News
According to a report by Commercial Times, as NVIDIA’s GB200 is set to officially ship by the end of this year and gradually ramping up in 2025, it is anticipated to usher in a new era of liquid cooling.
TrendForce predicts that the liquid cooling penetration will rise from 11% in 2024 to 24% in 2025. This growth is expected to be fueled by advancements in AI, which we may witness in 2025.
The expansion of liquid cooling will benefit several Taiwanese companies, including Delta Electronics, Asia Vital Components, Auras Technology, LOTES, and Fositek.
The growing global awareness of ESG principles, coupled with cloud service providers (CSPs) accelerating the deployment of AI servers and the upcoming official launch of NVIDIA’s Blackwell platform all contribute to the growth of liquid cooling technology, as the report in Commercial Times indicated.
According to the report, the market is optimistic that the significant growth in NVIDIA Blackwell’s penetration will drive a shift from air cooling to liquid cooling solutions.
Currently, the Hopper architecture remains dominant, commanding 90% of the market share this year. On the other hand, Blackwell is projected to capture only 4% of the market by the end of 2024. However, as Blackwell ramps up sales next year, it is expected to replace Hopper and establish itself as the new mainstream.
From the perspective of thermal design power (TDP), NVIDIA’s GB200 NVL72 racks, with a TDP of approximately 140 kW, will require liquid cooling solutions to manage heat dissipation, with Liquid-to-Air (L2A) technology anticipated to become the predominant method.
While immersion cooling is an option, it faces technical challenges that may prevent it from becoming mainstream in the next 3 to 5 years, as the report stated.
The report highlighted that the market price of a single NVL72 rack is 28 times that of a traditional server. As a result, an increasing number of companies are entering and competing in the liquid cooling market.
Currently, for those Taiwanese companies in the supply chain which may benefit from the market opportunities, Asia Vital Components and Cooler Master lead in providing cold plates, while Cooler Master and Auras supply manifolds, and Vertiv and Delta Electronics provide coolant distribution units.
On the other hand, quick disconnect (QD) components are critical for preventing leakage and have frequently been reported as out of stock. Taiwanese companies like LOTES and Fositek are reportedly in the validation phase, and by the first half of 2025, they are expected to join the list of QD suppliers, to compete with major companies such as CPC, Parker Hannifin, Denfoss, and Staubli.
Read more
(Photo credit: NVIDIA)
News
According to a report from Commercial Times citing sources, it’s revealed that NVIDIA has executed changes to the Blackwell series’ 6-layer GPU mask. Therefore, the process can now proceed without re-taping out, as production delays being minimized.
The report noted that NVIDIA’s updated version of B200 is expected to be completed by late October, allowing the GB200 to enter mass production in December, with large-scale deliveries to ODMs expected in the first quarter of next year.
Previously, as per a report from The Information, NVIDIA’s GB200 was said to be experiencing a one-quarter delay in mass shipments. Another report from the Economic Daily News further suggested that the problem likely lies in the yield rates of advanced packaging, which mainly affected the non-reference-designed GB200 chips.
Industry sources cited by Commercial Times addressed that NVIDIA’s Blackwell chip used to be facing instability in metal layers during the HV process, which was then resolved by July.
In addition, since the issue reportedly occurred in the back-end-of-line process, a new tape-out was deemed unnecessary. Still, as CoWoS-L capacity remains a bottleneck, the advanced packaging for GB200 this year is expected to adopt CoWoS-S.
Read more
(Photo credit: NVIDIA)
News
According to a report from Economic Daily News, on the evening of August 21st, Foxconn announced plans to expand its investments, increasing capital in its subsidiaries located in the U.S., Mexico, India, and Europe. The total investment amounts to roughly USD 840 million.
First, Foxconn announced earlier that its subsidiary, Cloud Network Technology USA Inc., has acquired shares of Foxconn Assembly LLC. The transaction is valued at USD 253 million.
Sources cited by the Economic Daily News suggests that this move is looking to boost the production capacity of its plant in Houston, Texas. Foxconn currently manufactures AI servers in three locations across North America: Mexico, Wisconsin, and Texas. This indicates that Texas is gradually becoming a key hub for AI server production.
Secondly, Foxconn announced that its subsidiary, Cloud Network Technology Singapore Pte. Ltd., has acquired shares of FII AMC MEXICO S. DE R.L. DE C.V. The transaction is valued at USD 241 million.
It is speculated by the Economic Daily News that this move is primarily aimed at increasing the production capacity of Foxconn’s subsidiary, FII (Foxconn Industrial Internet), in its Mexico plant.
FII previously stated that the initial production of the GB200 servers would start in Taiwan, with the related capacity already in place.
The first overseas production line for the GB200 servers is reportedly to be set up at the Mexico plant, which is already producing AI servers, with small-scale production of the GB200 expected to begin as early as the third quarter.
Thirdly, Foxconn announced that its subsidiary, Foxconn Interconnect Technology Limited, has acquired 197 million ordinary shares of Foxconn Interconnect Technology Singapore Pte. Ltd., valued at approximately EUR 180 million (roughly USD 200.53 million).
Per Economic Daily News, it is speculated that this move is related to Foxconn’s subsidiary, FIT (Foxconn Interconnect Technology), which previously announced the acquisition of shares in the German Auto-Kabel Group to strengthen its presence in the automotive electrification sector and expand its customer base.
Lastly, Foxconn announced that its subsidiary, Foxconn Singapore Pte Ltd, has acquired 1.203 billion ordinary shares of Foxconn Hon Hai Technology India Mega Development Private Limited, valued at approximately USD 144 million .
Reportedly, it is speculated that this investment aims to boost the capital of Foxconn’s Indian subsidiary.
As Foxconn is preparing for mass production of the iPhone 16 Pro and Pro Max in India, this year marks the first time Apple is integrating AI applications (Apple Intelligence) into the latest iPhone 16 Pro series.
Read more
(Photo credit: Foxconn)
News
At its earnings call on August 14th, Taiwanese tech giant Foxconn addressed the progress of the highly anticipated launch of NVIDIA’s next-generation AI server, the GB200.
According to a report by Commercial Times, Foxconn confirmed that the development of the server cabinets is on schedule, and the company will be among the first suppliers to deliver, with shipments expected to begin in the fourth quarter. The company also expects significant growth throughout the year, driven by strong demand for AI servers.
Addressing concerns about the progress of its AI server business, Foxconn spokesperson James Wu stated that driven by strong customer demand, the company’s AI server revenue grew by more than 60% quarter-over-quarter in Q2, accounting for over 40% of its total server revenue.
Wu reaffirmed that the robust demand for AI servers is expected to continue, and the company maintains its forecast that AI servers will contribute 40% of its total server revenue for the year.
Additionally, Foxconn has observed strong demand from various types of customers for the new generation of AI cabinet solutions, which is expected to significantly contribute to its server revenue in 2025.
Previously, per a report from The Information, NVIDIA’s GB200 is said to be experiencing yield issues, leading to a one-quarter delay in mass shipments. When asked about potential delays in the shipment of GB200 AI servers, Wu responded that the development timeline for the GB200 cabinets is on schedule.
In the second quarter, Foxconn reported revenue of NTD 1.55 trillion (roughly USD 48 billion), with a gross margin of 6.37% and earnings per share (EPS) of NTD 2.53.
For the first half of the year, its revenue reached NTD 2.87 trillion (roughly USD 88.9 billion), marking a 4% year-over-year increase. Gross margin, operating margin, and net profit margin were 6.37%, 2.83%, and 1.98%, respectively, all showing improvements compared to the same period last year, which stood at 6.21%, 2.58%, and 1.66%.
Cumulative EPS for the first half was NTD 4.12. Looking ahead to the third quarter, Foxconn expects operations to gain momentum as they enter the traditional peak season, with both quarter-over-quarter and year-over-year growth anticipated.
Read more
(Photo credit: Foxconn)
News
According to a report from The Information, NVIDIA’s “world’s most powerful AI chip,” the GB200, is said to be experiencing yield issues, leading to a one-quarter delay in mass shipments.
As per sources cited by a report from the Economic Daily News, it’s suggested that the problem likely lies in the yield rates of advanced packaging, mainly affecting the non-reference-designed GB200 chips.
The supply of the reference-designed GB200 chips remains stable, with Foxconn being the sole contract manufacturer receiving an adequate supply of these chips. Foxconn is set to ship according to the original schedule in the fourth quarter.
Furthermore, the sources cited by the same report point out that Foxconn is currently the only manufacturer able to meet the scheduled shipment of the GB200 in the fourth quarter. This is primarily due to Foxconn securing NVIDIA’s reference-designed GB200 chips orders, which are prioritized for shipment amid the supply shortage.
The term “reference-designed” refers to the GB200 AI servers ordered by NVIDIA for production at Foxconn and other manufacturers. These products are made according to NVIDIA’s reference designs and are not customized. Once produced, they can be sold to cloud service providers (CSPs) and other clients.
In contrast, “non-reference-designed” refers to customized versions of the GB200, which are tailored to specific customer requirements. The current yield issues are affecting the production of these non-reference-designed items, with the priority given to shipping the reference-designed products first.
Following the reports addressing the tight supply of GB200, customers are said to be scrambling to secure their orders from Foxconn due to its ample chip supply. Foxconn, traditionally silent on customer and order details, will reveal the latest status of its product lines during the press conference on August 14th.
The GB200 was originally scheduled for mass shipments starting in the fourth quarter of this year. However, over the weekend, reports emerged about yield issues, pushing the mass shipment timeline to the first quarter of next year, causing a stir among the market.
Read more
(Photo credit: NVIDIA)