Bank of America analyst Vivek Arya recently published a note predicting that the semiconductor market will grow to $3.2 trillion in 2030 from $1.7 trillion this year, an 88% increase. The analyst noted that despite some recent rhetoric about slowing artificial intelligence (AI) investments, his team has seen no slowdown in orders, capacity commitments, or pricing, all of which remain strong.
Within the semiconductor space, Arya sees memory becoming the largest market, growing 92% from $937 billion today to $1.8 trillion by 2030. Core semiconductor sales, meanwhile, are projected to rise 83% to $1.35 trillion, while server sales are forecast to jump 136% to $848 billion.
Let's look at three semiconductor stocks poised to benefit from this spending.
Image source: Getty Images.
1. SK Hynix
Bank of America sees memory as having the largest absolute dollar growth through 2030, increasing by an astounding $863 billion and accounting for over half the market's overall growth. That should greatly benefit memory makers like Micron Technology (MU +3.92%) and SK Hynix (SKHY +2.46%). Which memory maker is the better stock to own over the next several years likely comes down to the pricing of ordinary DRAM.
Demand for high bandwidth memory (HBM), which gets packaged with graphics processing units (GPUs) and other AI chips to reduce latency, is the driving force behind the memory market surge. However, this specialized memory already commands premium pricing and strong margins, while regular DRAM (dynamic random access memory) and NAND (flash) prices have seen much larger increases.
NASDAQ: SKHY
Key Data Points
If the overall DRAM market starts to come more into balance, SK Hynix is likely the better stock to own than Micron, as it is the HBM leader and a much larger percentage of its sales comes from HBM. That's why I would choose to play this expected surge in memory sales.
2. Nvidia
When it comes to compute, Nvidia (NVDA +1.34%) is one of the top stocks to own. Its GPUs remain the primary chips used to train AI models, and it continues to have a wide moat in this area, with most early AI code written on its CUDA platform.
On top of that, it has positioned itself as a leader in the inference market. Its "acquisition" of Groq allows it to use its language processing units (LPUs) for the memory-heavy decode phase of inference and its GPUs for the compute-intensive pre-fill phase.
NASDAQ: NVDA
Key Data Points
Combined with its world-class networking portfolio, this allows Nvidia to offer complete end-to-end servers for distinct AI tasks, whether that be training, inference, storage, or agentic AI. Nvidia continues to see demand outpace its capacity, and its new server systems should add another layer of growth to its story. Meanwhile, investors can pick up the stock at a bargain valuation, with it trading at a forward P/E of just 17 times.
3. ASML
In conjunction with its semiconductor forecast, Bank of America also sees wafer fabrication equipment spending soaring 129% from $156 billion to $360 billion by 2030. The biggest beneficiary of this will be ASML (ASML +3.08%).
NASDAQ: ASML
Key Data Points
The Dutch company holds a monopoly on EUV (extreme ultraviolet lithography) technology, which is needed to make both advanced logic chips, such as GPUs and central processing units (CPUs), and advanced memory, such as HBM. Its older DUV (deep ultraviolet lithography) machines, meanwhile, are used for manufacturing the less critical components of these chips. On top of that, the company's newest high-NA EUV machines, which cost double what EUV machines sell for, are set to be a future growth driver, with commitments coming from the big three foundries.
ASML plans to increase its EUV capacity by 30% next year and by a further 30% in 2028. With GPU, CPU, and HBM demand all soaring, the company should continue to see demand for its machines rise in the coming years. That makes the stock a buy.
