Taiwan's heavy-electrical and electromechanical industry is entering a new growth phase as demand expands beyond its long-standing reliance on Taiwan Power (Taipower). While Taipower's grid modernization remains the sector's foundation, semiconductor investments, AI data center construction, and overseas infrastructure projects are creating multiple growth drivers. At the same time, project-based business models are making revenue recognition increasingly uneven despite record order backlogs.
Nvidia CEO Jensen Huang has delivered a clear message ahead of his trip to Japan: the chipmaker's growth is still accelerating, even with quarterly revenue approaching US$100 billion.
As AI moves from experimentation to deployment, enterprises in Singapore and across ASEAN are shifting their focus from cloud adoption to multicloud integration, cybersecurity, and sovereign data control. Cloudmile chief commercial officer Jeremy Heng said the next phase of AI will be defined less by model hype than by the infrastructure needed to make systems secure, flexible, and commercially useful.
Singapore's relatively neutral stance on cloud adoption is helping make performance, compliance, and value the main factors in global companies' decision-making. As AI moves from pilot projects to production, businesses across ASEAN are rethinking which cloud, model, and location best fit each workload.
Huahsu, a semiconductor materials supplier established in 1997 through a joint investment by Japan's Asahi Kasei and Wah Lee, has launched a new capacity expansion plan as demand for AI, HPC, and advanced semiconductor packaging accelerates. The company recently held a completion ceremony for its second high-resolution dry film photoresist processing plant.
India is broadening its technology strategy from attracting chip factories to building semiconductor equipment, materials, AI capabilities, supply chains and trade partnerships as it seeks a larger role in global electronics.
MiTAC Computing Technology, a subsidiary of MiTAC Holdings, showcased its server lineup and high-density liquid-cooled rack solutions at the 2026 World Artificial Intelligence Conference (WAIC) in Shanghai.
Humanoid robots are very hyped these days. They are performing ever more impressive feats, from pulling off coordinated dance performances to running half-marathons. Yet one investor believes that the robots ultimately winning the commercial race will be the "boring" ones with solid, real-world market potential.
The humanoid robot industry is converging on a "big brain, small brain" architecture, with AI compute shifting from the cloud to the edge, and even to hands, feet, and other endpoints. DIGITIMES Intelligence predicts that Nvidia's CUDA will keep it dominant in the robot "big brain" layer for now, but automotive chipmakers and field-programmable gate array (FPGA) vendors can still target the "small brain" and endpoint edge-compute market to break into the humanoid robot ecosystem.
AI development is driving larger data transfers and higher GPU efficiency demands, pushing memory toward customization and prompting South Korean industry watchers to call for a shift to a "memory foundry" model. Sungkyunkwan University professor Seokjoon Kwon said at the Nano Korea 2026 forum that memory makers must move from mass production to order-driven design tailored to customer needs.
The competitive race to dominate artificial intelligence is driving technology giants to over-build computing capacity by roughly half again more than is economically efficient — and the same contest, financed with debt and circular equity ties, is quietly manufacturing the conditions for a sector-wide bust. That is the central argument of a new Bank for International Settlements (BIS) working paper that puts formal numbers on a warning the institution has been sounding for weeks.
More coverage
