Taiwan Semiconductor Manufacturing, popularly known as TSMC, is the world’s largest semiconductor foundry with an estimated market share of almost 62%, leading the pack by a big margin as second-place Samsung controls just 11% of this market. That’s why TSMC’s latest earnings report for the second quarter of 2024, which was released on July 18, gives us a healthy insight into the state of the semiconductor market.

TSMC’s revenue increased 33% year over year to $20.8 billion, while earnings per share jumped nearly 30% from the same period last year. The numbers were better than consensus estimates as analysts would have settled for $1.41 per share in earnings on revenue of $20.3 billion. More importantly, TSMC’s revenue forecast of $22.8 billion for the current quarter points toward stronger year-over-year growth of 32%.

The Taiwanese giant has also raised the lower end of its 2024 capital expenditure forecast. It was earlier forecasting capital expenses to land between $28 billion and $32 billion this year. It has upgraded that range to $30 billion to $32 billion, indicating that it now intends to spend an additional $1 billion in capital expenditures this year.

The above numbers bode well for Nvidia (NVDA 0.69%), which has emerged as one of TSMC’s top clients amid the artificial intelligence (AI) boom. Let’s look at the reasons why.

TSMC’s capex boost is good news for Nvidia

Nvidia was reportedly TSMC’s second-largest customer last year, accounting for 11% of the latter’s top line. Apple remains TSMC’s largest customer as the smartphone giant has been utilizing the Taiwan-based foundry giant’s manufacturing facilities to churn out the processors it deploys in iPhones, iPads, and MacBooks.

However, there is a good chance that TSMC could get more business from Nvidia this year thanks to the former’s focus on expanding its manufacturing capacity by way of an increase in its capital expenditure budget. TSMC’s management pointed out on its latest earnings conference call that “between 70% and 80% of the capital budget will be allocated for advanced process technologies.”

These advanced process technologies refer to chip nodes that are 7 nanometers (nm) or smaller in size, and these are the nodes that are used by Nvidia to manufacture its popular AI chips. For instance, Nvidia’s H100 Hopper graphics processing unit (GPU) was manufactured using TSMC’s 5nm process. The chip giant’s latest Blackwell GPUs are reportedly manufactured using TSMC’s 4nm process.

Additionally, the production line of TSMC’s 3nm chips is reportedly full until 2026 thanks to robust demand from Nvidia and others. So, it isn’t surprising to see why TSMC has decided to increase its full-year capital expenditure forecast, and it plans to spend a big chunk of its capex on advanced process nodes consumed by Nvidia.

What’s more, management indicated on the conference call that it is set to more than double its chip-on-wafer-on-substrate (CoWoS) packaging capacity this year and intends to double the same again next year. CoWoS is an advanced packaging technology used for manufacturing AI chips. Nvidia is one of the leading buyers of this tech from TSMC, reportedly accounting for half of the latter’s CoWoS capacity.

So, an improvement in CoWoS capacity bodes well for Nvidia as it should ideally be able to manufacture more AI chips and fulfill the huge demand it is witnessing. In its May earnings conference call, Nvidia management pointed out that the demand for its H200 AI GPU and the new Blackwell chips “is well ahead of supply, and we expect demand may exceed supply well into next year.”

Now that Nvidia’s foundry partner is looking to invest more money into expanding its capacity, it could deliver stronger-than-expected growth by churning out more chips.

Rising revenue and earnings estimates point toward solid growth

TSMC’s latest results are not the only indication that investments in semiconductor capacity are improving to meet the growing demand. Semiconductor manufacturing giant ASML Holding‘s recent results also indicate that foundries are set to invest more money to improve the production of advanced chips.

Given that the demand for AI chips is exceeding supply, it is not surprising to see that analysts have been increasing their growth expectations from Nvidia, especially considering that its supply line is likely to improve. This is evident from the chart below.

NVDA Revenue Estimates for Current Fiscal Year data by YCharts

The company is expected to deliver $120 billion in revenue in the current fiscal year, followed by $161.5 billion in fiscal 2026 and $189 billion in fiscal 2027. This improvement in Nvidia’s top-line growth estimates has also filtered down to its bottom line, with the company’s earnings expected to grow at an impressive pace in the current fiscal year and beyond.

NVDA EPS Estimates for Current Fiscal Year data by YCharts

Nvidia finished the previous fiscal year with $1.30 per share in earnings, and the chart above indicates that its bottom line could more than triple in three years. That’s why investors looking for a growth stock now would do well to buy Nvidia as it is trading at 45 times forward earnings, which is a small discount to the U.S. technology sector’s average of 46, especially because it could deliver stronger-than-expected growth in the coming quarters based on the indications from TSMC’s latest results.