Howard Liu
Reporter, Tech
Howard Liu is a Beijing-based technology and business reporter at the South China Morning Post, covering the semiconductor industry, AI and global tech supply chains. Previously at Caixin Media, he focused on China’s AI, semiconductor and cloud computing sectors. A graduate of Columbia University, he enjoys running, swimming and hiking.
Howard Liu is a Beijing-based technology and business reporter at the South China Morning Post, covering the semiconductor industry, AI and global tech supply chains. Previously at Caixin Media, he focused on China’s AI, semiconductor and cloud computing sectors. A graduate of Columbia University, he enjoys running, swimming and hiking.
Areas of Expertise:
AI, semiconductorLanguages Spoken:
English, Mandarin
Innolight on Thursday posted a 174 per cent surge in third-quarter net profit to US$1.29 billion, as revenue jumped 172 per cent.
The market retreat overshadows eye-catching earnings, as investors weigh near-term margin pressures against long-term AI promises.
Musk warns of AI neocloud failures, but Huang says robust demand would solve the sector’s growing liquidity crunch.
Report urges Samsung, SK Hynix to sharpen cost competitiveness in commodity memory while doubling down on cutting-edge sectors.
Advertisement
China’s remote access alone is expected to swell to 2.8 million H100-equivalents by the end of 2027, up from roughly 1 million this year.
In the period to 2030, the agentic share of the region’s AI software revenue is projected to rise to 12.3 per cent from 1.7 per cent.
The conference, focusing on sovereign AI, automation and infrastructure, comes days after Jensen Huang’s meeting with German Chancellor.
The firm also forecasts a 742 per cent revenue surge year on year.
The company has reportedly filed dozens of patent families related to the technology, which can improve chip performance.
GPU designer MetaX, listed in Shanghai and may soon list in Hong Kong, says it turned a profit in the first nine months thanks to more orders.
The negotiations underscore TSMC’s push to expand production and satisfy insatiable demand for AI chips from key clients like Nvidia.
Shanghai-based Langdao Silex displays self-developed EUV light-source system, marking crucial step in Beijing’s chip self-reliance drive.
Country’s top NAND maker bets on enterprise SSDs and AI infrastructure to drive next phase of growth, Counterpoint says.
Investors seem to be weighing strong earnings prospects against the risk of slower memory price growth as supply catches up with demand.
US chip production relies on ‘vulnerable inputs hidden deep in upstream global supply chains’, SIA says.
The investment underscores the growing importance of advanced packaging materials in the semiconductor supply chain.
Upstream laser chip suppliers bore the brunt, suffering steeper declines than downstream optical transceiver manufacturers.
DUV tools, though not as advanced as EUV, can be adapted to manufacture 7-nm logic chips and advanced memory for AI processors, report says.
Expanding China’s intelligent computing capacity fourfold in four years will require advances in power-grid management, domestic chips and software.
As it begins massive output increase, China’s top memory supplier sets ‘aggressive’ target for local chipmaking equipment, source says.
SemiAnalysis estimates that China has more than 24GW of delivered data centre capacity, with another 20GW in the pipeline.
US lawmakers seek broader powers to restrict technology products as Lenovo faces a patent investigation.
Biwin Storage Technology invests US$672 million in latest example of a storage firm looking to upgrade its business model.
Kirin 9050 Pro chip, the first based on Tau framework with 7nm technology, outperforms Apple’s 3nm A17 Pro in tests, research firm says.
After failing to unseat the QR code years ago, Beijing is betting that a “micro” version of its home-grown barcode can gain a foothold in the global semiconductor supply chain.
The firm’s Ascend chips are ahead of Nvidia in domestic market, Huawei’s Eric Xu Zhijun says.
The Chinese tech giant is leveraging its UnifiedBus technology to link hundreds or thousands of chips into single computing unit.
Company says ‘performance-doubling’ Ascend 960DT training chip to debut in first quarter, followed by 960PR inference chip in third quarter.
Its SuperPoD will face off against offerings by US rivals in the global computing power arena, including Nvidia’s NVL system.
