@tugot17 · Long NVDA · The Analysts

Nvidia moving beyond the cuda-moat to things that are hard to replicate through “Claude make no mistakes“

NVIDIA LPU supports 3 types of disaggregated inferencing: 1. Rubin Prefill + LPU Decode for the fastest interactivity 2. Rubin Prefill + Rubin Decode Attention + LPU Decode FFN for the middle of the curve 3. Rubin Prefill + Rubin Decode Verification + LPU Drafter for the

Read asLong NVDA

Since the call—

unrankedThe analyst's rank today

Medium · 0.86Confidence

Mon, Aug 31Tweeted

Tue, Sep 1 closeCounted from

not in the basket todayBasket

Rank per close

—

Sep 27

—

Sep 28

—

Sep 29

—

Sep 30

—

Oct 1

#156

Oct 4

#147

Oct 5

#113

Oct 6

#101

Oct 7

#146

Oct 8

#394

Oct 11

—

Oct 12

—

Oct 13

—

Oct 14