Inventec, a major ODM manufacturer, has been actively sprinting to the AI layout in recent years, and has the reputation of "the first brother of server motherboards" in the market, and its range of orders is also quite large, and it is also the supplier of AI server motherboards for Amazon AWS, mainly for a variety of AI server models from NVIDIA and AMD. With the upcoming Computex date, the company also confirmed that AMD's internal executives will appear at Inventec's booth at Computex on June 4 to share Keynote and unveil the latest sustainable AI achievements of the two parties.

Inventec management team. Reporter Wu Kangwei / Photography

AI server

As a member of the AI server foundry chain, Inventec not only holds the supply of NVIDIA's strongest B100 AI server water cooling project and its motherboard design, but also has a large proportion of NVIDIA's latest super AI chip GB200 order. In fact, Inventec not only supplies NVIDIA A100, B100, GB200 GPU and other models, but also successfully starts with orders for AMD MI300 series GPU architecture AI servers, it is reported that MI300 series products have started shipping since the first quarter of this year, which actually adds a lot of new momentum to AI server revenue.

In fact, Inventec has achieved a beautiful report card in recent years with the assistance of AI servers, and Chairman Sam Yeh recently revealed that 2024 is the first year of Inventec's "AI transformation", looking forward to 2024, the group will continue to expand production in the Czech Republic, Thailand, Malaysia and other countries, and continuous progress is also a very important step for Inventec, and the most important task in 2024 is to fill the production capacity.

Inventec revealed that the AI server market as a whole still focuses on AI server products for high-end training, while Inventec's AI server shipments are dominated by motherboards, with a gross profit margin of more than double digits, and its server motherboards account for 2% of the market. This year's shipments will focus on the H100 and A100 models, and next year will be the L40S inferred type, and the new factory in Thailand is expected to start producing laptops and servers in the fourth quarter.

Consolidated revenue

Inventec recently announced that its consolidated revenue in April reached 51.048 billion yuan, an increase of 2.7% month-on-month and 38.5% year-on-year, rushing to the highest point in nearly 22 months, and the consolidated revenue in the first four months of this year reached 181.554 billion yuan, an annual increase of 15.6%, in line with market expectations. Looking forward to the future, the internal is optimistic that this year, under the strong demand for AI from cloud service providers in the United States, the overall performance will increase by double-digit percentage year-on-year, further subdivide the server models, AI server revenue will grow by multiples, general-purpose servers, enterprise server revenue will increase by a mid-single-digit percentage year-on-year, and the future development is quite optimistic.