DeepSeek has taken global AI markets by storm, erasing $1 trillion of value from major tech companies on Jan. 27.

Founded by Chinese hedge fund High-Flyer, the DeepSeek AI model cost $5.6 million to develop compared to more than $1 billion spent to create U.S.-based AI models like ChatGPT and Google’s Gemini, according to a Bank of America Global Research paper published Jan 27. 

Businesses and banks will evaluate the geopolitical risks and compliance challenges tied to DeepSeek’s Chinese origins, as well as its lower cost, Abhivyakti Sengar, senior analyst at consultancy Everest Group, told Bank Automation News. 

“In today’s AI landscape, innovation is only part of the equation. Trust, security and regulatory alignment are equally important for long-term success,” Sengar said. 

Chip-making behemoth Nvidia lost 16%, or $600 billion, of its market capitalization on Jan. 27 as investors reevaluated their AI positions in the equity markets with a cheaper alternative of DeepSeek being available, according to a Nasdaq report released Jan. 28. 

Security concerns 

While many banks and FIs might be skeptical about using OpenAI or Google’s AI models, they would be even less likely to use a solution from China, despite significantly lower cost, Jon Tvrdik, chief executive and founder at digital banking engagement service provider WaveCX, told BAN. 

Many banks and financial institutions might get a better return on investment by using DeepSeek AI model, but “the geopolitical context surrounding DeepSeek will make this a nonstarter at first for most FIs,” Tvrdik said. 

Banks handle highly sensitive data, and solutions from entities perceived as less secure will face significant regulatory scrutiny and data privacy fears, Tvrdik said.  

More competition 

This week, Chinese e-commerce giant Alibaba also released its own AI model called Alibaba Qwen, which the company has touted as better than DeepSeek or Meta’s Llama models, as reported by Bloomberg today. 

With the launch of new models, AI competition will heat up and push major providers to lower the costs of their offerings, the Bank of America research report stated. 

“Over the longer term, we expect advances at the model level to drive accelerated enterprise AI adoption and usage as chatbots, copilots and agents become simultaneously smarter and cheaper,” the report stated. 

Register here for Bank Automation Summit 2025, taking place March 3-4 in Nashville, Tenn. View the full event agenda here.