CONNECT WITH US
Tuesday 7 April 2026
Global AI chip suppliers compete as TSMC remains top foundry partner
As the artificial intelligence (AI) era advances, approximately 133 companies are actively developing or selling AI chips, according to a SEMIEcosystem report citing Jon Peddie Research. Major suppliers include Nvidia, AMD, Broadcom, and Google, alongside numerous startups focusing on edge AI solutions.
Tuesday 7 April 2026
Google's chip revisions raise questions for MediaTek's growth plans

Recent reports suggest that Google has once again made engineering changes to its Tensor Processing Unit, or TPU, pushing the chip's tape-out to around mid-2026. The product in question—known as the v8x and designed by MediaTek—has raised fresh concerns about whether MediaTek can scale its application-specific integrated circuit, or ASIC, business as planned this year.

Tuesday 7 April 2026
Anthropic secures 3.5 GW of next-gen compute via landmark alliance with Google and Broadcom

Anthropic, Google, and Broadcom today announced a massive expansion of their strategic partnership, unveiling a multi-year roadmap that secures approximately 3.5 gigawatts (GW) of next-generation AI computing capacity for Anthropic.

Tuesday 7 April 2026
Analysis: NVLink licensing reshapes semiconductor alliance; potential Broadcom truce
In September 2025, Nvidia CEO Jensen Huang made a rare joint livestream appearance with Intel CEO Pat Gelsinger to announce a US$5 billion equity investment in Intel. In March 2026, Nvidia followed up with a US$2 billion investment in Marvell Technology. Why Huang is investing in potential competitors so aggressively remains a question.
Tuesday 7 April 2026
With Vera, Nvidia stops borrowing and starts building its own CPU muscle
Since artificial general intelligence (AGI) depends heavily on the CPU, Nvidia CEO Jensen Huang aims to build the most efficient "AI factory" by tightly controlling CPU development. At GTC 2026, Nvidia aggressively promoted its next-generation self-developed Arm-based CPU designed specifically for agentic AI.
Monday 6 April 2026
Smartphone order cuts hit IC design as recovery pushes to 2027
Market sentiment toward the 2026 smartphone outlook has turned increasingly pessimistic. On the supply side, rising component costs have expanded beyond memory to include across-the-board chip price hikes, making it unclear whether the initial round of handset price increases seen in early 2026 will be the only one.
Saturday 4 April 2026
Memory interface chip maker Montage sees DDR5 surge drive 2025 growth
Montage Technology reported full-year 2025 revenue of CNY5.46 billion (US$750 million), up 49.9% year on year, while net profit rose 58.4% to CNY2.24 billion, broadly in line with expectations. Adjusted net profit reached CNY2.02 billion, up 62.0%, reflecting solid underlying demand.
Thursday 2 April 2026
IBM and Arm collaborate on dual‑architecture hardware to broaden enterprise AI deployment
IBM and Arm announced a collaboration to build dual‑architecture hardware aimed at running AI and data‑intensive workloads with more flexibility, reliability, and security, potentially affecting enterprise infrastructure worldwide by expanding software choice, easing workload portability, and influencing how organizations deploy mission‑critical applications across cloud and on‑premises environments in the near term.
Thursday 2 April 2026
Samsung reportedly supplies Exynos modem chips to Fibocom for 5G module mass production
Samsung Electronics has supplied its Exynos modem chipset to Chinese wireless module maker Fibocom, marking a notable expansion of its system semiconductor business into external clients, according to inews24 and Greened.
Thursday 2 April 2026
Intel expands SambaNova stake, governance questions surface

Intel plans to invest a further US$15 million in AI chip startup SambaNova Systems, lifting its stake to about 9% pending regulatory approval, according to Reuters. The move follows a US$35 million investment in February that raised Intel's ownership to 8.2% from 6.8% a year earlier and formalised a strategic collaboration between the two.

Thursday 2 April 2026
Nvidia invests US$2 billion in Marvell to integrate NVLink Fusion with ASICs
On March 31, Nvidia announced a US$2 billion investment in Marvell and plans to further integrate its NVLink Fusion technology with Marvell's XPU services for customer use. Although Nvidia revealed partnerships with several ASIC service providers around NVLink Fusion technology earlier in 2025, this direct investment signals a closer collaboration between Nvidia and Marvell. The move raises questions about how the two companies will expand their presence in the cloud AI market and whether ASIC customers will embrace this integrated solution.
Thursday 2 April 2026
South Korea's market volatility hands Taiwan a rare edge in AI
During a lecture hosted by the Chinese National Association of Industry and Commerce (CNAIC), DIGITIMES Chairman Colley Hwang analyzed the East Asian industrial landscape. While headlines often focus on the chip wars between the US and China, Hwang shed light on a quieter, more structural divergence: the widening "resilience gap" between Taiwan and South Korea, as manifested through the lens of currency.
Thursday 2 April 2026
Chinese companies capture nearly 41% of domestic AI accelerator server market
China's AI accelerator server market is gradually shifting away from using Nvidia chips to domestic GPU and AI chip makers, with Chinese companies capturing almost 41% of the market last year, according to an IDC report.
Thursday 2 April 2026
Commentary: What's the real game behind Arm's chip venture?
Amid AI-driven shifts, Arm launched its AGI CPU in March 2026 to address system-level optimization lacking in highly customized data center CPUs. Partnering with Meta and supported by OpenAI, Arm seeks to offer a standardized solution that enhances ecosystem efficiency without directly competing with clients.
Thursday 2 April 2026
Intel buyback signals shift beyond austerity as chipmaker regains confidence
Intel's US$14.2 billion buyback of its Ireland fab stake signals a shift beyond austerity, reflecting improved finances, renewed confidence in AI-driven CPU demand, and a strategic move to regain full control of key manufacturing capacity amid persistent global semiconductor supply constraints.
Thursday 2 April 2026
Denso targets US$54 billion revenue, Rohm deal faces Toshiba-Mitsubishi alliance hurdle
Japan's auto parts supplier Denso on March 31 unveiled a mid-term business plan through March 2031, targeting revenue of JPY8 trillion (approx. US$54 billion) and a return on equity (ROE) of 11%. The strategy underscores a shift toward semiconductors as a core growth driver, alongside vehicle electrification and intelligent systems, signaling ambitions beyond its traditional role as an automotive supplier.
Thursday 2 April 2026
Interview: How Europe's two-year AI lag is becoming a security liability
While North America and Asia race to deploy "physics-bending" AI hardware, Europe has opted for a "wait-and-see" approach, shielded by thick layers of regulation. In a DIGITIMES Asia interview, CONTEXT World senior analyst Aaron Smith argues this is creating a vulnerability window. As Europe moves cautiously to ensure regulatory compliance, it is inadvertently extending the life of aging, insecure legacy systems. Meanwhile, advanced AI cyberattacks are intensifying, adding pressure to European enterprises.
Thursday 2 April 2026
Top 10 Chart: Markets signal AI shift as Taiwan IC designers decouple from early-year revenues

Taiwan's IC design sector is undergoing a decisive shift in 2026, with capital markets effectively validating an industry transition toward artificial intelligence (AI), even as recent revenue data still reflects older demand cycles.

Wednesday 1 April 2026
Fujitsu plans 1.4nm AI chip Japan-based production with Rapidus
Fujitsu plans to develop a 1.4nm artificial intelligence (AI) chip and outsource its production to Rapidus in Japan, according to Nikkei. The chip is designed for use in servers and related systems, with a focus on low power consumption and domestic manufacturing.
Wednesday 1 April 2026
China memory chip maker Longsys unveils SPU architecture for edge AI storage systems

As AI shifts from cloud training to edge inference, the memory stack is moving beyond data access toward system-level coordination, reshaping controller design, supply chain roles, and value distribution.

Wednesday 1 April 2026
Nvidia's 800V-to-12V push faces industry skepticism
Following Nvidia's promotion of its 800V-to-12V solution at the GPU Technology Conference (GTC) 2026, major power semiconductor and IDM players have successively launched related voltage converters and peripheral chip solutions. These emphasize their capability to directly convert 800V to 12V or 6V within the power supply system, thereby reducing the number of voltage conversion stages and enabling shorter, more direct power transmission paths.
Wednesday 1 April 2026
Nvidia's US$2 billion bet on Marvell turns rivals into ecosystem partners
Nvidia's US$2 billion investment in Marvell signals a shift toward hybrid AI infrastructure, balancing dominance with openness while intensifying competition across custom silicon, networking, and next-generation interconnects.
Wednesday 1 April 2026
China memory chip designer GigaDevice posts record 2025 earnings, locks in US$825 million DRAM supply
GigaDevice Semiconductor reported stronger full-year 2025 results, supported by AI-driven demand and a recovery in the memory cycle. Revenue rose 25.12% year on year to CNY9.2 billion (US$1.33 billion), while net profit increased 49.47% to CNY1.65 billion. Adjusted net profit grew 42.57% to CNY1.47 billion. The company plans a dividend of CNY7.5 per 10 shares.
Wednesday 1 April 2026
Arm, Tesla chip push reshuffles supply chains, lifts AI memory demand
Global technology and semiconductor companies are moving to build independent chip ecosystems to meet rising artificial intelligence (AI) demand, a shift that stands to reshape semiconductor supply chains.
Wednesday 1 April 2026
China GPU maker Biren triples revenue on AI data center demand

Biren Technology reported 2025 revenue of CNY1.04 billion (approx. US$150.47 million), up 207.2% year-over-year, supported by demand from domestic data centers and AI enterprise customers. Gross margin rose to 53.8%, up 0.63pp. The company's BR10X general-purpose GPU remained the main revenue contributor, while its next-generation BR20X is set for launch in 2026.