$GLW having a presentation on optics. This slide.. Everywhere you see yellow you see Corning.
"And when we add the optical scale up network to interconnect all the switches in every GPU rack, well we add a lot more yellow."
🔥 Amid reports that #Foxconn# may have accelerated #CPO# rack deliveries to #NVIDIA#, the company said at its earnings call that silicon photonics CPO switches remain on track for Q3 mass production, with 2026 shipments expected to hit 10,000 units.💡More: 🔗
👀 #TSMC#'s COUPE on Substrate tech is expected to hit mass production in 2H26 — and if #CPO# gains traction in AI servers, #NVIDIA# may look to lock in high-end substrate capacity early via long-term deals. 💡More: 🔗
Next year… I’m expecting there to be many articles about FAU + component bottlenecks.
Especially as the new CPO architecture led by $NVDA + $TSM starts to scale.
Then a lot of these names like FOCI (~$2.8B MC) or Nextronics (~$246M MC) that I’m mentioning today will be in the center of it.
Despite many of these “commodity” labels… (just look at transformers/NAND)
And I’ll do a “Did you listen anon post” like $AXTI.
We’ll see if this is right.
Miami Outtakes: Everyone is Optimistic About Clarity, Agentic AI and Tokenization with Hyperliquid Policy CPO @adam_minehardt, Centrifuge Co-Founder @itsbhaji & Anchorage Digital CEO @nathanmccauley
🖥️ NVIDIA Rubin: Four Simultaneous Hardware Overhauls
Liquid cooling shifts from premium to baseline. Power architecture moves to 800V HVDC. PCBs scale to 20+ layer HDI with cable-less designs. Copper yields to CPO at the interconnect. Every layer of the AI server stack is changing at once.
Inside the Rubin-era hardware and supply chain transition 👉
#TrendForce# #SelectedTopics#