But the company is unlikely to keep rising forever.
“The semiconductor industry is always cyclical, so we should expect to see ups and downs in the future,” said Chris Miller, a history professor at Tufts University’s Fletcher School.
Miller, author of the New York Times bestselling book “The Chip Wars: The Battle for the World’s Most Critical Technology,” told BI that semiconductor giants like Nvidia and TSMC “are benefiting enormously from building AI.”
But that may not necessarily be the case, he added.
“The surge in spending on AI infrastructure is driving a lot of growth in the chip industry,” Miller said. “If interest and investment in AI slows, chip industry growth will slow.”
What is the Semiconductor Super Cycle?
Essentially, a semiconductor super cycle means there is a period of sustained growth and increased demand, which can last for several years.
But that rosy growth won’t last forever: semiconductor manufacturing is capital-intensive, making it extremely difficult to match supply and demand.
And companies wanting to meet the surge in demand have to invest heavily to ramp up production, which certainly increases supply levels but also risks creating an oversupply if demand fades.
This means that investors could be in trouble if they time the semiconductor supercycle wrong.
For now, the chip industry is being buoyed by a surge in AI investments.
As a result, Nvidia, which has been doing active business with Meta’s Mark Zuckerberg and Tesla’s Elon Musk, has made huge profits.
In an interview with The Verge in January, Zuckerberg touted Meta’s growing inventory of Nvidia chips, adding that he expects the company to have more than 340,000 Nvidia H100 GPUs in its possession by the end of 2024.
But some parts of the industry, such as those made with chips for consumer electronics, are already seeing a slowdown. In April, chipmaker TSMC said it was lowering its growth forecast, citing weak sales of smartphones and PCs.
The surge in data center construction has also boosted demand for semiconductors, and Miller said semiconductor companies could be hurt if data center investment slows.
To be sure, not everyone thinks the AI-driven supercycle has peaked yet.
“Digitalization, AI and rising electricity demand due to the use of renewable energies are driving demand for raw metal materials,” Thomas Lupf, chief investment officer at VP Bank Asia private bank, told BI.
“It’s still too early to declare a new super cycle,” he said, adding that there are still investment opportunities in sectors such as industrial metals and mining, which are being boosted by demand for AI.
Joshi Anant, principal in EY-Parthenon’s strategy and transactions team, told BI that greater adoption of chips will make demand more sustainable.
“Growth in semiconductors is not hype, it’s reality,” she said. “For example, in automotive with electric vehicles, in life sciences and pharmaceutical with product innovation, and in consumer goods and retail with consumer recommendations and personalization.”
Increased competition will only make things harder for incumbents like Nvidia.
Aside from the semiconductor super cycle, investors may also want to keep an eye on competitive moves from chip designers like Nvidia.
“NVIDIA is currently the leader in making chips for training and deploying AI,” Miller said, “but its success is already attracting competitors.”
Unlike TSMC, Nvidia designs its chips but doesn’t manufacture them itself, which means it’s exposed to multiple threats, including from large tech companies like Meta and Google, and from startups like Groq, Miller added.
“The barriers to entry into chip manufacturing are very high as it requires huge capital investment to set up a fab,” said EY-Parthenon’s Ananth.
“However, when it comes to chip design, the space will be prone to disruption as more technology companies invest in chip design capabilities that are specific to their products and reduce time to market,” she added.
Amazon, Meta and Google are all designing their own AI chips.
And Nvidia isn’t standing still: In May, Nvidia CEO Jensen Huang said the company would release new chips on a “one-year cadence” instead of the traditional two.
But that doesn’t mean companies will all be competing for the same piece of the pie, which may be big now, but there’s no telling when it might shrink, Miller said.
“The key question is how much the big tech companies will continue to spend on building out AI infrastructure,” he said.
A Goldman Sachs report released last month estimated that technology companies are expected to pour more than $1 trillion into AI over the next few years.
However, the investment bank warned against being overly optimistic about the return on investment in AI, given that the technology is “very expensive.”
