Expired50% confidenceFactExact time
AI model parameter counts have been doubling every few months, forcing GPUs to increasingly read data from slower external standard memory.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
7/2/2026
First Seen
This claim has expired. It is kept for the record and no longer describes the current state.
Valid until: 9/30/2026(expired)
Sources
Marvell: The Hidden Giant Behind AI Infrastructure — Four Technical Aces Explained
bilibili中天评论6/6/2026
Related Claims
Unverified每个中间推理token与输出token消耗等量的GPU算力，Extra High档位单次请求FLOPs可能是普通生成模式的数十倍62% similarUnverifiedModels with larger parameter counts and more reasoning steps may consume 10–50 times the GPU compute of lightweight models.62% similarUnverified顶级模型每次调用的算力消耗可能是轻量模型的数十倍61% similarUnverifiedWhen AI clusters scale to tens of thousands of GPUs, overall efficiency can plummet to below 40%, with more than half of computing power consumed in data transmission and waiting.61% similarUnverified推理能力的引入将模型能力的翻倍时间从原来的约7个月缩短至3.5个月61% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/49984API
curl https://kongchang.com/api/v1/knowledge/claims/49984MCP
get_claim(id=49984)