During an interview with CNBC, Nikesh Arora stated that token costs must decrease by as much as 90% to enable large-scale AI adoption. He noted that while a recent improvement in token efficiency by OpenAI's latest model is a positive step, further reductions are necessary.
Arora pointed out that rising token costs are straining AI budgets and making it challenging for businesses to implement AI tools effectively. This sentiment is echoed by other industry leaders, such as Palantir's CEO Alex Karp, who criticized the current token model and suggested that open-weight models could be a viable alternative.
The high costs are prompting companies to explore cheaper solutions, including models from Chinese firms. Meanwhile, AI spending is on the rise, with major tech companies like SpaceX and Amazon seeking substantial funding to support their AI initiatives.
Arora believes that as the market adjusts to these spending levels, budgets will eventually stabilize as technology becomes more efficient, but the demand for AI solutions remains strong