Researchers at Google published a paper revealing that a median Gemini text prompt consumes 0.24 watt-hours of energy, lower than previous estimates. This gives Alphabet an advantage in deploying AI due to cost efficiency. Despite concerns about AI search costs impacting margins, energy consumption has decreased by 33 times in the last year.

Alphabet’s economic moat is strengthened by its cost advantage and network effect. Despite a 40% increase in shares, they are considered undervalued with a fair value estimate of $237. Token usage has grown significantly while margins expand, credited to gains in model efficiency. However, more complex prompts may require more energy than traditional search.

Read more at Morningstar: Technical Paper on Energy Consumption Supports Our Cost Advantage Thesis