TickerTain — Make Analysis Fun Again

←

▶ 28:06 · $NVDA · $GROQ · General Compute CEO: Groq architecture fundamentally flawed for large models, 80 hops per token

now playing · $NVDA · $GROQ

$AMD···bullish· mediumjason goodison

General Compute CEO: AMD inference deals with Anthropic, Meta, OpenAI make obvious sense on TCO

AMD is neck-and-neck with Nvidia on inference performance while offering lower total cost of ownership, making deals with Anthropic, Meta, and OpenAI economically rational.

$SAMBA-NOVAbullish· high· posjason goodison

General Compute bets company on SambaNova SN50: 5x faster inference, 10T params per rack, 20kW draw

SambaNova's SN50 chip enables 1-2,000 tokens/sec on frontier models with 10T parameters on a single air-cooled 20kW rack, delivering the speed and efficiency General Compute needs…

$CBRS···neutral· mediumjason goodison

General Compute CEO: Cerebras fast but model size limited, cost explodes at scale

Cerebras achieves exceptional inference speed but hits a wall on large model scaling where costs become prohibitive, making it unsuitable for General Compute's frontier model targ…

$NVDA···bearish· highjason goodison

General Compute CEO rejects Nvidia for inference: margins too high, architecture wrong for fast tokens

Nvidia's healthy margins and circular financing deals subsidize demand but its Vera Rubin/Groq architecture requires 80 back-and-forth trips per token for large models, making it…

$GROQbearish· highjason goodison

General Compute CEO: Groq architecture fundamentally flawed for large models, 80 hops per token

Groq's attention-FFN disaggregation forces 80 round-trips per token on 80-layer models, creating insurmountable latency for large model inference despite fast single-chip performa…

←