High Yield

@chipwise_tech How could imec do it without me?

@highyieldYT Are you helping installing the new EUV machine?😊

Hmm 🧐

@mualphaxi @_dedede223 Thanks, wasn’t visible for me on the dark background. Found it now and already ordered!

@highyieldYT @_dedede223 Pink link here

I have been dreaming of this day for a long time. Arena is now a book publisher, and our first volume, "Silicon" is open for preorders. It's quite unlike anything you've seen: a coffee table book capturing the ecstatic beauty of silicon technology. arenamag.com/silicon

@mualphaxi @Logic_Engine @_dedede223 I was in Belgium today, visiting imec in Leuven. Could have picked one up ;)

@Logic_Engine @_dedede223 @highyieldYT They are printed in Belgium but shipped from Texas. Globalism! Some day we'll be able to afford a European fulfillment operation but can't right now. Sorry...

@mualphaxi @_dedede223 Still can only select US or Canada as shipping option.

@_dedede223 @highyieldYT Okay we're changing. EU preorders now available, too!! Same page - $30 shipping charge.

@dylan522p I want a signed issue

I'm the author of the first chapter in the book, "Teaching Sand to Think" Julia kicked ass with this book, it's visually stunning and has so many awesome contributors

@BasicAppleGuy Would be cool to use actual die shots

A18 Pro, M5 Pro, & M5 Max Silicon Inside Stickers Now Available. basicappleguy.com/siliconinside

I spotted a next-gen ASML scanner in the imec clean room 👇

It’s like a ghost town in here 👻

@lithos_graphein @dwarkesh_sp I’ll get so see a 5200 tomorrow. Can’t wait!

@dwarkesh_sp i've been waiting for this day my whole life 🥲

EUV machines are the most complicated tools humans make. Their supply chain has over 10,000 individual suppliers, and any one of them not scaling fast enough can bottleneck the entire AI industry. An EUV tool fires lasers at a tiny tin droplet three times in precise sequence, blasting it hard enough to emit EUV light. That light bounces off 18 multilayer mirrors onto the wafer. Meanwhile, the two platforms inside the machine - one holding the stencil, one holding the chip - are flying back and forth at 9Gs in opposite directions. The successive passes have to land on top of each other to within 3 nanometers. If any part of this is off, yield goes to zero. Take just one component. The mirrors are mostly supplied by Carl Zeiss, who have probably fewer than a thousand people working on them. In turn, Carl Zeiss rely on machines from Switzerland to deposit each of the layers, and use a coating process co-developed with a different German company. None of these companies have woken up. They’re gradually increasing production, but nowhere near the levels necessary for what the labs want by the end of the decade. @dylan522p predicts production can't scale beyond about 100 EUV machines per year by 2030, no matter how much money gets thrown at the problem. In the medium term this is the key bottleneck on scaling.

Nvidia DLSS 5 off DLSS 5 on

@HyperTechInvest Did they say they are supplying for NVIDIA?

$MU enters high-volume production of HBM4 for NVIDIA Vera Rubin Micron said it has started volume shipments of its 36GB 12-high HBM4 in the first quarter of 2026, and the product is designed for NVIDIA Vera Rubin systems The company says the chip delivers over 11 Gb/s pin speed, more than 2.8 TB/s of bandwidth, around 2.3x the bandwidth of HBM3E, and over 20% better power efficiency Micron also said it has shipped samples of its 48GB 16-high HBM4, which increases capacity per HBM placement by 33% versus the 36GB 12-high version Beyond HBM, Micron said its 192GB SOCAMM2 is now in high-volume production. It is designed for NVIDIA Vera Rubin NVL72 systems and standalone NVIDIA Vera CPU platforms, enabling up to 2TB of memory and 1.2 TB/s of bandwidth per CPU Micron also said it is the first company to mass-produce a PCIe Gen6 data center SSD. Its Micron 9650 is built for AI training and inference workloads, supports up to 28 GB/s sequential read throughput and 5.5 million random read IOPS, and is optimized for NVIDIA BlueField-4 STX architecture. The company says it offers up to 2x the read performance of Gen5 and 100% higher performance per watt

@IanCutress I would have invited you Ian

The benefits of being invited under NVIDIA's analyst program: access to the chip. :(

@farajalrashidi I actually did the raw edit for the entire video like that. It only works well with the Speed Editor. Otherwise I wouldn’t even try.

@highyieldYT Most performative set up I’ve seen so far

This is what peak productivity looks like 🚀

@vctregea @tgod34748 @Silver_x86 It’s super thin though.

@highyieldYT @tgod34748 @Silver_x86 well the heatsink if almost the size of the logic board so yeah i think it’s pretty big for a chip with that TDP and provides the cooling it needs, that’s why the difference in performance during load with the m1 mbp was minimal

Just edited a long-form video on the MacBook Neo. The full Final Cut Pro for Mac running on an iPhone chip isn’t the best. Presumably due to the media engines. My base M1 Air is a lot faster in Final Cut. But surprisingly, this wasn’t my biggest hurdle while editing (cont.)

@vctregea @tgod34748 @Silver_x86 I wouldn't call it big (I have a M1 logic board right next to me), but yes, it does have a heatsink. And even more important, it doesn't stack the memory on top of the SoC.

@highyieldYT @tgod34748 @Silver_x86 the M1 Air, although it has no fans, it has a pretty big heatsink on top of the chip that dissipates it’s heat

@vctregea @tgod34748 @Silver_x86 The A18 Pro has a media engine.

@tgod34748 @Silver_x86 both the M1 and the A18 Pro do not have media engines

@tgod34748 @Silver_x86 Afaik, the A18 Pro does have a media engine while the base M1 does not.

@Silver_x86 Not at all. My M1 MacBook Air with 8GB of RAM handles much more intensive projects perfectly fine. This is something else in the A18 Pro that’s limiting it, likely being the media engines.