Taiwan directs too much public research and development (R&D) funding to its national research institutes, according to Chenming Hu, a former chief technology officer of Taiwan Semiconductor Manufacturing Company (TSMC). In his new book, Find the Solutions, Hu argues that this funding does not achieve the same multiplier effect as investment in universities.

Oracle has signed partnerships with Cerebras and other chipmakers to host their chips in its cloud infrastructure. The Cerebras deal gives the chipmaker a new large customer to anchor its post-IPO trajectory, while allowing Oracle to reduce its reliance on Nvidia's expensive chips, even though Oracle also announced a purchase of Vera Rubin platforms for its data centers.

AMD is backing one of its largest AI customers with capital and chips. Alongside an update on its hardware work with Meta, the chipmaker said on October 27 that it will invest US$1 billion in Meta's equity offering, joining Nvidia in buying into the social media company's fundraising for AI infrastructure. Critics describe this pattern as circular financing.

Google's tensor processing unit (TPU) business has gained one confirmed and one reported major customer within a week, adding pressure on Nvidia, whose graphics processors still dominate AI computing even as its biggest buyers look for alternatives.

Taiwanese IC design leaders MediaTek, Realtek, and Parade Technologies will hold their third-quarter 2026 earnings conferences in succession this week, offering the market an early read on conditions heading into late 2026 and early 2027. Parade will go first on October 28, followed by Realtek on October 29 and MediaTek on October 30.

Anthropic recently disclosed details of its computing partnerships with Google, Amazon Web Services (AWS), Microsoft, Advanced Micro Devices (AMD), and other providers in its S-1 filing for an initial public offering (IPO), revealing distinct cooperation models and chip technologies across its partner ecosystem.

NXP Semiconductors' latest results point to a business gaining from more complex vehicle electronics and industrial computing, while its emerging data center operation offers another route to growth. The next test is whether those opportunities can sustain demand as the company adds manufacturing capacity and customers weigh their inventory needs.

NXP Semiconductors is heading into 2027 with management arguing that its growth now rests on several structural drivers rather than a cyclical rebound. On the company's October 27 earnings call, executives said software-defined vehicles (SDV), AI at the industrial edge and data center control systems would sustain double-digit revenue growth next year.

Chinese AI chipmaker Cambricon Technologies reported a 131.49% increase in net profit attributable to shareholders for the first nine months of 2026, while revenue more than doubled. Third-quarter net profit rose nearly 150% from a year earlier.

China-based IC design firm Goke Microelectronics will raise prices for its Known Good Die (KGD) packaged products by US$0.5 to US$1.5 per unit from November 2026, depending on capacity. The move marks the second price adjustment known to customers in 2026, as AI-driven memory demand keeps double data rate (DDR) memory supplies tight and pressure spreads to niche memory and integrated packaging products.

Taiwanese startup Skymizer has formally announced a strategic partnership with Egis Technology to develop Flow (T12), a new-generation AI inference chip. Fabricated on TSMC's 12nm node and paired with LPDDR5 memory, the solution requires just three PCIe cards and 750W of power to run 3-trillion-parameter large language models (LLMs). This represents roughly one-tenth the power consumption of a comparable Nvidia H100 deployment.

Speaking at GTC Berlin and on SpaceX's third-quarter 2026 earnings call, respectively, Nvidia CEO Jensen Huang and Tesla CEO Elon Musk reached a rare consensus on a pivotal topic: GPUs are no longer rapidly depreciating consumables. Instead, their effective operational lifespans far exceed initial market expectations.

At its GTC conference in Berlin, Nvidia discussed the RTX Spark notebook, various implementations and partnerships involving physical AI, AI agents, and CUDA-X, as well as the DSX POD and European sovereign AI. Following its recently announced additional US$2 billion investment, Nvidia has committed nearly US$10 billion to Europe's technology ecosystem. The company said that building Europe's AI economy will require the region not only to consume AI, but also to produce it.

Infineon Technologies and ASMedia have announced a partnership to speed up the adoption of 20Gbps USB, driven by demand from machine-learning cameras and other physical AI devices that need to move large amounts of data quickly.

Broadcom has reached a deal with several financial firms for them to finance the purchase of US$63 billion worth of its chips for Anthropic. The agreement highlights how the line between supplier and lender is becoming blurred among some chipmakers, as well as Wall Street's growing involvement in funding arrangements as AI infrastructure spending continues to expand.

Expanding investment in AI data centers is putting pressure on memory supply and demand across the automotive industry. As semiconductor companies prioritize high-bandwidth memory (HBM) and server products, automotive memory faces rising prices and tighter supplies, making automakers' direct procurement, long-term partnerships, and in-house chip design increasingly strategic.

Nvidia CEO Jensen Huang's visit to Kazakhstan highlighted plans to build an AI computing hub for Europe, with projects spanning GPU clusters, data centers, talent, and physical AI. The initiative depends on power infrastructure, foreign technology, and Kazakhstan's ambition to develop and export its own AI capabilities across Central Asia.

After the GTC Berlin conference, Nvidia CEO Jensen Huang and Silicon Valley technology investor Gavin Baker spoke with Yahoo Finance for about half an hour. Huang confirmed Nvidia GPUs can remain in service for up to eight years and said their high residual value could help unlock financing for AI infrastructure.

Nvidia opened GTC 2026 in Berlin on October 21 with a keynote highlighting AI infrastructure, security architecture, and Physical AI, while a Greenpeace protest briefly interrupted proceedings. A robot demonstration also exposed limits in motion control, instruction comprehension, and execution stability as Europe weighs innovation, energy demand, regulation, and competitiveness.

China’s semiconductor and electronics supply chains are reporting a wave of price increases, from Goke Microelectronics raising quotes for hybrid-packaged memory products to Hygon considering higher CPU and deep computing unit (DCU) prices. The moves reflect rising costs for raw materials, components, and upstream supplies, but whether they can help China escape deflationary pressure remains uncertain.

The rapid expansion of AI data centers is driving demand beyond GPUs, memory chips, and advanced packaging. As power consumption in GPU clusters rises, data centers need not only more chips but also more efficient power delivery systems. Power plants and electrical grids outside data centers are also under pressure to close supply gaps, potentially tightening power semiconductor supplies and spreading shortages to other applications.

This week's coverage highlights shifting technology supply chains and rising demand across foldable smartphones, memory, energy storage, robotics, and AI infrastructure. Apple lifted its iPhone Duo shipment target after strong preorders, while memory makers faced conflicting forecasts over Chinese capacity and sustained AI demand. Taiwan advanced industrial microgrids and saw its first semiconductor strike, as Texas Instruments reported tightening analog supply. Chinese energy storage suppliers reportedly retreated from the US amid reciprocal technology and ownership restrictions.

HCLTech and Japan's Rapidus have inaugurated development centers in Bengaluru and Hubli, India, expanding their semiconductor engineering collaboration across Asia. The facilities will support advanced chip work, while a planned talent pipeline could connect Indian engineers with Rapidus' manufacturing ambitions in Japan, highlighting the cross-border partnerships shaping semiconductor supply chains.

Cadence Design Systems is betting that chipmakers will want AI models trained on its own design tools and run wherever customers choose. That puts it on a different route from rival Synopsys, which is building its specialized model with OpenAI on OpenAI-hosted infrastructure. On October 26, alongside its third-quarter results, Cadence introduced Maestro, a portfolio of domain-specialized AI models that power its AI super agents for chip and system design.

Cadence Design Systems used its third-quarter earnings call to argue that demand for chip design tools is no longer tied only to the AI hardware buildout. It pointed to analog and memory chipmakers, system companies, and new entrants such as space and finance firms designing their own silicon. Analysts on the October 26 call pressed two questions that will shape the coming year: how much growth comes from agentic AI rather than hardware, and whether Cadence can match the long-term targets rival Synopsys set a month earlier.