AI Assistants Now Cost More Than a Human Employee—Why?

• AI has become very expensive. The best AI assistant from OpenAI can set you back a whopping $20,000 per month. 

• AI promises to save time and labour by a great deal. But in the very cases of costly blasts, the primary mechanism to sum up the situation becomes pure demand, not strong and focused operational costs. It is a point of being free beings for any organisation to judge its stringency concerning this notion.

But AI was supposed to make life easier and cheaper, right? It turns out not. Instead of relieving businesses of costs, AI has become one of the most overpriced services around. Overpriced to the point that some sneering from seasoned Wall Street guys might not be out of place.

AI at $20,000/month? Seriously?

Yes, it’s no joke. OpenAI, the company behind GPT, may launch a series of AI “agents” against the price of your daily coffee.

Here’s how the prices of such agents are comparatively high:

• A general AI assistant is up for $2,000/month.
• Need a coding assistant? It’s going to be $10,000/month!
• And the junior PhD researcher is yours for a pitiful sum of $20,000/month.

At present prices, ChatGPT Pro is being sold for $200 a month, and OpenAI is losing money with this amount.
It is a very 20th-century picture through and through! Still, fascinatingly enough, some very big players are getting their hands on this. SoftBank, for instance, earlier this year pledged all of $3 billion toward AI agents at OpenAI, gaining sheer confidence in the opinion that these agents can perform as any highly skilled professional could.

Why Does AI Cost So Wildly Differently Right Now?

This is what it feels like to get your wits around pricing AI right now—it’s like getting into the Wild West: rolling dice.

Altogether, AI is a minor cost to augment an already existing product, with secret price hikes. By contrast, some providers set prices on a tasks-completed basis.

The price of the AI coding assistant from Cognition, called Devin, is only $500 a month, which is a small percentage of OpenAI’s $10,000 price for its own AI typing assistant. What is causing the immense price discord?

One reason is simply the high cost of maintaining them.

• A data centre underpinning an AI algorithm may cost somewhere between $500,000 and more than $1 billion a year.
• Specialised AI followers can be priced from $10,000 to $30,000 apiece.
• A single server rack stocked with these chips will cost you an arm and a leg, typically over $500,000, before you even switch the power on.

At a glance, you know that AI is not exactly cost-effective to develop, but do the eye-popping price tags fit the bill?

Is It Worth the Money?

That is the question. $20k worth of value for any AI assistant being handled by businesses—is it there?

AI’s proponents say that this very object changes the game: it can deliver grinding functionality for a business 24 hours a day, seven days per week, never taking a holiday, requiring no benefits, and processing any amount of information at lightning speed beyond the reach of any human.

Nevertheless, let’s face it: AI is nowhere near perfect.

• It makes errors.
• It hallucinates the facts sometimes.
• In an industry like finance or healthcare, where accuracy is non-negotiable, anything more than that can lead to potential risk by relying overly on AI.

Is an AI model developed by a PhD the actual replacement for human experts? This question remains open-ended.

The Key Takeaway 

AI is moving quickly, and businesses want to cash in. While, of course, there are explicit costs, a great part of this reflects far more on “what the market will bear” rather than what it takes to deliver the service.

But the finding by PYMNTS Intelligence is that businesses are making money off AI investment, and so they see a rationale going forward for the next many months. So, a premium is acceptable.

Will these prices hold up over time? Likely no. If the history of how technologies evolve is anything to go by, as competition runs its course and technology advances, pricing may be expected to stabilise downwards in AI.

Until then, might it be worth thinking twice before writing monthly checks worth tens of thousands towards these AI tools? Is it a breakthrough technology or an overhyped fad? Only time will tell.