Abstract:
If this report is basically true, Intel may usher in a new and extremely profitable development era. It is said that almost all leading companies in the US chip technology field, including Amazon, Apple, AMD, Google, Tesla, Microsoft, Nvidia and Qualcomm, are evaluating Intel's upcoming 14A process node.
The news comes from investment bank Piper Sandler (reported by technology media WCCFTech). The report pointed out: "The performance of various data of the 14A foundry process is improving, and it is currently waiting for core customers to implement it. All major customers in the United States (such as Amazon, Apple, AMD, Google, Tesla, Microsoft, Nvidia, Qualcomm, etc.) are evaluating Intel's 14A process. There are three reasons: first, the technical data performed better than expected; second, TSMC's production capacity has been sold out by 2028, and the industry is looking for alternative foundry and packaging suppliers other than TSMC; third, the trend of localization of production capacity driven by political pressure."
"From the end of 2026 to the beginning of 2027, Intel needs to convert these intended customers into core foundry customers before production capacity ramps up in 2028. We believe that the US$20 billion stock issuance completed by Intel last month is a signal that it is confident in winning core customers."
Piper Jaffray’s report added: “This funding will be used to support equipment investment in Fab 14A, which has a planned production capacity of 20,000 wafers per month and a construction cost of approximately US$10 billion. It should be noted that the main body of Fab 62 is nearing completion, but production equipment still needs to be installed.”
In addition to the general trend of the industrial chain returning to the United States as one of the core driving forces, the US$20 billion stock issuance is also another key piece of information. As Piper Jaffray said, this funding will provide support for the construction of a new generation of 14A wafer fabs.
In addition, Intel CEO Chen Liwu has previously made it clear that a clear sign of whether Intel truly values 14A new technology is whether the company starts investing in building fabs.
Chen Liwu said in July last year: "If we are unable to win important external customers and cannot reach key customer delivery nodes for the 14A process, then it may no longer be economically beneficial to continue to develop and produce 14A and subsequent advanced process nodes in the future."
"In this case, we may suspend or terminate the research and development of 14A and subsequent nodes, as well as multiple manufacturing expansion projects."
Recently, the CEO said that the 14A process is progressing smoothly, and the first batch of so-called "risk trial production" chips is expected to be put into production in 2028. Intel believes that this timetable will allow it to directly benchmark TSMC's advanced process technology. Chen Liwu said: "This is an extremely important breakthrough."
For additional background, Intel’s 14A node is its next-generation chip production technology. Theoretically, "14A" refers to the minimum size of the internal components of the chip at this process node, which is 14 Angstroms, equivalent to 1.4 nanometers.
But in fact, the actual size of the internal components of modern chips has long been decoupled from such marketing naming. The actual size is usually three to four times the nominal node value.
Similarly, the process values of different manufacturers are not directly comparable. In other words, TSMC’s 2nm and Intel’s 2nm are not the same standard. Therefore, although you may think that Intel's existing 18A node (nominally 1.8 nm) and TSMC's N2 node (nominally 2 nm) are at a similar level, in fact most industry analysts believe that Intel's 18A performance is only equivalent to TSMC's earlier N3 (3 nm) node.
It is currently unclear the specific benchmarking situation between Intel's 14A node and TSMC technology. Chen Liwu believes that 14A can compete with TSMC's process and may be the default benchmark for TSMC's N2 node. But by 2028, TSMC’s next-generation A16 node may be the mainstream competitor on the market.
In any case, TSMC currently occupies an absolutely dominant position in the advanced chip foundry market, and Intel still has ample market space to enter. If Intel can really win even a small part of the above-mentioned potential customers, it will be a huge benefit to it.
But the problem is that as early as March 2021, the then Intel CEO announced plans to transform its customer foundry business and no longer only use its own wafer factories to produce self-designed chips. It is now the second half of 2026, and we are still waiting for Intel to announce the key core customer that will allow its foundry business to truly gain a foothold.
Comments