Shulou(Shulou.com)11/24 Report--

CTOnews.com, February 6 (Xinhua)-- senior Forbes writer Patrick Mulhead (Patrick Moorhead) released a column called "Intel's Future Is Prettier Than Its Past Quarter And Guide", arguing that it is inappropriate for Intel to measure a semiconductor transformation that usually spans 4-5 years with a quarterly earnings report.

CTOnews.com is accompanied by an introduction by Patrick Mulhead (Patrick Moorhead): he specializes in covering the technology industry for more than 30 years, writing about disruptive companies, technologies and usage models.

Before officially opening this article, CTOnews.com summarized the headlines of the domestic and foreign technology media on Intel's fourth quarter 2022 results:

Reuters

China Business Daily

CTOnews.com, an American stock research agency, attached the main data of Intel's financial report:

Fourth quarter of 2022:

Intel's total revenue was $14.042 billion, down 31.6% from a year earlier and $20.528 billion in the same period last year.

Intel posted a net loss of $661 million, down 114.3% from a year earlier, and a net profit of $4.623 billion in the same period last year.

Adjusted net profit was $400 million, down 92% from a year earlier and $4.7 billion in the same period last year

Net revenue from Intel's most profitable customer computing business was $6.625 billion, down 36% from a year earlier, compared with $10.333 billion a year earlier.

Revenue from Intel's data center and artificial intelligence (AI) business was $4.304 billion, down 33% from a year earlier, compared with $6.426 billion a year earlier.

The whole year 2022

Intel's full-year revenue was $63.1 billion, down 20% from a year earlier. $79 billion in 2021

Intel's full-year profit was $8.014 billion, down 60% from a year earlier. $19.87 billion in 2021

CTOnews.com noted that at the close of trading on the second day after Intel announced its results (January 27, 2023), Intel shares fell 6.4%, wiping out about $8 billion in market value. In the face of such ugly results, a number of Wall Street analysts have given a "disaster level" assessment.

Touyuan: DoNews and to make matters worse, Intel CEO Pat Kissinger (Pat Gelsinger) said that the first quarter of 2023 results will once again fall short of expectations.

Intel's revenue estimates for the first quarter of 2023 are between $10.5 billion and $11.5 billion, well below market expectations of $14 billion.

Moorhead wrote in this commentary that Intel's latest quarterly results are not beautiful enough, and many Wall Street institutions have focused on analyzing the reasons for this quarter's "catastrophic" results, but there is little mention of Intel's long-term future.

Moorhead believes that this is because Wall Street is dominated by quarterly reports, and most news articles have an even shorter time span. The spread of this news to spectators who are not familiar with the semiconductor industry will only magnify the huge challenges facing Intel and ignore the future development of Intel.

Moorhead, who specializes in covering the semiconductor industry for more than 30 years, believes that Intel's semiconductor transformation will take at least 4-5 years, which is usually longer.

The main point of the attached article of CTOnews.com is that the semiconductor transition cycle is long, and the main argument is that

Products and markets: proof that it takes time to enter new markets and succeed

Architecture: it takes time to completely change the way the product is architected and make it competitive

Manufacturing: proof that it takes time to become a leading node and competitive universal contract factory

Semiconductor transition cycle is long Moorhead pointed out in the article that Intel's transformation cycle is limited by the development of silicon technology, which is doomed that the transformation can not be achieved overnight, Intel will need at least 3-5 years to turn a profit.

Another important factor is that humans are not good at discovery and perceptual transformation.

Both Apple and AMD have experienced such dark moments when they gambled to turn things around against a backdrop of widespread gloom.

Before AMD released the Ryzen processor, few people saw the potential of AMD in many crises, and Moorhead believes Intel is in the same moment.

Moorhead observes the transformation of semiconductor enterprises through three aspects.

Products and markets: the products developed and the markets involved

Architecture: a basic approach to designing IP blocks and encapsulation

Manufacturing: the ability to mass-produce chips

Intel is currently working on these three areas at the same time, and it will naturally take time to achieve transformation in all three areas.

Products and markets: proof that it takes time to enter new markets and succeed Intel's biggest change in terms of products and markets is the expansion of its total availability market (TAM), expanding consumer standalone GPU, data center GPU, data center HPC SoC, consumer AI chips, and data center AI cards.

Intel currently has a very low share of these products and markets, while NVIDIA (NVDA) dominates at $471 billion, and Intel's competitor AMD also has a presence in these markets.

If Intel can succeed in these markets, it will be a big help to boost the overall size of its business. Intel is a new player in these markets, but it is not easy to shake up strong competitors, and it will take time.

CTOnews.com small class:

The total addressable market (TAM), also known as the total availability market, is a term commonly used to refer to revenue opportunities available to a product or service. TAM helps prioritize opportunities by acting as a quick indicator of the potential potential of a given opportunity. One way is to estimate how much market any company can get without competitors. A broader change is to estimate the size of the market that can theoretically be provided with a particular product or service.

Architecture: it takes time to completely change the way products are architected and make them competitive to prove that Intel is pushing ahead with architectural reforms involving IP blocks (block) such as computing, GPU, AI, I / O chipsets, and improving and innovating the packaging of these chipsets.

All Intel's future logical designs will no longer use monolithic architectural design (Monolithic Architecture Design), in which everything on a bare chip (die) is made by a process or by a fab, but will use distributed design (Distributed Designs), using 2D, 2.5D and 3D packages to assemble various chipsets.

These chipsets are not entirely produced by Intel. For example, some of Intel's GPU chipsets can be produced by TSMC. Moorhead believes that Intel's CPU chipset will not be handed over to TSMC at this stage, but he believes that if Intel does not have sufficient technological advantages, it may be handed over to TSMC.

Moorhead firmly believes that Intel is an industry leader in 3D packaging technology, which is one of the "killer's mace" for Intel to turn around in the future.

Sapphire Rapids is one of Intel's first products with high-capacity, fully distributed architecture. It is because of the modular design that Intel has repeatedly delayed its launch.

Intel's first fully modular PC product, called Meteor Lake, has confirmed that production will begin in the second half of 2023.

Moorhead approached several original equipment manufacturers (OEM) who were excited about the product and thought it was strong enough to compete with Apple. CTOnews.com netizens, are you looking forward to it?

Moorhead believes that it is not easy for Intel to really improve distributed design and needs to revolutionize its design, testing and packaging methods and technologies to ensure that its IP is competitive, shared and cost-effective.

Intel has been using monolithic architecture for the past 30 years, and it will take time for Intel to shift to modularization in all designs.

All of Intel's designs are distributed, and chipsets are mass-produced through competitive processes. It should be interesting to see how Intel performs.

Manufacturing: proof that it takes time to become a leading node and competitive universal foundry Moorhead believes that the biggest strategic change since Pat Kissinger (Pat Gelsinger) took the helm of Intel has been to establish wafer foundry business IFS (Intel Foundry Services) and compete with TSMC, Samsung and GlobalFoundries.

It obviously takes time to prove that IFS wants an industry-leading node advantage. Intel announced in February last year that it would buy Israeli chip company Tower Semiconductor Ltd. (Gaota Semiconductor) for $53 a share in cash, totaling $5.4 billion. It will also take some time for Intel to digest and absorb high-tower semiconductors.

First of all, it is commendable that the contract manufacturing business can bring huge returns.

TSMC, for example, had a net profit margin of 47.3% in the latest quarter. If Intel's IFS business can form a scale, it can occupy a competitive leading position in PPA (performance, power, area) in cutting-edge design.

Chief executive Pat Kissinger (Pat Gelsinger) announced that it was fulfilling its promise to deliver five nodes within four years, or "ahead of schedule".

Intel disclosed that:

Intel 4 has been "manufacturing ready" (ready to go into production) and Intel 3 has "on track" (on track). Both nodes use EUV, which should significantly reduce costs by eliminating double-pattering.

18A / 20A: provide "silicon is running in fab" for major potential OEM customers. These two nodes deploy technologies with major breakthroughs such as RibbonFET and PowerVia. The complete PDK should be shared with key customers a week before its launch.

From the customer's point of view, Intel is also doing well:

MediaTek buys $4 billion worth of products during the Intel 3 process life cycle

Intel has added a "leading provider of cloud, edge and data center solutions" to customers

A total of 43 Intel chips are being tested.

Has established cooperation with 70% of the largest contract manufacturing customers.

Moorhead believes that it will take several years for Intel's contract manufacturing business to develop, and Intel is now entering a "short-term pain".

Summary

Moorhead believes that Intel's semiconductor transformation will take time, and it is unreasonable for Intel to use a quarter to measure a transition spanning 4-5 years.

Intel is expected to deliver 5 nodes within 4 years, which is undoubtedly a "shot in the arm" for Intel's future development. He believes that in addition to considering the huge risks currently facing Intel, investors should pay more attention to the long-term effects of "5 in 4".

Moorhead admits that Intel's earnings figures for the fourth quarter of 2022 and the first quarter of 2023 will be ugly, and he won't whitewash them.

Moorhead just wants investors, media and viewers to have a more objective perspective and a deeper understanding of Intel's transformation efforts. Moorhead believes that according to the five-year development cycle, the relevant transformation results of Intel will not be apparent until the fourth quarter of 2025.

Related readings:

1. "Intel's Future Is Prettier Than Its Past Quarter And Guide"

2. Intel Reports Sharp Loss for 2022

3. "Intel Reports Fourth-Quarter and Full-Year 2022 Financial Results"

4. "Intel shares drop on weak quarterly results and expectations for another quarter of losses"

5. Intel, it's time to wake up.

6. "Intel's share price has not reached its final low."

7. "Intel releases fourth-quarter results analyst:" Historic collapse "

8. Intel is caught in another "cold winter" when it is reported that it is a record loss.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.