Is it time to panic or panic? Wall Street Weighs in on How DeepSeek Will Shake Up the AI Business
AI stocks fell on Monday as the investment community latched onto news of a potentially more cost-effective Chinese AI model — and the West began to question the gargantuan spending on AI infrastructure.
Chinese AI startup DeepSeek last week released a new generative AI model called R1 aimed as a competitor to OpenAI. Analysts have cited DeepSeek’s models as more cost-effective: DeepSeek recently revealed that it spent just $5.6 million to train its latest model, V3, while OpenAI did. More than 100 million dollars To train the GPT-4 model.
NVDA fell as much as 17% on Monday, shaving hundreds of billions from the AI chip giant’s market cap and leaving short sellers with more than $5 billion, according to S3 Partners data.
Meanwhile, rival Advanced Micro Devices ( AMD ) fell 6 percent, and Broadcom ( AVGO ) fell more than 18 percent. The broader markets were also hurt: The S&P 500 (^GSPC) was down 1.9% Monday morning, while the tech-heavy Nasdaq (^IXIC) was down 3.4%.
D-Davidson analyst Gil Luria told Yahoo Finance in an email that the performance of DeepSeek models developed in China could have a significant impact on US AI companies. “Models from DeepSeek are performing at a fraction of the cost of the most advanced models built in the US.”
“This means US hyperscalers such as Microsoft ( MSFT ), Amazon ( AMZN ), Google ( GOOG ) and others may be investing heavily in data center construction,” he added. “If companies can run AI models with minimal computation, they won’t need data centers with hundreds of thousands of NVDA GPUs.”
Raymond James analyst Srini Pajjuri echoed those sentiments in a note late Sunday.
“DeepSeek clearly doesn’t have access to as much compute as US hyperscalers and has somehow managed to develop a model that looks very competitive,” Pajjuri wrote.
US President Donald Trump announced last week that Stargate, a venture backed by OpenAI, SoftBank and Oracle (ORCL), will invest $100 billion in US AI infrastructure (i.e. data centers) immediately and an additional $400 billion over the next four years. . Shortly after that announcement, Meta ( Meta ) said it would increase its capital spending to $65 billion by 2025.
CFRA analyst Angelo Zino said Monday: “We think investors should take innovation from China seriously, as the current capex spending/pace of technology upgrades calls into question the need.”
“We think increased risks may cause multiples for major chipmakers to be compressed, with the most vulnerable being NVDA, MRVL, AVGO, AMD and MU,” he said.