Intel's 2017 earnings performance was robust, adding to a 23% growth this year, in addition to the original PC market, which was aggressively expanding the data center market. Experts believe that 2018 will be a turnaround time for Intel, perhaps not the most fertile year, but optimistic expectations will be at least as good as 2017 growth performance. According to the Motley fool, Intel is the world's largest GPU manufacturer, thanks largely to the high market share of the lower-order integrated graphics solutions markets. As for the high-end independent GPU (discrete GPU) market by NVIDIA and Super Micro expansion card (Add-in-board; AIB) led. Intel has recently teamed up with the Ultra Micro Radeon graphics processor in the latest x86 chipset, creating a new core and Visual Operations Division (core and Visual Computing Group), by the Ultra Micro former Radeon owner Raja Koduri as department general manager and GPU chief architect and senior VP. Intel's move will pose a threat to nvidia and hyper-micro. In addition, although Intel's Xeon processor currently takes up 99% of the ultra-high market in the data center marketplace, nvidia is catching up. The latter claims that its Tesla GPU can outperform Xeon Phi in some highly efficient computing (HPC) applications, so it is possible for corporate customers to postpone updating the CPU's history to wait for new GPU purchases, hurting the growth of Intel's data center business. In response to Nvidia's challenge, Intel plans to launch a new Xeon Phi Chip (Knights Mill) and integrates the recently acquired Altera Programmable logic gate Array (FPGA) to significantly improve the effectiveness of deep learning (deep learning). Although Intel's ability to block Nvidia's foray into the HPC market is unknown, it is expected to be a strategic focus for Intel in 2018. In addition, Intel 2017 's most bright-growing department is the internet of Things (IoT), memory and programmable chips are up to 2-digit growth rates, so it is expected that Intel will continue to focus on these high-growth businesses in the future, including the introduction of customized chipsets for wearable devices, drones and other IoT devices, and the acquisition of Automotive, IoT and computer Vision (computer vision) technology. Analysts say it is worth continuing to observe the performance of Intel's Movidius computer vision chip, which is now being used in a variety of cameras and drones, as well as an unmanned platform developed by Intel's subsidiary Mobileye. Mobileye recently announced with Fiat Chrysler and BMW work together. Another concern is the new 3D NAND and 3D xpoint memory, developed by Intel and microns, which should help Intel strengthen its non-volatile memory business. Given the increasing difficulty of following Moore's law (Moore's laws), Intel formally changed the original chip development model in 2016, from 2 to 2 stages for a period of ' tock ' pendulum strategy, extended to 30-month 3-stage ' process (s), Architecture ( Architecture), optimizes (optimization) ' loops. In the 14 nm process, Intel's previous generation Broadwell chip was released in 2014, and the current generation of the first chip Skylake was released in 2015, followed by Kaby Lake, Kaby Lake R and Coffee Lake in 2017. 2018 Intel is expected to publish a 10-nanometer skylake chip cannonlake. The comments suggest that the release of the latest generation of Cannonlake chips will help Intel pull off the gap with its competitors ' Ultra Micro (AMD). The latter launched in early 2017 to compare the efficiency of Kaby Lake but more affordable Ryzen chip.
