A quick look around shows how much of the world is going digital. Over the past decade or two, cellphones, electric vehicles, smart home devices, and many other goods have been transformed by technology.
Semiconductors have been undervalued but crucial to the digital revolution. Modern electronics require semiconductors. Unfortunately, COVID-19 and its supply chain consequences have generated a semiconductor shortage in recent years.
However, one company provides investors a compelling long-term prospect. The world's largest semiconductor foundry, Taiwan Semiconductor Manufacturing (NYSE: TSM), or TSMC, makes individualized sophisticated chips for top firms.
A major tech development was TSMC's foundry model. Instead of mass-producing semiconductors, TSMC makes unique ones. Companies can't buy TSMC's semiconductors at a retailer, but they can contract the company and say, "We need a semiconductor for this purpose," and TSMC usually delivers.
TSMC works on iPhones, Tesla on electric vehicles, Nvidia on GPUs, and dozens of more companies use them for everyday technologies. Since TSMC perfected the foundry model, other semiconductor companies like Intel have followed suit, but not as well.
Looking at TSMC's revenue from the past two years shows that the semiconductor industry is cyclical. TSMC declares results in NT and USD because it is situated in Taiwan. Revenue was steady in NT and down 1.5% in U.S. dollars in Q4 2023. An overall reduction in smartphone sales represented for 43% of TSMC's last quarter revenue and 38% of its 2023 revenue, limiting revenue growth. The problem is short-term, therefore investors shouldn't worry too much.
In 2024, the International Data Corporation mobile phone tracker predicts 2.8% smartphone sales growth from the negative. TSMC should benefit from an expected 1.2 billion smartphone shipments in 2024.
Those who have followed tech in the past year have seen the AI frenzy. TSMC underpins the AI ecosystem, even though it doesn't directly deal with AI. Data is the foundation of AI.
Big data is needed to train and power AI apps like ChatGPT, Bard, and others, but external hard drives and cloud storage aren't enough. Large data centers with superior processing capabilities store and process it. This is why Nvidia's demand has risen recently. They make GPUs for data centers. Guess who makes GPU chips? As expected, TSMC. The AI pipeline fails without TSMC chips. Though not the only manufacturer, they make the most modern and efficient.
In its previous results call, TSMC's CEO predicted a 50% compounded annual growth rate for AI-related semiconductors, which account for a minor portion of its revenue. He predicts AI-related semiconductor revenue in the high teens by 2027.
stay turned for development