Why is Morgan Stanley saying bluntly that investing in TSMC is tantamount to taking "dead money" in the semiconductor industry?

As the demand in the semiconductor industry continues to be tight, TSMC, Samsung, Intel and other wafer factories continue to increase their investment and expand production, which also makes the capital market highly excited. Investment banks have given positive ratings and rising target prices. At a time when the market is highly expected that the semiconductor industry will become a "tuyere" that will last for several years, a research report recently released by, Morgan Stanley poured a rare cold water on it.

According to an updated rating by Zhan Jiahong, an equity analyst at Morgan Stanley, TSMC was downgraded to neutral on the 18th, while the target price was also lowered from T $655 to NT $580, a full 23% downside from the market consensus of NT $749. In a nutshell, the, Morgan Stanley team believes that although TSMC is in the tuyere, it may not be able to maintain a high level of profitability under the pressure of advanced processes.

Very communicative is that the, Morgan Stanley team in this report also referred to the investment in TSMC in the next 12-18 months as "dead money" (dead money).

There are still expectations in the process, we should pay more attention to the bottom line of cost.

The Morgan Stanley team said that simply looking at the progress made by IBM in breaking through 2nm and TSMC in the 1nm process, 2021 will still be a key year for Moore's Law. But for TSMC, a problem is emerging: the bottleneck of profitability comes faster than the process.

According to estimates, after reaching the 5nm process milestone, as the company's bargaining power with key equipment suppliers such as Asma weakens gradually, capital intensity will rise significantly. According to the preliminary quotation of TSMC 3nm wafers, the cost of transistors can not be further reduced. In addition, a series of alternative technologies, such as advanced 3D packaging, can also improve chip performance and divert to the needs of 3nm/2nm processes.

The Morgan Stanley team said that although the market generally uses the 53 per cent gross margin in 2020 to predict the follow-up trend, TSMC's gross margin in 2022 and 2023 is likely to fall below 50 per cent, less than the 52 per cent widely expected. This is still based on the judgment of the semiconductor upward cycle, and if it enters the industry downward cycle in 2022 as expected by Morgan Stanley, it will further prompt investors to adjust their expectations for TSMC's long-term profit growth.

The valuation is close to the ceiling.

Judging from Pamp E of a series of semiconductor industry companies, as a contract manufacturer that enables advanced technology to be "applied", TSMC's compound annual growth rate is also very difficult to break through the 20% threshold, which also means that the company will remain in the range of 20-25 times Pamp E, and it will be difficult to enter the valuation range of "technology innovators" such as Asma and Infineon.

The Morgan Stanley team said the latest T $580 forecast hinted at a forecast of 19x EPS in 2023, which already included the contribution of Intel 3nm CPU outsourcing orders, but also needed to be slightly reduced to reflect the decline in profit margins. The team currently forecasts an annualized compound growth rate of 15% for EPS from 2020 to 2023, which makes the 12-month forward Pmax E to 26 times looks a bit high. The subsequent rise in bond yields and changes in the industry cycle have made the decline in Phammer E a more likely direction.

Zhan Jiahong's team also mentioned potential "alternative investment" options in the report. Currently, the global semiconductor team gives buy ratings to equipment makers Asma, Pan Lin Group and Tokyo Electronics. For the Asia-Pacific region, Zhan Jiahong's team continues to be bullish on Samsung and believes that MediaTek and UMC will be key alternative investments.