Faced with the 10nm dilemma, Intel launched this combination of punches

Source: Content from "Digital Times", thank you.

At the Data-centric Innovation Summit in the United States in early August, Intel once again responded optimistically to the difficulty of 10nm technology, stabilizing investors' confidence. "In the AI ​​era, customer needs are diverse, and it is impossible to meet customer needs with CPU alone. We will provide the "best" product portfolio and sustainable solutions, just like we have done as a leader in the past 20 years. I am full of confidence." Navin Shenoy, Intel's executive vice president and general manager of the Data Center Group, said.

Chip giant Intel's 10nm process CPU will not be launched until the second half of 2019 at the earliest. This will make AMD, a competitor that has already launched a 7nm CPU, ahead of Intel in process technology for the first time. This major event has caused the outside world to worry about whether Intel's market leadership will be lost.

Goldman Sachs analyst Toshiya Hari pointed out that Intel's delay in launching 10nm process products is more serious than most people expected, affecting Intel's market share and leading to a decline in revenue, gross profit margin and operating profit margin. Nomura Securities directly pointed out that AMD will eat up 15% to 20% of Intel's data center processor market share. Nomura Securities also estimated that Intel may lose $2.62 billion to $3.49 billion in 2018 (based on an estimate of 15% to 20%). Since AMD's total revenue for the whole year of 2017 was only $5.33 billion, AMD's revenue and gross profit margin will be greatly increased.

However, judging from its stock price, revenue growth and the latest strategy announcement at the Data-centric Innovation Summit, Intel's future is not as bad as the outside world claims.

As Navin Shenoy said, Intel offers the broadest product portfolio. Although Intel's market share in desktop and laptop CPUs may be affected, Intel is gradually reducing its dependence on PC chips and turning to an operating model based on diversified businesses such as artificial intelligence, data centers, autonomous driving, 5G, FPGA and the Internet of Things.

In the first quarter of this year, the revenue of Intel's PC-related Client Computing Group (CCG) department has dropped to less than half of the total revenue.

"There is absolutely no one-size-fits-all solution," Navin Shenoy stressed. As the types of customers increase significantly, in addition to the advancement of chip manufacturing processes, integration between chips, software, systems and open source frameworks is also needed.

Therefore, in addition to the shipment of 10nm chips in the next one to two years, integrated warfare is also crucial for Intel.

Navin Shenoy believes that Intel not only needs to upgrade the CPU chip itself, but also needs to strengthen its integration capabilities with memory chips, semi-customized chips and software.

From Navin Shenoy's tone, we can tell that Intel believes that the outside world is too focused on the progress of the 10nm process and has ignored Intel's progress and efforts in other areas.

For example, Intel will work to integrate the latest memory technology Optane DC Persistent Memory into the CPU. In the field of FPGA chips, Intel relies on continuous mergers and acquisitions to enhance its technology research and development capabilities. The latest merger and acquisition case is the acquisition of eASIC in mid-July, allowing Intel to provide chip technology between FPGA and ASIC.

In terms of software, Intel has opened up its open-source Deep Neural Network (DNN) compiler nGraph, allowing various DNN models to run on various Intel chips.

In terms of deep learning chips, Intel uses the Nervana NNP-L1000 processor to compete with NVIDIA products.

The above examples are only a small part of Intel's integration war. The blueprint provided by Navin Shenoy shows a much larger and more ambitious architecture.

However, can the revenue generated by the integration of these chips make up for the revenue from the difficult production of 10nm chips? And will an overly broad layout lose focus? These are the thorny issues that Intel must face in the future.

Today is the 1679th issue of content shared by "Semiconductor Industry Observer" for you, welcome to follow.

Moore Elite is a leading chip design accelerator that reconstructs semiconductor infrastructure to make it easier for China to make chips. Its main businesses include "chip design services, wafer packaging and testing services, talent services, and incubation services". It covers more than 1,500 chip design companies and 500,000 engineers in the semiconductor industry chain, and has precise big data on integrated circuits. It currently has 200 employees and is growing rapidly. It has branches and employees in Shanghai, Silicon Valley, Nanjing, Beijing, Shenzhen, Xi'an, Chengdu, Hefei, Guangzhou and other places.

Click to read the original article to learn more about Moore's elite