CONNECT WITH US
Jul 21
Samsung, SK Hynix prioritize DDR5 as margins near HBM levels

Samsung Electronics and SK Hynix are reportedly allocating additional or flexible DRAM production capacity to server DDR5 and other conventional memory products while maintaining high-bandwidth memory (HBM) volumes already committed to customers.

Samsung Electronics executive chairman Lee Jae-yong, SK Group chairman Chey Tae-won and Naver founder and board chairman Lee Hae-jin are reportedly preparing to meet Nvidia CEO Jensen Huang near Silicon Valley, with a roundtable tentatively planned for July 24.
Nvidia details Vera CPU for agentic AI workloads
Jul 22, 12:32
Nvidia has outlined the design of its Vera CPU, positioning the chip for agentic AI systems that rely heavily on the processor for code execution, tool calls, retrieval, and data handling. The company said the architecture is intended to improve responsiveness, throughput, and latency consistency for developers and operators worldwide as AI infrastructure becomes more CPU-dependent.
Parts of ASML's next-generation high numerical aperture (High NA) extreme ultraviolet (EUV) machine have arrived in New York state, where they will be used for chipmaking R&D. State officials tout this as an important step in establishing a US-based research center for cutting-edge lithography as chip miniaturization begins to reach the limits of Moore's Law.
Samsung Electronics is expected to rely heavily on Qualcomm chips in its next Galaxy phones, according to market reports, a sign that the company is still prioritizing flagship performance and AI features over wider use of its in-house Exynos processors. The expected shift comes as memory prices rise and pressure builds on smartphone costs.

Chinese chip designer GigaDevice is investing another CNY500 million (approx. US$73.9 million) into its wholly owned subsidiary to support its DRAM research and industrialization project, continuing a series of capital injections aimed at expanding its memory business.

SK Hynix said in a July 22 disclosure that it had neither pursued nor decided on an acquisition of Intel's under-construction semiconductor site in New Albany, Ohio, denying a JoongAng Ilbo report that the companies were negotiating over the campus.

TSMC's planned 2027 wafer price increases are reinforcing confidence that AI chip demand remains strong while prompting fresh discussion over whether Intel and other semiconductor companies will gain greater flexibility to adjust their own pricing.

The AI industry may be riding a wave of momentum, but two clouds have recently unsettled the outlook: growing public opposition to data center construction across the United States, and a dramatic leap in the rankings by Chinese-developed model Kimi K3, which has renewed fears of AI overinvestment. NVIDIA CEO Jensen Huang addressed both concerns directly on July 21.

Nvidia's newly detailed Vera Rubin NVL72 is being pitched less as a faster chip and more as a cheaper unit of AI output. This shift matters most for the power-constrained data centers now defining the ceiling of the industry's growth. In its July 21 announcement, Nvidia said the rack-scale system delivers roughly 10 times more tokens per megawatt than its current Grace Blackwell NVL72, and about one-tenth the cost per million tokens of the GB200 NVL72. This framing positions electricity, not silicon, as the scarce resource.

Chinese semiconductor equipment suppliers are beginning to secure production orders for through-glass via (TGV) manufacturing tools, marking a shift from prototype validation toward commercial deployment as the industry moves to develop glass substrates for next-generation AI chip packaging.

Siemens said it has agreed to buy Precision Innovations in a deal that would add AI-powered chip planning tools to its electronic design automation business. The move could affect how semiconductor teams worldwide plan, test, and refine system-on-chip designs as complexity rises and development cycles come under pressure.