Customer specification changes are pushing Lite-On’s updated 800V power rack toward mass production at the end of 2027, DIGITIMES reports. Lite-On (2301 TT), a Taiwanese AI server power supplier, says cloud customers are revising cooling and electrical protection requirements. The changes call for larger power modules and transformers, along with adjustments to the shelf layout for battery backup units. Small-volume shipments for customer power testing are still planned for Q4 2026. The updated specifications are expected to be finalized by the end of Q1 2027, with a larger production ramp in 2028. Lite-On still sees AI server power orders through 2028.

JUST IN: Samsung is reportedly targeting an exit from its 8-inch foundry business by the end of 2029, a year earlier than the previously reported 2030 timeline, according to Korea’s Dealsite, cited by DIGITIMES. DIGITIMES also reports that Samsung has outsourced some production to foundries including $UMC as it shifts resources toward 12-inch manufacturing. Moving existing products to another foundry can take around 6–12 months, industry sources cited in the report estimate, as process conditions need adjustment and customers must qualify the new production source. With global 8-inch capacity shrinking, some customers may be ordering more from Samsung now to build inventory ahead of the transition. We think that could pull purchases forward, making stronger near-term orders an imperfect measure of underlying demand.

Bing Bong is bringing the energy to London as @phantom powers the After Dark experience at Breakpoint. Solana culture is taking over Soho, and the full clip captures the excitement.

JUST IN: Micron’s $MU Taoyuan union says its push for a profit-sharing system has been unsuccessful and will skip labor talks scheduled for Nov. 17. In a Nov. 12 open letter, the union said it would instead hold a protest press conference at Taiwan’s ruling party headquarters that day. It also plans to continue organizing a union across Micron’s Taiwan sites. The dispute remains unresolved. Today’s announcement provides no confirmation that work on the reportedly paused Taoyuan fab expansion has resumed.

Google’s TPU v9 could use fewer PCB layers than Amazon’s Trainium 4 while requiring more HDI build-up steps. We think total layer count alone would understate the manufacturing work involved. Taiwan sell-side research points to a 3-28-3 HDI design for Trainium 4 in Q3 2027 and a 6-12-6 design for TPU v9 in Q4 2027. That works out to 34 total layers for Trainium 4 and 24 for TPU v9. But TPU v9 would have six HDI build-up layers on each side of its inner stack, versus three for Trainium 4. Those outer layers are added through sequential lamination, with laser-drilled microvias connecting them. More build-up layers add processing steps and place greater demands on alignment and interconnect reliability. For AI PCB suppliers, the two designs would create different production requirements. High-layer-count capability and the ability to reliably manufacture deeper HDI structures both matter as cloud providers develop their next custom chips. $GOOG $AMZN

My limit order is set for $NVDA $130 is an insane entry price and i think its a pretty good possibility that it gets down there I am buying the dip. But send it down more first