DeepSeek, a China-based company, disrupted the world of generative artificial intelligence (GenAI) earlier this year with its R1 model, which offered impressive performance at a fraction of the cost of its competitors.
The model, priced at approximately $6 million, raised eyebrows in the industry, particularly in light of the dominance held by major players such as OpenAI, Anthropic, and Google since late 2022.
Generative AI assistants like OpenAI's ChatGPT, Anthropic's Claude, and Google's Gemini had long been seen as the frontrunners in the sector, benefiting from multi-billion-dollar investments in engineers, data centres, and cutting-edge AI chips. However, DeepSeek's R1 model, powered by less-advanced chips, proved that high-performance AI could be delivered at a significantly lower cost.
While some specialists have raised doubts about the true cost of the R1 model, its introduction sparked discussions around the future of GenAI assistants. There were growing suggestions that the market for AI models was becoming more competitive and that the technology might soon be viewed as a commodity, driven by innovation and market dynamics.
"The first company to train models must expend lots of resources to get there," said Angelo Zino, senior equity analyst at CFRA. "The second mover can get there cheaper and more quickly," he added, highlighting the advantage that newcomers like DeepSeek could have over established giants.
At the HumanX AI conference in Las Vegas earlier this week, Thomas Wolf, co-founder of Hugging Face, echoed the sentiment that launching GenAI models had become less costly. "I feel like we are moving to this multi-model world, which is a good thing," Wolf remarked, referencing the relatively muted reception of the latest version of ChatGPT.
However, Kevin Weil, chief product officer at OpenAI, challenged the idea that all models were now equal. "That's actually not true," Weil asserted. "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, with 400 million users, remains in a strong position, as Weil explained, due to the ability to use massive traffic data to continually enhance its models.
Fen Zhao, research director at Alpha Edison equity firm, emphasised that OpenAI had an edge due to its popularity and recognition. "OpenAI has the Google advantage of being the thing that's in everybody's minds," Zhao noted.
Despite the competition, Jeff Seibert, CEO of the accounting and AI start-up Digits, predicted that OpenAI would remain at the forefront of advanced use cases, though he acknowledged that for many other applications, the difference between competitors would be less significant. "But for a lot of stuff, it won't matter as much," Seibert added.
Seibert advised entrepreneurs to design their technology in a way that allowed them to switch between different GenAI models, offering flexibility in the rapidly evolving industry.
Meanwhile, improvements in chip technology and new optimisation techniques have contributed to lowering the cost of developing large language models (LLMs) like those behind ChatGPT and Gemini. An open-source approach, adopted by some LLMs, has accelerated innovation by enabling anyone to experiment with and improve the software.
Zino, the senior equity analyst at CFRA, suggested that the valuations of closed-model startups such as Anthropic and OpenAI had likely peaked, with their first-mover advantage fading. OpenAI's valuation had surged to $300 billion in February after receiving a $40 billion investment from SoftBank, but Zino warned that "as the first-mover advantage dissipates, their valuation may face pressure."
Jai Das, managing partner at venture capital firm Sapphire Ventures, questioned how OpenAI could sustain its growth given its significant cash burn. "If you're burning a billion dollars a month, which I think OpenAI is, you have to keep raising money," Das said. "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, bringing its valuation to $61.5 billion, underlining the fierce competition in the AI sector and the intense race for investment.
