The artificial intelligence (AI) gold rush ignited by “generative” systems such as the overnight chatbot ChatGPT promises to be lucrative for those who can grasp the technology‘s potential, but like the early days of any gold rush , it has allowed sellers of tools such as pickaxes and shovels, which are essential for gold panning, to make a profit first.
Huida released its results after the closing hours last Wednesday (May 24) local time in the United States. Its revenue performance in the last quarter was better than market expectations. What surprised investors was that the company’s revenue forecast for this quarter was 11 billion. The U.S. dollar plus or minus 2%, more than 50% higher than analysts’ forecasts, and the company’s announcement that it will increase the supply of AI chips to meet the continued soaring demand. Inspired by this, Huida’s stock price soared 30% the next day, and its market value approached US$1 trillion. CEO Jen-Hsun Huang said after the financial report was released that the AI market is experiencing the so-called “iPhone moment.”
Other chip companies, ranging from peers like Advanced Micro Devices (AMD) to manufacturers like TSMC, have been embraced by the wave of AI mania, but the beneficiaries are not just these big names, but also some computer infrastructure suppliers, such as those Businesses engaged in cable manufacturing, air-conditioning equipment, and data center space leasing, as well as software companies that assist AI model operations and organize data.
The “AI Picks and Shovels index” of 32 such companies has risen 40 percent since ChatGPT launched last November, according to Refinitiv Datastream and The Economist. , the tech-heavy Nasdaq rose just 13%.
●Chip manufacturer
With the rapid rise of AI, the demand for computing power has also surged, and the most obvious winners are chip manufacturers, such as TSMC. Since Huida announced its financial report, TSMC’s stock price has risen by 10%, and its market value has increased by about US$20 billion.
●Chip Design Company
Chip design companies are also beneficiaries. According to a UBS survey, in the next one to two years, AI will increase demand for graphics processing units (GPUs) by US$10 billion to US$15 billion. Based on this estimate, Huida’s data center business, which accounts for 56% of the overall revenue, is expected to double its annual revenue.
AMD will launch a new GPU later this year. Stacy Rasgon, an analyst at asset management company Bernstein, said that although AMD is far behind Huida in the field of GPU design, the scale of this AI boom means that even if AMD can only get some “scum”, it can also benefit from it. .
●Chip packaging company
Companies related to chip packaging are also beneficiaries, such as Besi Semiconductor of the Netherlands. According to New Street Research analyst Pierre Ferragu, the Dutch company controls three-quarters of the high-precision bonding market. Shares of Best Semiconductor have more than halved this year.
●Network Equipment Company
GPUs in the data center need to communicate with each other, so more advanced network equipment, such as switches, routers and dedicated chips, is required. The market for such tools is expected to grow 40 percent a year over the next few years, reaching nearly $9 billion by 2027, according to research firm 650 Group.
Nvidia’s licensing business, which also includes such tools, accounts for about 78% of global sales. And its rivals, including Arista Networks, have also begun to attract investors’ attention, with their shares rising nearly 70% in the past year. In addition, communications chip maker Broadcom said its annual revenue from such semiconductors will triple to $800 million in 2023.
●Server manufacturer
IDC analyst Peter Rutten said server assembly plants are also beneficiaries of the AI boom. Dell’Oro Group predicts that the proportion of servers dedicated to AI in global data centers will expand from the current less than 10% to about 20% within five years, and the proportion of equipment capital expenditures for such servers will increase from the current From about 20% to 45%. Server makers such as Taiwan’s Wistron (3231) and Inventec (2356) will be beneficiaries.
Other smaller server makers should also get a piece of the pie. Wiwynn Technology (6669), another Taiwanese server maker, recently said that AI-related projects account for more than half of its current orders. Its U.S. peer Super Micro said AI products accounted for about 29 percent of company revenue in the three-month period to the end of April, up from the previous 12-month average of 20 percent.
●Software company
These AI hardware all require professional software to operate, some of which are provided by hardware companies to customers, such as Huida’s software platform CUDA. There are also companies developing applications that allow AI companies to manage data and large language models, such as Datagen, Pinecone, Scale AI, HuggingFace and Replicate.
●Data center landlord
The final group of AI infrastructure winners are data center landlords such as Equinix, Digital Realty. As the demand for cloud computing surged, the vacancy rate of these companies continued to decline. By the second half of last year, the data center vacancy rate fell to 3%, a record low.
(Ke Wanxiu, Times Information)
further reading
Can the AI chain still catch up?Experts warn Wistron and other 6 key
AI bubble? Funds are about to collapse?Experts reveal key facts
Aliens using AI to observe humans?Experts expose shocking insider
The post Not Huida, TSMC AI Gold Rush and these hidden winners appeared first on Business Times.