In a landmark event for the semiconductor industry, Intel’s stock market capitalization was surpassed by AMD on Friday, July 29, local time in the United States. After the stock market closed on Friday, Intel’s market capitalization was $148 billion, while AMD’s was $153 billion.
As the two largest companies in the processor industry, Intel and AMD have been competing for a long time, and in the past 20 years of competition, AMD has almost always been pressured by Intel, and many times the industry has joked that the only purpose Intel has allowed AMD to exist is to ensure that Intel does not encounter antitrust litigation. However, time has changed and today we see that AMD’s stock market value has surpassed that of its old rival Intel, a scene that almost no one would have expected a few years ago.
AMD’s market capitalization overtook Intel’s after Intel released unsatisfactory second-quarter results, which were reported on Thursday local time, with the first quarterly loss in more than a decade and a sharp cut in full-year earnings estimates.
Among Intel’s multiple product divisions, in the second quarter, including PC chips, Intel’s client computing business unit revenue was $7.7 billion in revenue, a 25 percent decline. Although the PC market does have a decline in shipments due to the economic situation, but according to market research the decline was 13%, much lower than the decline in revenue of Intel’s related divisions.
On the other hand, Intel’s data center and artificial intelligence division revenue fell 16 percent, and Intel said the division’s revenue decline was mainly due to competitive market pressures. Next-generation products in the server segment (Sapphire Rapids) will also come later than expected and are now expected to come in 2023.
Intel shares fell 10% after the earnings release. AMD shares, on the other hand, rose 3%, mainly because Intel’s market share decline meant that AMD’s market share rose accordingly. At present, Intel’s P/E ratio is only 6.03, while AMD is 34.23. The difference in P/E ratio mainly reflects that investors are not optimistic about Intel’s overall prospects, while they are more optimistic about AMD’s development.
First, the problem behind the decline of Intel’s market value
Intel’s market value was overtaken seemingly suddenly, but Intel’s decline is from many years ago to sow the seeds.
Intel has not been able to maintain its dominance in the traditional PC and server markets, and the main reason Intel has not been able to maintain its dominance is that its original strong semiconductor manufacturing process is lagging behind the emerging forces represented by TSMC after many years. The position, in contrast to the use of TSMC process competitors AMD is riding the TSMC process to take advantage of the east wind. Currently, although Intel is no longer difficult in the development of new semiconductor processes, but whether it can catch up with TSMC process updates are still not optimistic. This is also the core reason why Intel is no longer leading in the traditional market.
In addition to the semiconductor process, Intel in the packaging process and the corresponding next-generation core design also failed to maintain the traditional advantages.
In AMD’s successful use of next-generation chiplet design and combined with advanced packaging technology to greatly improve yields and performance, Intel’s progress in this area is hardly smooth: the 2019 launch of the processor Lakefield using Intel’s own advanced packaging technology Foveros has been announced last year no longer developed.
The future of Intel’s further development and production progress in chiplet is also in doubt: Intel has announced that it will launch the ambitious Ponte Vecchio chip, which will contain more than 100 billion transistors and 47 chiplet production in different process nodes, but such a leap forward in the product rather than let people worry about taking too big a step. Because Intel has not previously designed and mass production of chip products using so many chiplet, so Ponte Vecchio can actually mass production and how the yield and yield will be before the official release of the product is very difficult to make people optimistic.
In addition to the core market advantage is no longer, another key reason for the decline in Intel’s market value is the failure to find new growth points. After missing the mobile computing market, Intel has been looking for other growth points, which is also an important reason for Intel’s frequent acquisitions of other companies. However, in the current emerging market, Intel can hardly be successful.
For example, in the hottest artificial intelligence market, the mainstream leader is Nvidia, Intel’s presence is not strong, although Intel has spent a lot of money in the past few years to acquire the corresponding companies (including Altera, Movidius, Nervana, Habana), but artificial intelligence has not become Intel’s new Instead, we have seen more and more of the AI server processor market being eaten up by AMD.
In the popular field of automotive electronics, Intel’s acquisition of Moileye, although the momentum in the market is good, but Intel did not bring strong synergies to the company after the acquisition of Mobileye, so the acquisition can only be described as qualified, not particularly successful at present.
Overall, we believe that the core factor of Intel’s market value decline is that its traditional core market technology advantage is no longer, while not finding new growth points. In the past few years, Intel has always given the impression that the sound of thunder is loud and the rain is small, the action is loud but specific to the product does not make people shine.
Second, AMD steadily achieves beyond
Unlike Intel, in the past few years, we have seen AMD has been a very strong focus, and not easy to spend energy to develop other popular markets, but almost all the energy on its core market (CPU and GPU) technology research and development and iteration, such deep plowing instead brought AMD a better return than the blossoming Intel.
In contrast to Intel’s problems with its own semiconductor process, AMD has chosen to hand over its model to TSMC, and this choice is in fact one of the main reasons why AMD is currently able to surpass Intel, because TSMC’s process has had a large advantage over Intel in the past few years, and even in the future evolution of semiconductor nodes, it seems that TSMC is now ahead of Intel It is also a probable event.
If it was a painful choice for AMD to give up its own semiconductor process, it seems that today this choice has finally brought AMD great rewards.
In addition to choosing the foundry model in the semiconductor process, AMD has in fact been experimenting with advanced technologies in research and development.
For example, Zeppelin chiplet technology eventually landed in AMD’s CPU and GPU technology, this technology in fact allows AMD in the CPU and GPU design field in a more advanced position relative to Intel, and because chiplet and advanced packaging will become the most critical technology in the field of high-performance processors in the future, so AMD’s strong investment in this area At present, AMD is also in a strategically advantageous position.
Unlike Intel in the field of advanced packaging, AMD in this field is steady and stable, able to define a reasonable technology roadmap and continue to iterate its technology in accordance with the roadmap plan, so AMD’s execution in technology than Intel more reassuring to investors.
If Intel has been looking for new growth in the past few years, then AMD’s new growth point is how to eat Intel’s original share in the server market. This AMD do very successful, and AMD’s success is to rely on the processor and GPU technology to steadily iterate and rely on TSMC’s leading semiconductor process, and ultimately achieve cost and performance for Intel beyond.
In the emerging field of artificial intelligence such as AMD is not without layout, but Intel wants to start a new kitchen by acquiring artificial intelligence to accelerate its chip company to overtake, AMD’s strategy is to improve its GPU in the field of artificial intelligence to accelerate performance and Nvidia head-on competition.
This strategy is the same as when AMD entered the server processor market to compete with Intel, that is, first to play the role of a backup supplier and major server customers, and through continuous iterative performance to replace the original supplier’s share.
At present, we see AMD in the launch of the MI series of GPUs for artificial intelligence and high-performance computing, its underlying programming interface and Nvidia’s GPU are compatible, although there is still a gap in performance, but because of its compatibility has been gaining more and more cooperation with customers in the field of servers.
We believe that such a compatible strategy is in fact more likely to gain a larger market share than Intel’s strategy of using AI gas pedals to overtake the market, because the current share of dedicated AI accelerator chips in the server sector is still much smaller than the share of GPUs, and as AMD’s technology slowly evolves, the chances of getting more market share from Nvidia in the future are great The chance of getting more market share from Nvidia in the future is great.
Overall, AMD’s success in recent years comes from a correct choice (TSMC foundry) and a focused and steady strategy, with strategic determination rather than trying to bend the road to overtake the run faster.
Third, the future of the competitive landscape
Intel and AMD’s traditional battlefield is the PC market, but with the PC market gradually saturated, the next most critical market will be the server market, especially in the field of artificial intelligence.
As mentioned earlier, AMD is gaining more and more share in this field – on the processor side AMD is eating into Intel’s share, while in the AI acceleration field, although AMD has not gained enough share but with the maturity of its high-performance computing GPU technology, there is at least a chance to become an alternative supplier to Nvidia in the future.
From the processor side, AMD and Intel need to further compete in the technical aspects of the implementation, Intel needs to balance its process development and chip technology development (including the balance of self-developed processes and the use of TSMC foundry), while AMD can focus on the development of chip technology.
Intel although from the momentum is not as good as AMD, but after all, the bottom is still in, through the adjustment of the future to regain a dominant position is still not a small possibility. In the server field of artificial intelligence, the two are catching up with Nvidia, the future competitive landscape will be compatible with alternative strategies (AMD’s GPU acceleration) and bending strategy (Intel’s Habana acceleration chip) and the competition between Nvidia.
Intel also has layouts in other markets compared to AMD, but there is some uncertainty about whether these layouts will pay off. This is also a test of Intel’s ability in strategy and resource scheduling, because if multiple investments are not matched with good strategic capabilities, it will be a distraction.
For example, at the end of last year Intel intended to relist Mobileye so that it could collect funds to invest in its main processor business. We hope that Intel can regain forward momentum through such adjustments, so as to bring better products to the market.
