Marvell Technology is betting that the next phase of AI infrastructure growth will require substantially more networking and optical connectivity, raising its fiscal 2028 revenue outlook to approximately US$20 billion and setting a fiscal 2031 target of US$70 billion to US$90 billion. The company's strategy centers on extending existing hyperscaler relationships across switches, optics, and custom chips surrounding AI processors.

AI server supply chains are accelerating their shift toward regional manufacturing. However, the labor access, wage costs, and capacity expansion constraints of US manufacturing are forcing manufacturers to raise automation levels to balance delivery speed and production efficiency. That has pushed Foxconn and Nvidia to jointly invest in precision assembly automation research, with an eye toward reducing manual handling.

The AI server supply chain is poised for a stronger quarter as Nvidia's Vera Rubin and AWS's Trainium 3 gradually ramp up, with momentum expected to strengthen across the supply chain from ODMs to component suppliers.

At IROS 2026, one of the world's top robotics conferences, held in Pittsburgh at the end of September, a robot startup's booth featured a dual-arm robot catching, in midair, sticks dropped at random from above. The demonstration is one of many this year showing robots dancing, doing parkour, and even running half marathons—but the real question is what those performances actually prove about physical AI.

Catcher Technology said its existing AI products continued to expand shipments as the year-end period approached, while medical and drone businesses were entering mass production. The casing maker also said additional AI server projects are scheduled to begin mass production from the first quarter of 2027, with AI PC-related products expected to follow in the second quarter of 2027.

Transcend reported a sharp recovery in September 2026 revenue as shipments for industrial PC and networking customer projects picked up. The memory module maker said monthly consolidated revenue reached NT$6.04 billion (approx. US$190 million), up 24.8% from August and 290% from a year earlier.

The excitement around massive orbital infrastructure mainly stems from its tremendous potential in offering compute nodes powered by near-continuous solar energy beyond the spatial limits of terrestrial infrastructure.

Reflection AI, the Nvidia-backed startup that has cast itself as a US counterweight to China's open-model developers, has unveiled Beam, its first open-weight model, which it says rivals leading options in the US and China while costing less than many competitors.

Demand for agentic AI in commercial and consumer markets was rising rapidly, and the market appeal of consumer services such as Meta Muse was expanding. In agentic AI tasks, large language model (LLM) inference was mainly handled by GPUs or application-specific integrated circuits (ASICs), while many other functions relied on CPUs, which was expected to boost CPU demand. Consumer-facing harness applications were expected to adopt enhanced general-purpose servers, while enterprise core automation would use high-density CPU racks deployed inside AI data centers.

Cerebras' stock rebounded by 9% after OpenAI CEO Sam Altman gave reassurance that the two companies are close partners days after a report cast doubt on their relationship, causing the chipmaker's stock to drop 20% last week. The episode raises concerns over its current concentration on a few large customers.