DeepSeek’s AI Model Challenges Industry Giants
China-based DeepSeek entered the generative artificial intelligence (GenAI) arena earlier this year, introducing a cost-effective, high-performance model that is disrupting the dominance of established players like OpenAI.
The emergence of DeepSeek’s R1 model, reportedly developed for around $6 million using less-advanced chips, has led experts to encourage tech entrepreneurs to prepare for a dynamic landscape in AI models. This advice is underscored by rapidly evolving costs and capabilities.
Shifting Market Dynamics
Since late 2022, a few AI assistants, including OpenAI’s ChatGPT, Anthropic’s Claude, and Google’s Gemini, have been at the forefront due to massive investments in engineering, data centers, and high-performance AI chips. DeepSeek’s entrance into the market suggests that GenAI assistants could become a regular commodity, driven by innovation and competitive market forces.
“The first company to train models must expend lots of resources to get there,” said CFRA senior equity analyst Angelo Zino. “The second mover can get there cheaper and more quickly.”
At the HumanX AI conference in Las Vegas, Hugging Face co-founder Thomas Wolf noted the declining cost of launching GenAI models. He also highlighted the growing significance of a multi-model approach.
“I feel like we are moving to this multi-model world, which is a good thing,” Wolf said, referencing the recent reception of the latest version of ChatGPT.
Maintaining an Edge
At the conference, OpenAI chief product officer Kevin Weil countered the notion that all models are created equal, implying a continued advantage for industry leaders.
“That’s actually not true,” Weil said. “The days of us having a 12-month lead are probably gone, but I think we have a three- to six-month lead, and that is really valuable.”
OpenAI plans to protect that edge. The company, based in San Francisco and with a user base of 400 million, leverages data from its extensive traffic to refine its models continuously.
“OpenAI has the Google advantage of being the thing that’s in everybody’s minds,” said Fen Zhao, research director at Alpha Edison equity firm.
Jeff Seibert, the chief of the accounting and AI start-up Digits, agreed that OpenAI would remain ahead. However, he expected the gap to close eventually.
“For advanced use cases, yes, there will be a lot of advantages,” Seibert said, commenting on OpenAI’s position. “But for a lot of stuff, it won’t matter as much.”
Seibert advises entrepreneurs to design their technology to accommodate switching GenAI models to remain flexible in a quickly changing industry.
Cost Reductions and Investment
Improved chip utilization and new optimization techniques have decreased the cost of designing the large language models (LLMs) that power ChatGPT, Gemini, and their rivals. Some LLMs are open-source, which helps speed up innovation by allowing anyone to modify and enhance the software.
According to Zino, the valuation of closed-model start-ups like Anthropic and OpenAI has likely reached a peak, as their “first-mover advantage dissipates.”
In February, SoftBank, the Japanese investment giant, invested $40 billion in OpenAI, valuing the start-up at $300 billion, nearly doubling its value from the prior year.
“If you’re burning a billion dollars a month, which I think OpenAI is, you have to keep raising money,” said Jai Das of Sapphire Ventures. “I have a hard time seeing how they get to a point where revenues are higher than the amount of cash they burn.”
In early March, Anthropic raised $3.5 billion, valuing the company, which prioritizes responsible AI, at $61.5 billion.
