- Analyst firm CCS Insight is predicting a “cold shower” for generative AI in 2024.
- The company’s chief analyst told CNBC that he believes the technology has been overhyped and faces huge costs in deploying it.
- The firm’s prediction comes amid fears of a global chip shortage.
Loading Something is loading.Thank you for reading this post, don't forget to subscribe!
Thanks for signing up!
Access your favorite topics in a personalized feed when you’re on the go. download app
One analyst firm is predicting a “cold shower” for generative AI in 2024.
London-based tech analytics firm CCS Insight expects that the real challenges of cost, risk and complexity in AI may soon overshadow the current hype about the technology in 2024, CNBC reported on Tuesday.
“There is so much hype around generic AI in 2023 that we think it is overhyped and will have to jump through a lot of hurdles to get to market,” Ben Wood, chief analyst at CCS Insight, told CNBC.
“The cost of deploying and maintaining generative AI is very high,” Wood said. He said that although large companies like Google and Meta can absorb these costs, it will be too expensive for many organizations.
For context, AI depends on chips to run, and the firm’s prediction comes amid concerns about a global chip shortage.
The importance of these chips has also ignited a global battle for Nvidia’s semiconductor chips among industry giants – ranging from Elon Musk to Chinese tech giants.
In response to rising demand, trillion-dollar chip maker Nvidia announced in August that it plans to triple production of its $40,000 processor chips next year.
And AI company OpenAI is now exploring plans to make its own chips, Reuters reported on Friday, citing recent internal discussions at the company.
For context, Dylan Patel, principal analyst at semiconductor research firm SemiAnalysis, estimated that the company spends up to $700,000 per day on the computing power needed to keep ChatGPT running.
CCS Insight did not immediately respond to a request for comment sent outside regular business hours.