Google (NASDAQ:GOOG) (NASDAQ:GOOGL) unveiled its long-awaited Gemini large language model Wednesday
JP Morgan said a “significant innovation”
Analyst Doug Anmuth reiterated his overweight rating and $150 price target and noted that the three versions of Gemini —Ultra, Pro and Nano —will allow the LLM to do everything necessary, anywhere.
“The different sizes allow Gemini to be used efficiently across different use cases & hardware capabilities, from data centers to mobile devices, which should enable better cost management,” Anmuth wrote in an investor note. “Model sizes and cost concerns came up frequently in our industry discussions last week at AWS re:Invent, w/our view that there is increasing demand for somewhat smaller, more cost-efficient models, in addition to demand for a wide range of models to avoid lock-in.”
Anmuth said there is likely to be some pushback that Ultra will not be available until “early” 2024 and there are limited early use cases for Pro and Nano, along with “uncertainty” around monetization for search. However, he said he is encouraged that Gemini is going into consumer’s hands with Gemini Pro, which will power Google’s (GOOG) (GOOGL) Gen AI chatbot, Bard, which competes with ChatGPT, owned and operated by Microsoft (MSFT) backed OpenAI.
Anmuth also said the company addressed some concerns about the high cost of running Gen AI models via the three versions of Gemini, the newly announced TPU v5p chip and early use cases via Bard, search generative experience and the Pixel 8 Pro.
“While it remains early, the Gemini launch represents significant innovation for Google as we enter year 2 of commercialized and widely distributed availability of Generative AI,” Anmuth wrote.