Now Playing
Samsung's Price Hike, TSMC's $60B Bet & Intel's Equity Test
Audio is available on Spreaker — see link below.
Samsung's foundry price surge signals TSMC capacity is fully booked — and AI demand has outrun the entire advanced node supply chain. Today's briefing covers the cost cascade hitting GPU supply, Intel's dilutive $20B offering, and AMD's inference push.
Audio is available on Spreaker — see link below.
Samsung just raised its foundry prices by ten to fifteen percent. That's the real headline in semiconductor capacity right now, and the reason it matters has almost nothing to do with Samsung itself.
TSMC's own capital expenditure guidance reinforces exactly that read. The company raised its twenty twenty-six spending commitment to somewhere between sixty and sixty-four billion dollars, and paired it with a dividend increase.
Cut to Intel, and the picture is more complicated. The company closed a twenty-billion-dollar stock offering at ninety-five dollars per share, with twenty-three billion dollars in institutional demand.
Downstream from the foundry tier, the cost pressure is cascading in ways that are becoming visible in shipment data. Entry-level GPU supply is tightening sharply.
On the server side, Gigabyte launched its R165 series, a one-unit server built around AMD's EPYC Venice processors, with PCIe six-point-zero slots targeting AI inference and cloud workloads. The launch itself is relatively contained, but what it represents is worth noting.
The metrics worth tracking from here are specific. Watch whether Samsung's price increases hold or whether yield issues at SF4 and SF5 give customers a reason to wait for TSMC capacity.
Chapter summary auto-generated from the verified script. Listen to the full episode for the complete content.