Electronics manufacturing service provider Pegatron Corp (和碩) yesterday said revenue from its artificial intelligence (AI) server business is expected to grow by a triple-digit percentage this year, following a surge last year of those proportions driven by demand for AI applications.
Pegatron’s AI server business has entered a phase of rapid expansion, company co-chief executive officer Gary Cheng (鄭光志) told reporters at a company event.
In addition to its graphics processing unit-based server shipments, such as Nvidia Corp’s GB200 and GB300 models, Pegatron is also expanding into the design and production of inference servers and application-specific integrated circuit (ASIC)-based systems, Cheng said.
Photo: CNA
The firm is cooperating with several major customers in the inference area, with contributions expected to begin materializing in the second half of this year, he said.
Pegatron’s AI server customer base includes major cloud service providers (CSPs), small and mid-sized data centers, as well as enterprise customers, he said.
Training-focused projects typically generate greater revenue and are better suited to its capabilities, Cheng said, adding that the firm plans to step up deployments with large CSPs this year.
While the company has secured orders to make ASIC-powered servers, with shipments expected to begin this year, the projects would mainly focus on transformer-based training and inference applications, Pegatron co-chief executive officer Johnson Teng (鄧國彥) said.
ASIC projects differ from traditional server business, as some customers design their own AI chips and act as system solution providers, Teng said.
As a result, Pegatron’s role extends beyond board-level manufacturing to full system integration, including high-end “Level 10” and “Level 11” servers, as well as thermal and liquid-cooling solutions, which places greater demands on supply-chain coordination, Cheng said.
In the small and medium-sized enterprise and enterprise segments, Pegatron early last year secured a major Taiwanese customer and has seen steady progress in related projects, he said.
Pegatron’s new facility in Mexico, which is to focus on AI server production, entered trial production at the end of last year and is ready for mass production, he said.
Construction of a company plant in Texas is expected to be completed by the end of March, with trial production to begin soon, he said.
In Taiwan, the company last year built two new plants in Taoyuan, which are to begin production this year, Cheng said.
As local production in the electric vehicle (EV) sector becomes more pronounced across North America, as well as China and the rest of Asia, Pegatron has allocated most of its local EV product production to Mexico and North America to avoid high tariff requirements, Cheng said.
The company aims to maintain double-digit growth in its EV business this year, he said.
KEEPING UP: The acquisition of a cleanroom in Taiwan would enable Micron to increase production in a market where demand continues to outpace supply, a Micron official said Micron Technology Inc has signed a letter of intent to buy a fabrication site in Taiwan from Powerchip Semiconductor Manufacturing Corp (力積電) for US$1.8 billion to expand its production of memory chips. Micron would take control of the P5 site in Miaoli County’s Tongluo Township (銅鑼) and plans to ramp up DRAM production in phases after the transaction closes in the second quarter, the company said in a statement on Saturday. The acquisition includes an existing 12 inch fab cleanroom of 27,871m2 and would further position Micron to address growing global demand for memory solutions, the company said. Micron expects the transaction to
Nvidia Corp’s GB300 platform is expected to account for 70 to 80 percent of global artificial intelligence (AI) server rack shipments this year, while adoption of its next-generation Vera Rubin 200 platform is to gradually gain momentum after the third quarter of the year, TrendForce Corp (集邦科技) said. Servers based on Nvidia’s GB300 chips entered mass production last quarter and they are expected to become the mainstay models for Taiwanese server manufacturers this year, Trendforce analyst Frank Kung (龔明德) said in an interview. This year is expected to be a breakout year for AI servers based on a variety of chips, as
Global semiconductor stocks advanced yesterday, as comments by Nvidia Corp chief executive officer Jensen Huang (黃仁勳) at Davos, Switzerland, helped reinforce investor enthusiasm for artificial intelligence (AI). Samsung Electronics Co gained as much as 5 percent to an all-time high, helping drive South Korea’s benchmark KOSPI above 5,000 for the first time. That came after the Philadelphia Semiconductor Index rose more than 3 percent to a fresh record on Wednesday, with a boost from Nvidia. The gains came amid broad risk-on trade after US President Donald Trump withdrew his threat of tariffs on some European nations over backing for Greenland. Huang further
HSBC Bank Taiwan Ltd (匯豐台灣商銀) and the Taiwan High Prosecutors Office recently signed a memorandum of understanding (MOU) to enhance cooperation on the suspicious transaction analysis mechanism. This landmark agreement makes HSBC the first foreign bank in Taiwan to establish such a partnership with the High Prosecutors Office, underscoring its commitment to active anti-fraud initiatives, financial inclusion, and the “Treating Customers Fairly” principle. Through this deep public-private collaboration, both parties aim to co-create a secure financial ecosystem via early warning detection and precise fraud prevention technologies. At the signing ceremony, HSBC Taiwan CEO and head of banking Adam Chen (陳志堅)
RSS