Nvidia's RTX Spark Superchip takes direct aim at Intel and AMD's client compute dominance, while TSMC confirms 15% price hikes on N3 nodes and U.S. printed circuit board supply collapses to 4% of global share. Six stories shaping the AI hardware infrastructure layer.
Audio is available on Spreaker — see link below.
Nvidia just moved back into the PC chip business in a serious way. At Computex, the company launched the RTX Spark Superchip, a single package combining a twenty-core Grace CPU with a six-thousand-one-hundred-and-forty-four-core Blackwell GPU, aimed squarely at Windows ARM laptops and desktops.
Intel isn't staying quiet. One of their senior executives publicly flagged what they're calling a healthy paranoia about Nvidia's Windows ARM entry.
While Computex dominated the headlines, the more structurally significant development may be what TSMC's CEO confirmed separately. Fifteen percent price hikes on N3 nodes are coming by late twenty-twenty-six.
Intel also disclosed more detail on Crescent Island, its data center GPU built on the Xe3P architecture. The headline spec is four-hundred-and-eighty gigabytes of memory.
AMD extended its AM5 socket commitment through twenty-twenty-nine, adding two more CPU generations to the platform. The more interesting detail is the five-thousand-eight-hundred-X-Three-D revival.
Senator Elizabeth Warren has called Nvidia CEO Jensen Huang to testify before the Senate on export controls and China business. The concern is direct: that Nvidia chips are reaching Chinese military applications.
Chapter summary auto-generated from the verified script. Listen to the full episode for the complete content.