used in end devices as well as smart factory robots, enterprise drones and smart cameras in cities.

Nvidia has recently partnered with Microsoft to develop an AI cloud computing platform accelerated by GPUs. Nvidia’s Telsa has also been adopted by China-based Tencent’s cloud computing platform. Nvidia is planning to host GPU Technology Conference (GTC) 2017 in May and is expected to unveil more updates on its AI plans.

Qualcomm has been aggressively trying to expand in the AI market. Following its US$47 billion acquisition of the largest car-use chip maker NXP worldwide, Nvidia also recently formed collaboration with map content provider TomTom.

AMD CEO Lisa Su recently noted that the company will not be absent from the AI competition and the market is rather important to AMD. In addition to processor and GPU technologies, FPGA and ASIC are also expected to become major technologies. AMD has recently launched Radeon Instinct graphics card, specifically for AI and machine learning purposes.

AMD is the only company that has both GPU and x86 processor technologies. In addition to capability of satisfying datacenter’s variety of demand, the company is also able to push smart machine’s development as the company has been developing the efficient connection between the GPU and the CPU to meet AI’s technological demand.

Categories: IT + CE IT components, peripherals Server, IPC, cloud computing Software

Tags: AMD Intel Nvidia Qualcomm

Related stories
• Digitimes Research: DNN-architecture AI widely used in image recognition (Mar 22)
• AI helps Nvidia reach record revenues (Feb 10)
• Brand vendors to launch smartphones with AI applications to ramp up sales in 2017 (Jan 24)
• SerComm to offer AI-integrated devices in mid-2017 (Jan 20)