CONNECT WITH US
Wednesday 3 December 2025
Innoscience wins US GaN patent case as ITC rules no infringement
Innoscience said at midday on December 3 via the Hong Kong Stock Exchange that it had prevailed in its US patent dispute with Infineon, marking a decisive turn in a long-running GaN intellectual property battle. The US International Trade Commission delivered the key ruling, closing a case that had drawn industry-wide attention.
Wednesday 3 December 2025
AMD, HPE expand partnership around Helios rack-scale AI architecture
AMD and Hewlett Packard Enterprise (HPE) are expanding their long-standing partnership with a new plan to deliver open, rack-scale infrastructure for the AI era. HPE becomes one of the first adopters of AMD's "Helios" architecture, a full-stack, Ethernet-based platform built to streamline and accelerate the deployment of large AI clusters.
Wednesday 3 December 2025
Nvidia CFO: US$100bn OpenAI agreement still unsettled

Nvidia CFO Colette Kress said at the UBS Global Technology and AI Conference in Arizona that the company's potential investment agreement with OpenAI of up to US$100 billion has not been finalised.

Wednesday 3 December 2025
How Gelsinger's 'Made in America' vision challenges TSMC's Arizona ambitions
Pat Gelsinger describes himself as a born-again Christian who lives "at the intersection of faith and technology." For the former Intel CEO, reviving American semiconductor manufacturing isn't just business strategy—it's a mission he believes he was called to fulfill. This conviction has driven him to donate nearly half of his income to charity and maintain strict self-discipline, waking at 4 a.m. each morning. But the same unwavering sense of purpose that guides his personal life now shapes his vision for the industry, one that fundamentally challenges how foreign companies like TSMC are investing in the US.
Wednesday 3 December 2025
South Korea receives first 13,000 Nvidia GPUs under Jensen Huang's AI infrastructure pledge
Following Nvidia CEO Jensen Huang's commitment to supply 260,000 GPUs to South Korea, the government has received its first shipment of approximately 13,000 units, including the latest B200 GPUs and a range of earlier-generation models.
Wednesday 3 December 2025
Marvell to acquire Celestial AI to accelerate shift toward optical interconnects for AI data centers
Marvell's planned acquisition of Celestial AI signals a major shift toward all-optical interconnects in AI data centers, with the company positioning itself to capture a rapidly expanding market driven by multi-rack, high-bandwidth system architectures.
Wednesday 3 December 2025
AWS plays both sides as Trainium 4 leans into NVLink
In a bid to maintain its dominance in cloud computing and challenge Nvidia's chokehold on the artificial intelligence (AI) chip market, Amazon Web Services (AWS) used its annual re:Invent conference to reveal a torrent of new products, headlined by its next-generation custom silicon and a calculated partnership with its main rival.
Wednesday 3 December 2025
Marvell reports strong third-quarter growth driven by data center demand
Marvell Technology reported strong third-quarter results driven by accelerating demand for AI data center products, with revenue, profit, and key business segments posting sharp year-over-year gains. The company also raised its outlook for next year and announced the acquisition of Celestial AI to strengthen its position in next-generation data center infrastructure.
Tuesday 2 December 2025
Android phone sales falter in China's Singles' Day promotion, pressuring SoC shipments

China's smartphone sales during the 2025 Singles' Day period posted only modest growth, and nearly all of the gains came from Apple, according to new market data. The uneven performance has raised concerns that weaker Android demand could force MediaTek and Qualcomm to scale back shipments of their flagship mobile processors earlier than usual.

Tuesday 2 December 2025
MediaTek secures Google V7e chip order, eyes 2M units by 2027
Amid rising demand for Google's TPU chips, MediaTek has emerged as a key ASIC supplier with confirmed orders for its V7e product line set to begin shipping in the second half of 2026. Market sources indicate that MediaTek expects shipments of 300,000 to 400,000 units next year, surpassing earlier conservative estimates and positioning the company to reach its US$1 billion revenue target from this segment. Looking at the full V7e product cycle, total shipments could reach at least 2 million units, with peak volume anticipated in 2027.
Tuesday 2 December 2025
Nvidia's US$2 billion Synopsys stake signals a bid to strengthen CUDA ecosystem amid ASIC challenges
Nvidia's US$2 billion investment in Synopsys marks a pivotal attempt to extend CUDA and GPU-accelerated computing from AI training into the core of industrial and semiconductor engineering, tightening its grip on the entire AI server ecosystem, as ASIC servers, particularly TPU-based, are challenging Nvidia.
Monday 1 December 2025
PaleBlueDot AI seeks US$300m loan for Nvidia chips
US AI company PaleBlueDot AI is reportedly seeking a US$300 million loan to help Chinese social media platform RedNote (Xiaohongshu) use Nvidia's high-end AI chips through a data center located in Tokyo.
Monday 1 December 2025
India commits US$500 million to modernise SCL Mohali and boost domestic chip production
India will invest INR45 billion (approx. US$500 million) to modernise and expand the Semiconductor Laboratory (SCL) in Mohali, India's Union Electronics and IT Minister Ashwini Vaishnaw said during a review visit. The minister reaffirmed that SCL would "not be privatized" and would instead undergo a major upgrade aimed at significantly boosting domestic chip production and strengthening India's strategic semiconductor capabilities.
Monday 1 December 2025
China turns to Baidu's Kunlun as AI chip constraints intensify
Following the US ban on sales of Nvidia's high-end chips to China, expectations have grown that Baidu could rapidly emerge as a new force in China's AI chip sector and become the country's second major source of domestic compute capacity after Huawei.
Monday 1 December 2025
A coming quantum shock? Pat Gelsinger says AI's GPU era is nearing its end
In a recent interview with the Financial Times, former Intel CEO Pat Gelsinger delivered a bold forecast: quantum computing could reach mainstream adoption within just two years, potentially puncturing the current AI hype. He went further, asserting that GPUs, which currently dominate AI processing, may begin to be gradually replaced by 2030.
Monday 1 December 2025
Jensen Huang says Nvidia must run very fast as AI chip competition intensifies

Nvidia CEO Jensen Huang made his fifth visit to Taiwan in 2025 on November 27, reportedly to pay a personal visit to TSMC founder Morris Chang. Speaking publicly, Huang confirmed he had met with Chang, describing the semiconductor pioneer's health as "very good," and added that he would be leaving Taiwan later that day.

Monday 1 December 2025
China's Xiangdi Xian unveils Fuxi A0, the first mass-produced IMG DXD GPU
Imagination Technologies made a prominent showing at the 31st ICCAD-Expo 2025 in Chengdu, where strategic partner Xiangdi Xian Computing Technology debuted its next-generation GPU built on the Imagination DXD architecture. Revealed via the company's WeChat account, the Fuxi A0 is confirmed as both the world's only mass-produced IMG DXD product and the first DXD chip to reach tape-out.
Monday 1 December 2025
Weekly news roundup: China phone makers pull back as Huawei advances chip tech, Nvidia strengthens Taiwan ties
These are the most-read DIGITIMES Asia stories in the week of November 24 to November 30, 2025.
Monday 1 December 2025
Nvidia and Apple compete for TSMC capacity, targeting A16 and A14 nodes
Driven by strong AI demand, TSMC continues to expand its advanced process capacity. To meet long-term needs from customers like Nvidia, TSMC is not only ramping up 3nm production at its Southern Taiwan Science Park (STSP) wafer fabs but also planning three new 2nm fabs in the same area. This capital expenditure increase over the next three years will boost the entire supply chain, including factory engineering, equipment, and materials suppliers.
Monday 1 December 2025
Jensen Huang visits Taiwan and refutes ODM capacity concerns
Nvidia CEO Jensen Huang has recently been on the move, drawing significant attention. Aside from visiting TSMC founder Morris Chang and Quanta chairman Barry Lam during his Thanksgiving trip to Taiwan, there are also rumors that Huang made the trip to address ODM shipment issues with the GB300. Supply chain companies clarified that GB300 shipments are proceeding without problems, though pushing output to meet Nvidia's required volume will need an extra push.
Monday 1 December 2025
Commentary: Nvidia’s trillion-dollar test as the AI build-out hits a harder phase
Michael Burry, famed for predicting the 2008 financial crisis, has launched a paywalled Substack called Cassandra Unchained. In his debut post, he targets the AI boom and warns that Nvidia may now sit in the same systemic hotspot Cisco occupied during the dot-com era.
Monday 1 December 2025
Riding the AI wave, Abico Asia pivots toward robotics and semiconductors

As generative AI applications surge, global venture investment has shown a steady rebound, with AI and robotics emerging as two of the most sought-after sectors. Sensing the momentum of this technological shift, Abico Asia Capital has already repositioned its portfolio and will make AI, robotics, and semiconductors its core investment pillars for 2026. The firm also emphasized its commitment to identifying Taiwan's "hidden champions" and helping them upgrade and globalize their supply chains.

Sunday 30 November 2025
Taiwan's vertical integration cluster advantage limits industry relocation risk
Amid reports that Taiwan plans to invest US$400 billion in the US to secure better tariff treatment, avoid overlapping tax rates, and reduce Section 232 impacts under the Trump administration, a Deloitte Taiwan consultant says concerns over industrial relocation and hollowing out are largely unfounded.
Friday 28 November 2025
Google's TPU strategy mounts fresh challenge to Nvidia's GPU lead

Google's expanding Tensor Processing Unit (TPU) strategy is emerging as a serious challenge to Nvidia's long-running dominance in AI accelerators, particularly after a report from The Information revealed that Meta is in talks to begin using Google TPUs in its data centers in 2027 under a potential multibillion-dollar agreement.

Friday 28 November 2025
Commentary: Taiwan's IC design edge in ASIC development
Taiwan's IC design sector has rapidly expanded in the application-specific integrated circuit (ASIC) market, driven by growing demand for AI applications. MediaTek has entered Google's TPU supply chain, Alchip Technologies has secured new Amazon Web Services programs, and Global Unichip (GUC) continues to win additional projects. Market chatter indicates GUC has landed the full back-end turnkey order for Tesla's AI5 chip, expected to contribute revenue starting in 2027.