Nvidia Corp on Wednesday overshot Wall Street estimates, as its profit skyrocketed bolstered by its dominance in chipmaking, which has made the company an icon of the artificial intelligence (AI) boom.
Its net income rose more than sevenfold from US$2.04 billion a year earlier to US$14.88 billion in its first quarter. Revenue more than tripled from US$7.19 billion the previous year to US$26.04 billion.
“The next industrial revolution has begun,” Nvidia CEO Jensen Huang (黃仁勳) said on a conference call with analysts.
Photo by Joel Saget, AFP
Companies buying Nvidia chips would use them to build a new type of data center, designed to produce a new commodity: AI, Huang said, calling them “AI factories.”
Training AI models is becoming a faster process, as they learn to become “multimodal” — that is, capable of understanding text, speech, images, video and 3D data — and able to “reason and plan,” Huang added.
The company reported earnings per share — adjusted to exclude one-time items — of US$6.12, well above the US$5.60 that Wall Street analysts had expected, FactSet said.
Nvidia also announced a 10-for-1 stock split, a move that it said would make its shares more accessible to employees and investors.
The company increased its dividend from US$0.04 a share to US$0.10.
Nvidia shares rose 6 percent in after-hours trading to US$1,006.89 on Wednesday. The stock has risen more than 200 percent in the past year.
The company now boasts the third-highest market value on Wall Street, behind only Microsoft Corp and Apple Inc.
“Nvidia defies gravity again,” Emarketer analyst Jacob Bourne said of the quarterly report.
While many tech companies are eager to reduce their dependence on Nvidia, which has achieved a level of hardware dominance in AI rivaling that of earlier computing pioneers such as Intel Corp, “they’re not quite there yet,” Bourne said.
Demand for generative AI systems that can compose documents, make images and serve as increasingly lifelike personal assistants has fueled astronomical sales of Nvidia’s specialized AI chips over the past year.
Tech giants Amazon.com Inc, Google, Meta Platforms Inc and Microsoft Corp have signaled that they would need to spend more in coming months on the chips and data centers needed to train and operate their AI systems.
What happens after that could be another matter.
Some analysts believe that the breakneck race to build those huge data centers would eventually peak, potentially spelling trouble for Nvidia in the aftermath.
“The biggest question that remains is how long this runway is,” Third Bridge analyst Lucas Keh said.
AI workloads in the cloud would eventually shift from training to inference, or the more daily task of processing fresh data using already trained AI systems, Keh said.
Inference does not require the level of power provided by Nvidia’s expensive top-of-the-line chips, which would open up market opportunities for chipmakers offering less powerful, but also less costly, alternatives, Keh added.
When that happens, “Nvidia’s dominant market share position will be tested,” Keh said.
IN THE AIR: While most companies said they were committed to North American operations, some added that production and costs would depend on the outcome of a US trade probe Leading local contract electronics makers Wistron Corp (緯創), Quanta Computer Inc (廣達), Inventec Corp (英業達) and Compal Electronics Inc (仁寶) are to maintain their North American expansion plans, despite Washington’s 20 percent tariff on Taiwanese goods. Wistron said it has long maintained a presence in the US, while distributing production across Taiwan, North America, Southeast Asia and Europe. The company is in talks with customers to align capacity with their site preferences, a company official told the Taipei Times by telephone on Friday. The company is still in talks with clients over who would bear the tariff costs, with the outcome pending further
A proposed 100 percent tariff on chip imports announced by US President Donald Trump could shift more of Taiwan’s semiconductor production overseas, a Taiwan Institute of Economic Research (TIER) researcher said yesterday. Trump’s tariff policy will accelerate the global semiconductor industry’s pace to establish roots in the US, leading to higher supply chain costs and ultimately raising prices of consumer electronics and creating uncertainty for future market demand, Arisa Liu (劉佩真) at the institute’s Taiwan Industry Economics Database said in a telephone interview. Trump’s move signals his intention to "restore the glory of the US semiconductor industry," Liu noted, saying that
STILL UNCLEAR: Several aspects of the policy still need to be clarified, such as whether the exemptions would expand to related products, PwC Taiwan warned The TAIEX surged yesterday, led by gains in Taiwan Semiconductor Manufacturing Co (TSMC, 台積電), after US President Donald Trump announced a sweeping 100 percent tariff on imported semiconductors — while exempting companies operating or building plants in the US, which includes TSMC. The benchmark index jumped 556.41 points, or 2.37 percent, to close at 24,003.77, breaching the 24,000-point level and hitting its highest close this year, Taiwan Stock Exchange (TWSE) data showed. TSMC rose NT$55, or 4.89 percent, to close at a record NT$1,180, as the company is already investing heavily in a multibillion-dollar plant in Arizona that led investors to assume
AI: Softbank’s stake increases in Nvidia and TSMC reflect Masayoshi Son’s effort to gain a foothold in key nodes of the AI value chain, from chip design to data infrastructure Softbank Group Corp is building up stakes in Nvidia Corp and Taiwan Semiconductor Manufacturing Co (TSMC, 台積電), the latest reflection of founder Masayoshi Son’s focus on the tools and hardware underpinning artificial intelligence (AI). The Japanese technology investor raised its stake in Nvidia to about US$3 billion by the end of March, up from US$1 billion in the prior quarter, regulatory filings showed. It bought about US$330 million worth of TSMC shares and US$170 million in Oracle Corp, they showed. Softbank’s signature Vision Fund has also monetized almost US$2 billion of public and private assets in the first half of this year,