The cost of developing generative artificial intelligence (AI) that can power chatbots like OpenAI's ChatGPT has presented a new obstacle for Alphabet's Google. While ChatGPT-like AI requires a tremendous amount of computer power, it costs the organisation a lot of money. Inference in AI systems like ChatGPT requires billions of dollars' worth of chips, which is far more computational power than is typically required for search.
According to analysts quoted in a Reuters report, even if businesses like Google and Microsoft were able to make money from chat-based search ads, the technology could cost Alphabet several billion dollars in additional expenses.
The large language model AI will cost the business ten times more than a typical keyword search, according to Alphabet's Chairman John Hennessy.
Due to the required computing power, this type of AI costs more than traditional search. Chipsets costing billions of dollars are required to perform that computing. The price of electricity is also included.
Given the pressure to create AI-powered chatbots that can compete with Microsoft's Bing search engine, Alphabet's Google may find it difficult to afford ChatGPT-like AI.