Unverified50% confidenceFactExact time
The cost of training frontier models runs into the hundreds of millions of dollars and depreciation cycles are short, so such companies often exhibit the characteristics of high revenue growth coexisting with high losses when they go public.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
10/27/2026
First Seen
Valid until: 1/25/2027
Sources
Related Entities
Related Claims
UnverifiedPausing large model training is costly, involving enormous compute expenditures and potentially affecting the pace of product iteration and competitive market position.75% similarUnverifiedTraining a frontier model can cost tens of millions to hundreds of millions of dollars, and once its API outputs are harvested at scale, competitors can train substitutes with comparable capabilities at a cost far below the training cost, directly eroding the original model vendor's commercial moat.74% similarUnverifiedThe cost structure of large-model companies is characterized by "front-loaded heavy assets": the GPU compute, data center, and electricity expenses required for training and inference occur before revenue, and they rise rapidly as model scale expands.74% similarUnverifiedIn the cost structure of large model companies, training costs occur before the model is released and are a one-time but enormous computing investment; inference costs, by contrast, occur continuously with user call volume, and the more users there are and the more frequent the calls, the greater the expenditure.73% similarUnverified训练2.8万亿参数模型代价高昂，便宜和开源并不等于不需要算力，算力需求从闭源巨头向分散的企业定制场景转移73% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1103520API
curl https://kongchang.com/api/v1/knowledge/claims/1103520MCP
get_claim(id=1103520)