Gilad Shainer
SVP of Networking · NVIDIA
Position Evolution
3 tracked across this operator's appearancesSame operator, on the record, on the same topic, at different points in time. Each delta below is anchored to verbatim transcript spans verified against source — no paraphrases. This is the alumni-graph moat: SemiAnalysis cannot reproduce this query because they don't have the speaker-stable corpus.
Data center as single computing unit
Hardenedconfidence 88%Shainer consistently frames the data center as a unified supercomputer, but in the later appearance he grounds this with far more technical depth—explaining jitter, synchronization, and scale-up vs. scale-out distinctions. The conviction is the same but the stakes and specificity are higher.
"you have a scale- out computing infrastructure that connects all those racks together and scale up the data center and making the data center a single unit of computing."
Source on theCUBE ↗"the compute engine is the entire data center. And therefore you need to have a way to take those GPU ASICs and connect them together so they can exchange information between themselves and to form a single unit of computing out of those ASICs"
Source on theCUBE ↗Co-packaged optics for scale-out efficiency
New positionconfidence 85%In the earliest appearance Shainer speaks only broadly about bringing networking engines into enterprise data centers with no mention of co-packaged optics. By the latest appearance, co-packaged optics has become a central, named technical position with quantified energy and resilience benefits.
"The key thing is that we want to get those networking AI engines inside, and we want to make it in the easiest way for enterprises to consume that inside their data centers."
Source on theCUBE ↗"Co- packaged optics means that we take the optical engine that sits outside of the box and put it in the same package as the switch. So now the light goes all the way to the switch package and then from that point it's very close-"
Source on theCUBE ↗Reference architecture enabling fast deployment
Shiftedconfidence 78%Earlier, Shainer emphasizes reference architectures as the practical path to fast, accessible AI data center deployment. Later, the framing shifts to 'co-design' as the governing philosophy, moving away from prescriptive reference architectures toward a deeper engineering principle. The audience implication changes from IT simplicity to architectural discipline.
"if you're following that, you can build that new AI data center in weeks, and you don't need to have huge amount or a huge group of ITs to cover all those elements"
Source on theCUBE ↗"A supercomputer is not a collection of elements. It's a co- design that covers everything from the software to the hardware to build a single unit of computing."
Source on theCUBE ↗All theCUBE appearances (5)
theCUBE + NYSE Wired: AI Factories - Data Centers of the Future | Gilad Shainer, NVIDIA
GUEST · NVIDIA · SVP of Networking
theCUBE + NYSE Wired: AI Factories - Data Centers of the Future | Gilad Shainer, NVIDIA
GUEST · NVIDIA · SVP of Networking
theCUBE + NYSE Wired: The AI Factory - Data Center of the Future | Gilad Shainer, NVIDIA
GUEST · NVIDIA · SVP of Networking
theCUBE + NYSE Wired: The AI Factory - Data Center of the Future | Gilad Shainer, NVIDIA
GUEST · NVIDIA · SVP of Networking
NVIDIA GTC Washington, D.C. 2025 | Gilad Shainer, NVIDIA & Will Eatherton, Cisco
GUEST · NVIDIA · SVP of Networking