US cloud service providers (CSPs) are entering a new phase of AI investment, with continued growth in capital expenditures (capex), driving short-term demand for high-end AI chips. However, the lack of clear returns on these investments means tech giants are increasingly focusing on cost-performance ratios in their AI strategies.

The Chinese government has ramped up actions against Nvidia, including reopening antitrust investigations and abruptly removing all products from the company's Tmall flagship store. These measures are widely viewed as Beijing's response to recent US Department of Commerce sanctions.

Taiwan-based tier-1 high-density interconnect (HDI) board maker Compeq Manufacturing has opened its first production base in Southeast Asia located in Thailand. The facility will initially focus on producing printed circuit boards (PCB) for low-earth orbit (LEO) satellites with mass production set to begin in the first quarter of 2025.

Intel CEO Pat Gelsinger retired on December 1, 2024, leaving CFO David Zinsner and senior executive Michelle Johnston Holthaus to serve as interim co-CEOs. While the board works to identify a permanent successor, Intel faces one of the most significant crises in its history, searching for a qualified leader to navigate both internal instability and mounting external pressures increasingly challenging.

The Japanese government has accelerated plans to develop a next-generation supercomputer by the fiscal year 2030, succeeding the renowned Supercomputer Fugaku. This ambitious project, seeking to accelerate research and development, marks a significant step forward in Japan's technological landscape.

As artificial intelligence (AI) strains data center cooling limits, a seismic shift from traditional air systems to liquid cooling solutions is reshaping the US$5 billion thermal management market, industry analysts say.

FuriosaAI, a Seoul-based AI chip startup founded in 2017, is making waves in the AI hardware market with its innovative accelerator chips aimed at hyperscale data centers and enterprises. In August 2024, the startup unveiled its advanced AI chip, "RNGD." The chip is positioned to challenge industry leaders such as Nvidia with a 100% power efficiency advantage over Nvidia's H100 GPUs and competitive performance metrics. Furthermore, it recently gained attention for its testing and sampling progress with global enterprises like LG AI Research and Aramco.

The newly released book, The Nvidia Way: Jensen Huang and the Making of a Tech Giant, reveals key insights into Nvidia's corporate culture. A notable aspect is CEO Jensen Huang's preference for whiteboards over PowerPoint presentations during business discussions. Unlike meticulously formatted slides, which can obscure meaningful ideas, Huang believes whiteboards encourage clearer and more focused exchanges.

In a sweeping analysis of the AI sector, Sequoia Capital projects a widening competitive moat among leading cloud providers through 2025, as infrastructure buildout reaches critical mass. The venture firm's latest report, "AI in 2025: Building Blocks Firmly in Place," identifies a consolidation around five key players: Microsoft/OpenAI, Amazon/Anthropic, Google, Meta, and xAI.

Japanese company Nidec has announced its collaboration with Supermicro to develop a coolant distribution unit (CDU) for AI servers, which has already begun shipping. This new cooling solution boasts a performance that is 2.5x greater than existing devices of the same size and will be integrated into AI servers equipped with Nvidia GPUs.

Google has officially launched its upgraded AI model, Gemini 2.0, delivering enhanced performance and multimodal capabilities while paving the way for the future era of AI agents. Additionally, Google's sixth-generation Tensor Processing Unit (TPU) Trillium, first introduced at the I/O Developer Conference in May, is now operational to support Gemini 2.0's training and inference workloads.

Apple's cloud infrastructure progress has received much attention ever since the company began developing its Apple Intelligence technology. Currently, Apple Silicon chips are used for inference tasks, while model training is conducted in collaboration with major cloud service providers (CSP) such as Google and AWS.

Taiwan's National Center for High-performance Computing (NCHC), under the National Applied Research Laboratories (NARLabs), plans to build an AI supercomputing center in Tainan by 2025, which will surpass Nvidia's Taipei-1 in Kaohsiung in terms of computing power, according to NCHC Director General Chau-lyan Chang.

A new book, The Nvidia Way: Jensen Huang and the Making of a Tech Giant, featuring the first authorized interview with Nvidia CEO Jensen Huang, was officially released on December 10. The book provides an in-depth exploration of Nvidia's history through the lens of Huang's visionary leadership and strategic decision-making.

AWS, a subsidiary of Amazon, has significantly expanded its foundational model offerings, leading to speculation on whether AWS intends to reduce its dependence on Anthropic. Vasi Philomin, vice president of generative AI (GenAI) at AWS, stated that the goal is to provide customers with as many model options as possible, stressing that the partnership between AWS and Anthropic is a mutually beneficial relationship.

According to Pure Storage Taiwan Country Manager Jeffrey Chow, the AI boom is driving demand for storage, with many companies choosing to store their important data on-site instead of on the cloud.

Sharp and KDDI Corporation have announced a joint venture to develop an AI data center in Osaka. The two companies have entered into a Memorandum of Understanding (MoU) to transform Sharp's former Sakai facility into a state-of-the-art AI data center, marking a strategic initiative to address the growing demand for data infrastructure.

Donald Trump is expected to take office on January 20, 2025, likely being cautious with exporters running significant trade surpluses with the US. Meanwhile, US demand for electronic components and AI servers from Taiwan remains high, with Taiwanese exports to the US reaching US$101.5 billion from January to November 2024, compared to US imports in the same period totaling only US$43 billion.

Server ODMs demonstrated strong performance in the fourth quarter of 2024, driven by general purpose (GP) servers filling the gap as AI servers undergo a transitional period. ODMs project that AI servers will enter a renewed growth phase once production of Nvidia Blackwell GPUs ramps up in 2025.

Feng Ching Metal Corp., a Taiwanese manufacturer specializing in enameled and other insulated wires, anticipates stable sales growth in products catering to emerging energy applications, including AI data centers and electric vehicle (EV) wireless charging systems. According to company president Hsin-yu Chen, these trends are driving demand for high-performance insulation solutions.

Taiwan's PCB industry is experiencing a resurgence due to the rising demand for AI servers and low-Earth orbit satellites. This growth is primarily centered around rigid and flexible boards. However, global supply chain shifts are leading to a decline in Taiwan's PCB export values.

At the AWS re:Invent conference, an appearance by an Apple representative highlighted two key issues. It underscored the ongoing challenge of insufficient computing resources for major AI developers. Additionally, it suggested a new competitive landscape for companies that have invested in developing their high-end cloud chips in response to the growing market demand for increased computing power.

Taiwan's AI data center operator Zettabyte has announced that Pegatron will integrate Zware, its self-developed GPU software solution, to boost AI server performance and set a new standard for high-performance AI computing.

Hewlett Packard Enterprise (HPE) surpassed revenue and operating margin expectations in its latest quarter, driven by strong AI-server demand, though margin pressures raise questions about sustained profitability.

Foxconn reported a November revenue of NT$672.585 billion (approx. US$20.76 billion), surpassing the total earnings of the previous year in the first 11 months of 2024. Despite a dip in Smart Consumer Electronics and Cloud & Networking Products, there was growth in Computing Products and Components & Others, driven by strong demand for AI products and new Apple releases. The company expects further improvement in the second half of the year, especially in the fourth quarter.