The AI arms race just experienced perhaps its biggest shock yet. Enter DeepSeek, a Chinese AI startup shaking up the industry by offering a high-performance LLM at a fraction of the cost.
Wall Street wasn't ready- Nvidia's stock plunged nearly $600 billion, fearing what this disruption could mean for U.S. tech dominance. DeepSeek's latest V3 model was reportedly trained for just $5.5M, compared to the massive $100M needed for ChatGPT.
At the same time, AI's energy consumption is under scrutiny. Today, one day of global ChatGPT usage could rival the energy needs of small cities. Even Sam Altman, CEO of OpenAI, said in March 2024, "I think we still don't appreciate the energy needs of this (AI) technology". Yet DeepSeek claims its simpler processes and modular scaling drastically reduce costs and energy requirements - it claims to be 40% more efficient than ChatGPT.
But this disruption comes at an awkward moment. Just days into his second term, US President Donald Trump launched The Stargate Project, a $500 billion AI infrastructure investment to keep America ahead in the AI race. With SoftBank, Oracle, and OpenAI backing the initiative, the U.S. is doubling down on AI supremacy. DeepSeek's training model costs put a big question mark on the investment required to reach 'the future of AI'.
Trump himself acknowledged DeepSeek's rise, calling it "a wake-up call" for the U.S.
Is this the start of a significant AI power shift? The world is (deep)seeking answers.