When it comes to foundries, most of us have the mighty Taiwan Semiconductor Company (TSM) in mind, but there is also the Israeli company Tower Semiconductor (NASDAQ: TSEM) which is about to be acquired by Intel (NASDAQ: INTC). News of the deal dates back to February, and more recently discussion between the American computer chip giant and Samsung Electronics (OTC: SSNLF) in Seoul report some form of cooperation to build production capacity to deal with the current chip supply chain crisis.
Now, as evidenced by the blue chart of Intel outperforming TSMC from mid-March (in the figure below) when it announced spending $20 billion to build two additional chip factories ( called fabs) in Arizona, investors seem to trust Intel more.
My objective with this thesis is to assess whether, at a price/earnings multiple of 7.33x, it makes sense to invest in Intel, in light of the acquisition of Tower helping the company expand rapidly. its foundry ambitions, not to mention the Samsung collaboration.
First, I paint a picture of the industry outlook amid supply chain constraints.
An industry outlook marred by supply
The United States is home to the world’s top chip researchers who excel in the design of sophisticated silicon used in computers, smartphones and communications equipment, with companies such as NVIDIA (NVDA), Intel and Qualcomm (QCOM) being at the forefront of AI, 5G and Wi-Fi. It also has some of the biggest chip equipment manufacturers like Lam Research (LRCX) whose wafer fabrication machines are used by TSMC, Samsung and others.
Additionally, there are companies like Intel, Analog (ADI) and others that manufacture chips in the US, but that only accounts for 11-12% of the total global production. With Europe accounting for another 8-9%, East Asia, primarily through countries like Taiwan, South Korea and China, makes up the remaining 80%.
Normally synonymous with long-term and sustainable demand, the industry’s outlook for chips has been marred by the current supply crisis, which actually stems from the first wave of the pandemic in the first half of 2020 and has been detrimental to the American auto industry ever since. . Additionally, the latest Chinese lockdowns put in place on March 26 this year led to component shortages for computer networking players like Cisco (NASDAQ:CSCO).
To address the problem of overreliance on Asia, the US Senate approved the CHIPS For America Act, aimed at providing $52 billion in semiconductor manufacturing subsidies by the end of this month. march. The goal of this legislation is to create a more resilient regional supply chain, which in turn should reshape the semiconductor industry itself, resulting in a chain in which Intel, with its experience operating fabs, should play a key role. However, this experience is not enough and more is needed in terms of contract manufacturing expertise or the ability to produce chips designed by fabless companies. It is at this point that acquiring the Tower becomes useful.
Tower Value Proposition
Tower, based in Israel, focuses on trailing edge nodes (150nm to 300nm) for analog devices used primarily for power management and in IoT sensors. These are used as components in a wide range of electronic devices from household appliances to cars, and related shortages prevent many sectors of the economy from being fully operational.
These trailing edge nodes are at the other end of the spectrum from the leading edge or more sophisticated process nodes such as 5nm, 7nm, and 10mm that Intel primarily focuses on. Therefore, by acquiring Tower, Intel reinforces the complementarity of its product range and benefits from the associated revenue synergies.
In addition to the technological and financial dimensions, there is also the expertise around the semi-finished ecosystem, namely how Intel’s sales teams should sell to fabless games like Qualcomm and manage customer relationships. as they evolve into a large contract manufacturer manufacturing chips for its own needs. . For that matter, it currently outsources some production to TSMC and Samsung.
Therefore, with this acquisition, Intel first expands its semiconductor portfolio and becomes a de facto manufacturer for customers like Broadcom (AVGO), Teledyne (TDY), Skyworks (SWKS) and many more. Second, he gets first-hand knowledge of how to produce chips for others, as well as how to partner with chip design companies that work with end customers like automakers. Recruiting this specialized talent in today’s tight job market is not only expensive, but can be time consuming.
Third, there is also geographic expansion with Tower’s production sites in Israel (two fabs), the United States (two fabs), and Japan (three fabs). Intel’s expansion into Japanese territory bypasses TSMC’s plans to establish a presence there since October 2021, as the Taiwanese company, which accounts for around 50% of the world’s chip supply, has become essential to solving the shortage.
The chart below shows Intel’s current and future manufacturing, testing, assembly, and R&D presence around the world, which contrasts sharply with many people’s view of it as being exclusively based in the United States.
Now, Intel has planned to spend around $100 billion in the medium term in the US and Europe, which is what TSMC is spending to expand in Taiwan or internationally. Those are big bucks and one of the factors driving foundry spending is demand, with the semiconductor market expected to grow at a CAGR of 9.2% from 2022 to 2029.
Ratings and Key Takeaways
Yet it is important to consider the difference in costs between the United States and Asia as I had elaborated in my thesis on Global Foundries (GFS) and where I had shown how government subsidies play an important role in determining the feasibility of fabs projects, with stimulus measures. being essential given the high Capex required. In this case, Intel’s investments in a German MegaFab (table above) should benefit from 5.5 billion dollars as promised by the government of this country.
Tellingly, this amount is less than the $5.4 billion Intel paid for Tower. Now, thinking out loud, as a result of the deal, Intel could also benefit from the $3.46 billion the Japanese government is providing in subsidies to set up manufacturing operations. For Intel, acquiring and expanding Tower’s assets in Japan is much faster than having to build facilities from scratch like TSMC will have to do in that country.
Telling numbers, Tower’s annual revenue of $1.58 billion (for 2021) is expected to add to Intel Foundry Services (“IFS”) revenue of approximately $900 million, making it one of of the major operating segments after “Datacenter and AI” and “Network and Edge” by the end of 2022. The company can also count on a $1 billion run rate for IFS, which is expected to generate around 3 $.5 billion in sales by 2023, representing approximately 4.4% of Intel’s total revenue for 2021.
Therefore, Intel’s futures price on the 2.35x sales multiple for 2023 is rather low. Adjusting by 15.8% which is based on its undervaluation relative to the IT sector, I get a P/S of 2.72x. This translates to a target of $51.9 (44.84 x 2.72/2.35) based on a stock price of $44.84. The US company is also undervalued considering the price-earnings metric, as shown in the chart below.
As for Tower, given the offer price of $53 per share, its stock is still about $4-5 undervalued from its price of $48.4. In that regard, regarding regulation, given Intel’s previous presence in Israel, the deal was met with enthusiasm from Tower shareholders and the company had completed the regulatory process in two jurisdictions by April 28.
With this in mind, the acquisition not only advances the Intel IDM 2.0 strategy which emphasizes manufacturing capability, but also strengthens the company’s R&D with the Tower Design Center and Platform. photonic form of open foundry silicon. In the same vein, a collaboration with Samsung, possibly for expertise around state-of-the-art foundry outsourcing services, is a positive point. It might sound like a far-fetched idea, but it’s doable as with its foundry in Texas, where it makes logic (processor) chips for Intel, Nvidia, Tesla (TSLA) and others, Samsung just like Tower has experience of contract manufacturing on American soil.
The South Korean company is also expanding production in Austin with a $17 billion investment and given the tons of materials needed ranging from iron and steel castings to chip-making equipment, these companies could save money. of scale in the United States, including making some joint orders as inflation bites harder.
Finally, this collaboration which was discussed in person with the CEO of Intel traveling to Seoul to meet his Samsung counterpart is not to be taken lightly because it has a strong geopolitical flavor, and comes just after the visit of the US President in South Korea.