Sam Altman calls RAM in home PCs an inefficient waste of resources
OpenAI head Sam Altman said that scarce RAM is more reasonably allocated to data centres rather than installed in home computers. As Kod.ru reports, citing Futurism, he made the remarks on the Cheeky Pint podcast.
Altman noted that many are betting on local inference—that is, running AI models directly on their own computer—and expect that with the emergence of open models, data centre capacity will no longer be needed. He himself does not share this approach. "From a global allocation perspective, it's pretty terrible to have a piece of the world's limited DRAM supply sitting in everyone's computer," the OpenAI chief said. As relayed by Habr, he also added that on user machines this memory is utilized only 5% of the time.
According to Altman, when chips are in short supply, equipment should be utilized to the maximum, and this can only be achieved in the cloud. He acknowledged that many people's intuition suggests the opposite.
The statement came amid rising memory prices: according to Kod.ru, manufacturers are dedicating more and more capacity to contracts with cloud companies and AI servers, while SK Hynix previously warned of shortages lasting until 2030. The author of the post on Habr noted that, according to his data, OpenAI ordered 40% of the world's DRAM output from Samsung and SK Hynix, after which contract memory prices rose by 171%.
Altman's remarks drew criticism among gamers and PC builders. Reddit users wrote that computers are purchased for gaming, video editing, and running software without a monthly subscription.
Sources
- Хабр: Сэм Альтман: «Оперативная память в вашем компьютере — напрасная трата дефицитных ресурсов» media outlet
- Kod.ru: Сэм Альтман назвал оперативную память в домашних ПК пустой тратой media outlet
This article was prepared with the help of AI and translated from the Russian original. Spotted an error — let us know.
Published 31 October 2026, 09:01 · No updates · Russian original




Comments on this story live under its post in our Telegram channel. Comment on Telegram