Chip overlord Intel ushered in its 50th birthday last week, and Intel announced its first financial report after its 50th anniversary.
According to US GAAP, Intel's second quarter revenue for 2018 was US$16.692 billion, up 15% from last year's US$14.8 billion, and net profit was US$5.006 billion, up 78% year-on-year. Intel’s eye-catching performance exceeded Wall Street’s expectations, but the stock price fell after the earnings report. Why?
Intel and AMD are different
Specifically, Intel’s second-quarter earnings report, 51% of revenue comes from PCs, and 49% of revenue comes from data centers.
PC part:
Customer Computing Group (CCG) revenue was $8.728 billion, up 6% year-on-year
Data center part:
Data Center Group (DCG) revenue was $5.449 billion, an increase of 27% year-on-year
Internet of Things Group (IOTG) revenue was $880 million, up 22% year-on-year
Non-Variable Storage Solutions Group (NSG) revenues were $1.079 billion, up 23% year-on-year
Programmable Solutions Group (PSG) revenues were $517 million, up 18% year-on-year
For Intel's latest earnings report, Intel's interim CEO and CFO Bob Swan said: 'After 50 years of technology development, Intel's performance is expected to hit a record high for the third consecutive year. We have a unique advantage to take full advantage of processing, storage and mobile The market demand for data, and data has never been as popular or valuable as it is now. Intel is currently vying for a market opportunity of $260 billion, and second-quarter results indicate that we are winning. Because our businesses have continued to show Strong momentum, so we raised our full-year revenue and profit expectations.'
However, financial data and software company FactSet survey showed that analysts had expected the average data revenue of Intel Data Center Group (DCG) in the second quarter to reach 5.61 billion US dollars. In fact, the financial report shows that the group's revenue is 5.449 billion US dollars. DCG's failure to meet expectations is the main reason for the decline in Intel's share price.
The outside world seems to be dissatisfied with Intel DCG performance, but Intel said in its earnings report: DCG has invested in the strong demand of cloud and communication service providers to meet the explosive demand for data and the need to improve data-intensive performance such as artificial intelligence. Intel also said that DCG customers' preference for Intel's highest performance products continues, and the momentum of Intel Xeon processors will continue.
In contrast, Intel’s latest earnings report on AMD’s biggest competitor in the CPU market, according to GAAP guidelines, AMD’s second-quarter revenue was $1.76 billion, up 53% from $1.15 billion in the same period last year, up 7% from the previous month; net profit was 1.16 billion US dollars, a net loss of 42 million US dollars in the same period last year, an increase of 43%.
In terms of business, AMD's second quarter was server, and game controller manufacturers provided image processing and computing business revenue of $1.09 billion, up 64% year-on-year. Revenue contribution was mainly from Ryzen series processors and Radeon graphics cards. AMD CEO Lisa Su said in a earnings conference call on Wednesday that AMD's chip sales for the data center have grown substantially and there has been a good acceleration trend. The PC market is also better than most people expected. We are in second place. Excellent performance in the quarter, strong revenue growth, profit margin expansion, and the highest quarterly net income in seven years.
In sharp contrast to the decline in share prices after the announcement of Intel's earnings report, AMD's earnings report closed up 14.33% on Thursday, closing at $18.35, with a total market capitalization of $17.8 billion.
Intel's pressure
Short-term stock price changes don't mean anything, but Intel has indeed faced a lot of pressure recently. On June 21, Lei Feng.com (public account: Lei Feng) reported that 58-year-old Intel CEO and director Brian Krzanich will no longer hold his current position. Temporary takeover by CFO Robert Swan on the grounds that Krzanich has a close relationship with one of the company's employees, violating Intel's 'no office romance policy' policy for all management.
However, some analysts believe that Krzanich's departure may be related to Intel's underperforming data center business. At the first quarter of 2018, Bob Swan, the current Intel interim CEO, said that DCG growth is expected in the first half to the second half of this year. Mitigation, on the one hand, is based on Intel's current market position growth is more difficult, on the other hand, Intel will face more intense competition. Nomura Securities also pointed out that Intel is trying to prevent 15-20% of the data center processor share loss to AMD. In the report of Nomura Securities, Intel acknowledged in the response statement that it will face a 'more competitive environment' and said its current financial guidance has already reflected the increasingly fierce competition between AMD and other companies.
In any case, the lack of a CEO is not a good thing for Intel. But what's worse is that Intel's 10nm process has been delayed in mass production. Intel originally promised to deliver the 10nm process in 2015, but it has been delayed in the last quarter. At the teleconferencing conference, Intel announced that the mass production of 10 nanometer products will be postponed until 2019. By the second quarter of the earnings conference, Intel said that the 10nm processor will be launched in the second half of 2019. Therefore, 10nm has been repeatedly extended. The launch time of the new generation 10nm Xeon processor will also be delayed until 2020.
It should be noted that Intel's 14nm process began in 2014 and has undergone an upgraded version of 14nm+ and 14nm++. Intel has been using 14nm for almost four years now. In this regard, Krzanich previously explained that Intel's 10nm process is at 14nm. The goal of increasing transistor density by 2.7 times in the process is 'slightly radical'. In contrast, Intel's density increased by only 2.4 times when it was upgraded to 14 nanometers. Although the numerical difference looks small, the industry average is only 1.5-2 times.
As the creator of Moore's Law, Intel has long maintained its advantage in the face of competitors with this advanced semiconductor technology, so Intel is also actively solving the mass production problem of 10nm.
AMD's historic opportunity in the past decade
In the field of PC processors, AMD has always been Intel's biggest competitor, and the two competitors were only one year apart. Intel launched a series of activities to celebrate the company's 50th anniversary, and recently selected Intel's top 10 Innovative technologies and products. At this time, AMD, which was established only one year later than Intel, also listed its own series of innovative technologies, products, and called it the only company in the world with high-performance CPU and GPU.
In the history of AMD's innovation, 2017 is clearly an important year, because this year AMD launched a better architecture processor - Zen, the first Zen-based chip came out in March 2017, except for the highest frequency However, outside of Intel, the Zen architecture is generally performing well. From AMD's latest earnings report, the Ryzen series based on the Zen architecture currently accounts for 60% of its shipments.
Not only that, Su Zifeng also said that the 7nm process AMD has cooperated with TSMC and GlobalFoundries in the quarterly financial meeting after the second quarter earnings report. The 7nm Rome (second generation EPYC Opteron server processor) is currently being trial-produced. It was built by TSMC. Su Zifeng also reiterated that the Vega architecture new graphics card will be launched later this year, which is definitely the professional product line of Radeon Pro.
Previously, AMD has repeatedly emphasized that the 7nm Zen2 architecture chip has already been streamlined and is being tuned in the AMD lab. Especially on the EPYC platform, early verification with some core partners is underway and will be released as scheduled in 2019.
Therefore, even if the recent semiconductor process is different from the previous naming of semiconductor process nodes based on scientific metrics, and the more biased marketing is named, TSMC's 7nm is not exactly the same as Intel's 10nm process, so it is difficult to simply conclude TSMC's semiconductors. The process is already ahead of Intel, but TSMC is starting to mass produce 7nm products, which indicates that TSMC is successfully solving production challenges.
TSMC's 7nm process coupled with the Zen2 architecture's second-generation AMD Opteron server processor will put a lot of pressure on Intel. This may also be Intel's shift from almost 100% market share to the mission of 'not let AMD occupy 15 -20% of the market share', I think AMD is a significant cause of a serious threat to its lucrative data center business.
Intel has beaten AMD in the tick-tock tactics in the past ten years. AMD can't keep up with Intel in chip design. In the process of chip manufacturing process development, it is impossible to continue to follow Intel to improve semiconductor manufacturing process. The chip manufacturing business was sold to the United Arab Emirates investment company to establish Grofund.
However, Lei Feng.com believes that Intel's 10nm mass production problem has not been solved, and TSMC has been able to mass-produce 7nm. AMD also has a better Zen architecture. In the face of the fast-growing and lucrative big data market of 100 billion, how can Intel? Can you feel the pressure? But what's interesting is that at the end of last year, Raja Koduri, head of AMD's RTG graphics business unit, left to join Intel. This year, Jim Keller, chief architect of the Zen architecture who was resigned from AMD in 2015, was released. Keller) also joined Intel At this time, does the serious threat that Intel feels indicates that AMD has ushered in a historic opportunity to seize the Intel market in the past decade?
