CONNECT WITH US
Thursday 17 September 2026
Huawei Connect 2026: OceanStor M900 turns PB-scale KV cache into Huawei's next AI infrastructure layer
Huawei has unveiled the OceanStor M900 at Huawei Connect 2026, extending its AI infrastructure strategy into memory and storage with a petabyte-scale key-value (KV) cache system designed for AI agents and long-context inference.
LATEST STORIES
Friday 18 September 2026
ByteDance's second AI-agent phone drops forced automation for permission — yet app wall still up
The question facing China's "AI agent phone" category is no longer whether a model can drive a smartphone. It is whether the apps will let it. Judging by the first mass-produced device to try it, the answer is mostly not yet.
Friday 18 September 2026
ByteDance builds strong supply chain force to master emerging hardware ecosystems beyond core demand

The Chinese smart glasses sector is rapidly consolidating around the principle that "subtraction is king." Rather than rushing unproven hardware to market, ByteDance is adopting a deliberate, supply-chain-first strategy.

Friday 18 September 2026
Rumored Japan-US chip plant talks test whether two former chipmaking giants can rebuild at home

If the project under discussion goes ahead, it would be the first time Japan's US$550 billion tariff-linked investment pledge is directed toward chip manufacturing rather than the energy and industrial-materials projects announced so far — and the clearest sign yet that the two countries that once dominated chipmaking could rebuild some of that capacity jointly.

Friday 18 September 2026
China sees a second chance in VLEO as AI shifts the satellite race from chips to networks
Starlink's massive scale in conventional low-Earth-orbit (LEO) satellites has markedly reinforced the satellite giant's aggressive grip on the commercial space industry for the next decade or so.
Friday 18 September 2026
GaN is winning the AI power argument, not yet the cost battle

Gallium nitride (GaN) is emerging as a key power semiconductor for AI data centers, where rising rack power and the shift toward higher-voltage architectures are forcing suppliers to improve efficiency and power density. But despite rapidly expanding orders and falling costs, industry researchers say GaN remains at an early stage of adoption and is unlikely to displace silicon broadly for another two to three years.

Friday 18 September 2026
Huawei Connect 2026: Huawei scales supernodes into million-card AI systems

Huawei is extending its AI infrastructure strategy beyond individual accelerators and storage systems, unveiling an Agentic supercluster architecture designed to scale Ascend computing to as many as 1 million accelerator cards through UnifiedBus.

Friday 18 September 2026
Samsung Display reportedly taps Sunic for KRW300 billion RGB OLEDoS push
Samsung Display (SDC) is reportedly preparing to invest about KRW 300 billion in a 12-inch RGB OLED-on-Silicon (OLEDoS) production line, moving the next-generation microdisplay technology closer to mass production as the XR market shifts toward lighter glasses.
Friday 18 September 2026
Z.ai's next growth engine starts where compute scarcity ends

Chinese AI developer Z.ai has raised its 2026 year-end annual recurring revenue (ARR) target by 25% to US$3 billion after fresh financing eased the computing bottleneck that had constrained growth earlier this year.

Friday 18 September 2026
CATL looks beyond China's passenger-car squeeze as electric trucks become the next growth front
The expansive reach of CATL's entrenched battery grip on the EV market has persisted through years of undefeated market share, pinning many Chinese automakers against severe component constraints that have proved difficult to break away from.
Thursday 17 September 2026
AMD pivots to China developer ecosystems as Nvidia retreats

As Nvidia gradually retreats from China's high-end AI accelerator market amid strict US export controls, Advanced Micro Devices (AMD) is deepening its regional AI strategy. Rather than relying solely on traditional CPU and GPU sales, AMD is shifting its focus toward edge AI platforms, AI PCs, agent computers, local developer communities, and university partnerships.

Thursday 17 September 2026
Samsung bets on zHBM as 2.5D memory nears its limits
Samsung Electronics is betting on vertically stacked zHBM to move memory beyond the physical constraints of conventional 2.5D architectures, setting a system-level target of raising AI response throughput from about 100 to 1,000 tokens per second per user as workloads become more demanding.
Thursday 17 September 2026
Tier IV open-sources self-driving AI chip design to challenge Nvidia

Japanese autonomous driving developer Tier IV is developing an AI chip and will release its chip design and related toolchain as open source, allowing semiconductor manufacturers to use them to develop autonomous driving chips. The company aims to break Nvidia's dominant position in the autonomous driving chip market.

Thursday 17 September 2026
Samsung Galaxy S27 series rumored to feature 4 models, reshaping flagship lineup with display and camera upgrades
Samsung Electronics is expected to launch its Galaxy S27 series in early 2027, marking what could be its most significant product lineup restructuring in years. Market reports indicate that the newly added Galaxy S27 Pro will substantially narrow the specification gap with the Ultra model. Both high-end smartphones are expected to share top-tier features, including 60W wired fast charging, a 200-megapixel main camera, and the latest M16 OLED display. Samsung will continue to use telephoto cameras, screen sizes, battery capacities, and S Pen support to differentiate the models, establishing a new "dual high-end flagship" hierarchy.
Thursday 17 September 2026
Huawei Connect 2026: Ascend 960 puts industry-first NPO supernode at heart of Huawei's AI scale-up bet
Huawei has unveiled its next-generation Ascend 960 supernode, positioning near-packaged optics (NPO) at the centre of its effort to scale AI computing beyond the limits of individual accelerators. Huawei said the system is the industry's first supernode to adopt NPO and is designed to support large-scale training and inference for models with up to 10 trillion parameters.
Thursday 17 September 2026
South Korea's medical device city opens physical AI institute, aiming to train 1,400 workers in five years
Wonju, a South Korean city of about 360,000, roughly 90 minutes east of Seoul, opened a government-funded institute on Thursday, September 17, to train more than 1,400 workers in artificial intelligence over five years.