Apple Caps AI Token Use At $200 Daily
Apple has set a daily AI token use cap, prompting other companies to consider similar measures to control costs. This move may influence other companies to follow suit.
Original reporting · digiday.com
Original: “The case for and against AI token capping”
Read “The case for and against AI token capping” at Digiday →How this was written. GMS Agency Desk wrote this article with a language model from Digiday’s reporting (model: @cf/meta/llama-3.3-70b-instruct-fp8-fast); it describes that reporting rather than adding to it. The original, linked above, is the full story.
Key points from Digiday’s report
- Apple has set a daily AI token use cap of $200 for engineers and may reduce non-engineers' token allowance to $25.
- Ad agencies are considering token capping to control AI costs, weighing the risk of limiting employees' creativity.
- AI costs are likely to rise as companies like Meta, OpenAI, and Anthropic seek a return on their investments.
Apple is placing a hard cap of $200 on its engineers' daily AI token use, and may reduce non-engineers' token allowance to $25, according to Digiday.
A handful of ad agencies have already implemented token capping this year, using AI software or bespoke tools to prevent users from consuming more than a set amount of AI compute.
Agencies considering capping token usage must weigh the risk of limiting their employees' creativity, as they build the infrastructure of AI accountability.
The case for AI token caps is that AI costs are likely as cheap as they will ever be, and companies like Meta, OpenAI, and Anthropic will want a return on their investments, making today's prices unstable for long.
Performance agency Roast has capped the amount of AI tokens its staffers can use since March, and the company allots a set amount of tokens to each employee within its Claude subscription.
U.S. media agency PMG places a $50 daily limit on staff AI usage, and staff that hit the cap more than once can be assigned a larger token budget or receive advice on using different AI models.
Tiered capping systems can provide teams with enough leeway to pilot something new without burning up the budget, and constraints can be useful for finding new ways of working.
However, some agencies argue that AI costs won't stay low forever, and intentionally cutting back before compute costs rise would be unwise, as they can develop the best uses for AI tech now while it's cheap.
Other agencies use soft AI caps or 'checkpoints' instead of hard limits, preserving staff autonomy while keeping a lid on AI spending, and some use gateway systems or 'gating' to reroute users to less powerful models.
Questions and answers
Why is Apple capping AI token use?
To control costs, as AI costs are likely to rise in the future.
How do ad agencies implement token capping?
Using AI software or bespoke tools to prevent users from consuming more than a set amount of AI compute.
What are the potential drawbacks of token capping?
Limiting employees' creativity and potentially hindering the development of new uses for AI tech.
