Nvidia has faced a significant hit due to the emergence of DeepSeek’s latest reasoning model, which rivals the best American tech and costs much less. Several smaller AI chip companies have seen an uptick in demand as investors flock to the Chinese startup.
DeepSeek’s R1 model is an open-source alternative to proprietary models like OpenAI’s, making it attractive to developers who want affordable solutions. CEO Andrew Feldman of Cerebras Systems said the release of R1 generated a record spike in demand for its services, highlighting the growing opportunity for smaller companies to scale with DeepSeek.
The company claims its model rivals the best American tech despite running on lower costs and being trained without cutting-edge graphic processing units. However, industry experts have questioned these assertions.
DeepSeek’s innovation lies in accelerating the AI cycle from training to “inference” phase, where AI is applied to make predictions or decisions based on new information. This shift has sparked interest among smaller chip startups, who see room for expansion in the “inference” segment.
Analysts agree that DeepSeek’s accomplishments are a boost for AI inference and the wider industry. The company’s performance is attributed to engineering innovations that reduce inference costs while improving training costs. This trend explains Jevon’s Paradox, where cost reductions drive increased demand.
As the demand for AI continues to grow, smaller players will have more room to expand. Sunny Madra of Groq said Nvidia’s limited supply of chips creates opportunities for smaller companies to sell into the market aggressively. The emergence of DeepSeek’s R1 model has sparked a significant shift in the AI chip industry, with smaller companies poised to benefit from the growing demand for affordable solutions.
Source: https://www.cnbc.com/2025/02/07/deepseek-force-multiplier-for-smaller-ai-chip-firms-.html