All-In Pod: Local AI Will Dominate? For both cost and data sovereignty, All-In says the future will be every employee with 20k-40k of compute on their desk and every employee will be given the freedom to tokenmax. Although I previously did not see local compute as a threat to Neoclouds, if employees are willing to spend 20k-40k of hardware budget for each employee and update it every four years then this is a serious case to consider. Note that 20k-40k of hardware will get individuals the $20-$40 plan worth of Anthropic/OpenAI equivalent performance. However it's low compared to salaries and the IP value of a company. Conversely, employees who have more than 40k of OpenAI/Anthropic budget will need 200k+ of hardware. These employees are likely working on critical IP, not just customer
View on X ↗