AI token spend at some companies already exceeds or is approaching parity with engineering salary costs.

The steady-state AI spend will be $100k of AI tokens per engineer.

“I think it'll land at 100 grand per engineer equivalent. I think that's what we'll I think we'll give each of our best engineers $100,000 of tokens and in return we'll cut the size of our dev teams 30 40% effectively.”

For usage-based AI SaaS, LTV/CAC is insufficient; you must measure LTV to gross profit because token/usage costs are the largest cost item.

“LTV to CAC is usually not enough you want to probably look at like LTV gross profit specifically if you're you have token cost because you're the biggest cost item is usage at the end of the day.”

The most valuable tokens to spend are on agents where ROI per token can be directly measured, rather than on generic AI adoption.

“Tier three tokens the most valuable are this agents where you can get the ROI of each specific token and I can do that now.”

Companies will eventually spend $100-200 million on a single AI model query because the output of that query will be so valuable.

“He was like we will have companies spend 100 to 200 million on one model kind of result because that model result is so important like the output of one query.”

Decagon's token usage per conversation has increased over time because more model calls are added to improve quality, contrary to the assumption that optimization reduces token usage.

“Actually over time, the number of tokens we're using per conversation has gone up because we're actually doing more model calls to make the quality better.”

Spending $100,000 a year on tokens in 2026 gives you access to capabilities that will be mainstream and cheap by 2028, creating an 'alpha opportunity' to live in the future now.

“If you're willing to spend $100,000 a year right now in tokens, you are living the way somebody in 2028 is going to live because by then it'll be really cheap.”

AI spending currently represents approximately 3.8% of developer salaries.

“Which works out to 3.8% of developer salaries for the kind of you trillion dollar valuations to be justified you need to be in the 18 to 20% range.”

Tasklet's internal token spend is approximately 5-10% of its payroll.

“I would guess we're at about five like 5 to 10% of payroll right now in terms of internal token spent.”

Sierra's LLM spend is likely less than 10% of its revenues.

“My guess is their LLM spend is sub 10% of revenues use.”

Replit's cost to model providers is significantly less than the commonly cited 80% of revenue figure.

“For us it's not 80. It's it's it's way less than that.”