Loading page…
Loading page…
News stories updated
Nvidia reported $115.84 billion in fiscal third-quarter revenue, up 103% from a year earlier and 20% from the prior quarter, as demand for AI computing continued to defy mounting skepticism over infrastructure spending. Data-center revenue reached $107.6 billion, while GAAP net income rose to $65.3 billion and the company guided to approximately $130 billion of fourth-quarter revenue. The results substantially exceeded Wall Street expectations and sent Nvidia shares higher in after-hours trading. Yet the quarter also sharpens the central question facing the AI boom: strong chipmaker revenue does not necessarily prove that every company financing GPUs, data centers and cloud capacity will earn adequate returns. Read More...
Global fund managers made their sharpest monthly cut to technology exposure since 2006 as confidence in the AI investment boom weakened. Bank of America’s November survey found net technology overweight fell from 37% in October to 12% in November, while a record 38% of respondents said companies were overinvesting and 52% expected a major AI hyperscaler to cut capital spending within six months. Investors identified an AI bubble and a possible hyperscaler credit event as their leading tail risks, yet only 3% expected a global recession. The shift suggests skepticism is concentrating around the economics of AI infrastructure rather than the broader economy. Read More...
French optical-networking company Ekinops has launched PTM300, a compact data-center interconnect platform designed for rapidly rising AI and cloud bandwidth requirements. The 3RU system is rated for up to 32 terabits per second using as many as forty 800 Gb/s coherent interfaces, with manufacturer-claimed power efficiency of 0.08 watts per gigabit per second. Ekinops says PTM300 can operate over direct fiber, existing Ekinops optical systems or third-party infrastructure and supports open management standards. General availability is planned for December 2026, with first customer deployments expected in the first quarter of 2027. The announcement highlights how networking is becoming as important as compute in large-scale AI infrastructure. Read More...