As 2025 nears its close, the global artificial intelligence (AI) boom is again prompting debate over whether the sector is entering speculative territory. From Silicon Valley to Wall Street and across China's tech and investment circles, comparisons with the 2000 dot-com bubble have intensified. With China now a central participant in the AI landscape rather than an observer, the debate has gained broader industry attention.

Embedded World North America 2025 took place in Anaheim, California, from November 4-6, 2025. Advantech, alongside partners including MediaTek, Qualcomm, Nvidia, and NXP Semiconductor, showcased a range of edge AI application solutions.

Microip, a provider of custom ASIC and AI software design services, has established a new business team in North America to tap into the growing demand for edge artificial intelligence. The company aims to expand its presence in sectors such as agriculture, aquaculture, and livestock management, while also targeting emerging opportunities in maritime logistics and defense-related unmanned systems. Microip projects more than 40% revenue growth in the North American market by 2026.

Baidu used its World 2025 conference on November 13, 2025, to reset its long-term AI strategy, with founder and chairman Robin Li and EVP Shen Dou detailing how the company plans to expand from silicon to large-model infrastructure. Li addressed renewed debate over an "AI bubble," while Shen outlined Baidu's latest Kunlun chips and a five-year roadmap pointing to a deeper, end-to-end build-out of China's AI stack.

President Donald Trump's mid-May visit to Saudi Arabia, Qatar, and the UAE, joined by Nvidia's Jensen Huang and AMD's Lisa Su, sparked renewed investment interest across the Middle East. Washington framed the trip as a push to promote US AI software and hardware and strengthen its influence in fast-growing regional compute markets.

Tesla is extending its in-house chip roadmap beyond vehicles and humanoid robots, with CEO Elon Musk signaling that the company's forthcoming AI8 processor could power orbiting data centers for SpaceX, even as its promised Full Self-Driving (FSD) capabilities remain incomplete.

WPG Holdings has unveiled a comprehensive group restructuring plan that will consolidate its four subsidiaries—World Peace Industrial (WPI), Silicon Application Corporation (SAC), AIT Group, and Yosun Industrial Corp.—into two core business units led by WPI and AIT. The initiative aims to boost operational efficiency and achieve economies of scale, positioning the company for stronger growth.

Microsoft has accelerated its artificial intelligence (AI) chip development by licensing proprietary chip and system design innovations from OpenAI, CEO Satya Nadella disclosed during a podcast interview. The licensing agreement grants Microsoft rights to all system-level innovations and intellectual property (IP) emerging from its collaboration with OpenAI, facilitating faster progress on Microsoft's AI accelerators and computing architectures.

South Korean AI chip startup Rebellions is accelerating its global expansion, establishing a US entity and appointing two senior Silicon Valley executives to strengthen its position in the North American market.

Infineon Technologies is ramping up its focus on AI infrastructure, projecting EUR1.5 billion (approx. US$1.74 billion) in AI server-related revenue for fiscal 2026, double the previous year's level. Executives said the company's leadership across power management technologies and strong partnerships with GPU and server makers position it to capture accelerating AI-driven growth.

Infineon Technologies expects moderate growth in fiscal 2026 as geopolitical uncertainties, tariff risks, and currency fluctuations weigh on near-term performance. CEO Jochen Hanebeck said the company remains well-positioned for long-term growth, driven by structural demand and rising opportunities in AI and energy efficiency.

Nvidia is set to adjust its AI server ODM shipment model to exert tighter control over system assembly processes. Supply chain sources indicate that Taiwan's Wistron Group, Quanta, and Hon Hai Precision Industry (Foxconn) will handle assembly of Nvidia's next-generation Vera Rubin AI server racks up to Level 10 (L10), after which Nvidia will ship products directly to customers.

Infineon Technologies struck a cautious tone on its automotive outlook during its 2025 earnings call, with CEO Jochen Hanebeck highlighting ongoing challenges in China's electric vehicle (EV) market and competitive pressures on power modules, even as demand for microcontrollers and Ethernet solutions continues to strengthen.

Infineon Technologies ended fiscal 2025 with stable revenue and a rebound in profit, supported by strong demand for power and sensor products. The company expects moderate growth in fiscal 2026, with AI data center demand emerging as a key growth driver despite continued weakness in automotive and industrial markets.

Samsung Chairman Lee Jae-yong's right-hand man, Chung Hyun-ho, has suddenly retired. Chung led the Business Support Task Force (TF), and Lee is expected to take full control of Samsung's personnel and outline a new blueprint for the company's future to restore its former glory.

Samsung will continue investment in the Phase 4 production line at its Pyeongtaek campus, converting the facility from foundry to 1c DRAM production. The latest monthly capacity is estimated at 80,000 wafers. Additionally, the previously postponed Phase 2 (Ph2) line is expected to resume investment in 2026 as a foundry production line.

The AI boom has caused memory giants to concentrate production on HBM, resulting in a shortage of general-purpose DRAM and high prices. According to Chosun Biz, DRAM prices are expected to continue rising as a product of the supply shortage. Analyses indicate that SK Hynix's general-purpose DRAM operating profit margin could reach 70%, drawing attention to its future developments.

AMD unveiled its latest CPU and GPU product roadmap during investor day, confirming the launch of the Zen 6 architecture built on TSMC's 2nm process in 2026. The next-generation Zen 7 architecture also made its first appearance on the roadmap, with industry speculation placing its release between 2027 and 2028. Both core architectures will successively power future EPYC and Ryzen processor series.

Taiwan's top IC distributors WT Microelectronics (WTMEC), WPG Holdings, Edom Technology, and Supreme Electronics reported their October 2025 results, highlighting robust AI-driven demand that continues to power growth across the semiconductor distribution sector.

The Nexperia chip crisis is gradually calming down, but the disruption has already had significant impacts on the automotive supply chain. According to IC design companies, the situation will demand greater diversification in chip sourcing. In the future, more customers from different application fields are also expected to accelerate the adoption of multiple suppliers.

At the 2025 World Internet Conference Wuzhen Summit, the China Internet Development Report 2025 showed that as of the first half of 2025, China accounted for about 40.3% of global 6G patent applications, ranking first in the world. In the field of AI patents, China also leads globally with a 60% share.

AMD used its analyst day on November 11 to outline an ambitious long-term roadmap powered by accelerating AI demand, projecting sharp gains in revenue, profitability, and market share across its core businesses. CEO Lisa Su said AMD's latest guidance excludes China and reflects stronger-than-expected momentum in AI infrastructure spending, according to Bloomberg and Reuters.

Automotive microcontroller IDM vendors Texas Instruments (TI) and Microchip both fell short of market expectations for their fourth-quarter 2025 outlook. Although the financial results were slightly below forecasts, overall revenue in the automotive and industrial control sectors has not shown a significant decline.

Intel faces a strategic shift in its artificial intelligence (AI) efforts as Sachin Katti, the company's former chief technology and AI officer, has resigned to join ChatGPT developer OpenAI. Katti will lead the design and construction of computing infrastructure to support research and expansion of artificial general intelligence (AGI).

As global semiconductor competition intensifies between the US and China, a provocative proposal emerged from Seoul's National Assembly: Taiwan and South Korea should abandon decades of rivalry to create an alternative power center in the chip industry.