Microsoft spent $1 billion on OpenAI in 2019 and has invested another $10 billion in the company, indicating that it sees great potential in this new, powerful AI engine. The software giant has introduced an OpenAI service on Azure, which developers may utilize in their software designs, and is trying to integrate ChatGPT’s capabilities into its software products, from Office to Bing.
Every significant tech company will now invest extensively in AI to compete as AI appears to have reached a level at which it becomes table stakes in a wide range of enterprise and consumer applications.
If the AI wars start, the semiconductor companies that make AI work should profit more than Microsoft or other giant tech platforms. Here are four stocks poised to benefit.
Nvidia
Nvidia (NASDAQ: NVDA) dominates GPUs, which are needed for artificial intelligence. It’s intriguing that ChatGPT doesn’t even run on Nvidia’s latest chip, the H100, or “Hopper,” introduced late last year.
The October-shipping H100 is expected to do AI “training” nine times faster than the A100 and “inference” 30 times faster. The H100 claims 3.5 times improved energy efficiency and three times lower total cost of ownership.
Grace, Nvidia’s first CPU, is designed to work with Hopper and Nvidia’s networking-focused DPUs. These comprehensive AI systems should offer ultra-fast GPU processing, networking, and data mobility.
In 2023, Nvidia will benefit from Hopper H100 chips and AI warfare. After its 50% drop in 2022, the stock has traded at a more reasonable valuation.
TSM Semiconductor
ASML Holdings and Taiwan Semiconductor Manufacturing Company (NYSE: TSM) make Nvidia GPUs.
ASML monopolizes the extreme ultraviolet lithography (EUV) technology used to make cutting-edge semiconductors. EUV lasers make semiconductors with transistors smaller than 10 nm apart by top chip firms. The AI wars will force chip foundries to produce more cutting-edge chips since artificial intelligence applications need the most transistor-dense, power-efficient devices.
Creating a cutting-edge semiconductor requires more than lithography. Lithography, masking, etching, and deposition machines; metrology and inspection machines; and improved packaging are needed to make these complicated chips without flaws.
Taiwan Semiconductor’s experience as the world’s largest outsourced foundry gives it a competitive moat. TSMC’s size and scale advantage, as well as Intel’s current struggles due to the sluggish PC market, look to be increasing its lead in leading-edge chip production.
On a results call with analysts, management cited its high-performance computing segment for AI customers as the reason for its optimism. TSMC foresees a semiconductor market rebound in the second half of 2023.
ASML and TSMC cooperate, although both are competitors in generating the most advanced AI processing chips.
Micron Technology
Finally, all that data processing demands lots of memory and storage, which should boost Micron Technology (NASDAQ: MU).
As PC sales plummet and cell phones and consumer electronics struggle, the memory market’s delicate supply-demand balance has collapsed, causing Micron’s profitability to decrease.
Despite worsening performance, Micron’s stock has strangely remained around June levels. The market is forward-looking, and the stock trades just above Micron’s book value. The chipmaker is one of three global firms that mass-produce DRAM memory and one of five that mass-produce NAND flash storage.
With minimal competition, Micron and rival SK Hynix have announced dramatic investment cuts for 2023, which should help restore supply-demand equilibrium in the second half of the year.
Last year, Micron led this limited competition in technology. The memory and storage solution maker produced the first 1-beta DRAM chips and 232-layer NAND flash chips in six months. The company should profit in the second half of 2023 and beyond as demand for memory-intensive AI servers rises.
