AI Pricing Trends in 2025: Lessons from OpenAI's CEO

If you follow tech news and more specifically OpenAI's updates, you probably came across Sam Altman’s now-famous pricing-related tweet:

OpenAI Pro, priced at $200/month (10x the Plus plan), has seen such intense adoption that it's reportedly costing OpenAI more than it earns.

As I shared on LinkedIn, it’s unclear if this is entirely true or a clever marketing move to emphasize high usage and encourage Pro plan upgrades.

Either way, it’s a timely reminder of the pricing and packaging challenges and changes SaaS companies, especially in AI, are grappling with today.

AI Use Cases Are Expanding Rapidly

AI is undeniably "eating the world" with tools like ChatGPT, Claude, and Perplexity making waves across both B2C and B2B domains.

For instance, consumers use Perplexity for information searches (vs. Google) or ChatGPT for translation (vs. DeepL), while businesses adopt AI-based SaaS solutions.

As AI usage grows, customers increasingly expect these capabilities in their B2B software.

SaaS companies have no choice but to integrate and adapt.

And the ecosystem around AI pricing is adapting with new billing software emerging.

People Are Willing to Pay for AI-Based Use Cases

At $200/month, OpenAI Pro isn’t cheap, but users find enough value to justify the cost.

The return on investment drives this willingness to pay.

However, there may be a bias: higher prices can motivate users to maximize their usage, optimizing their expense.

This raises an important point for SaaS pricing strategy: lower-priced plans may lead to underuse, reducing perceived value and increasing churn.

This, in turn, limits the feedback needed to refine your product and achieve PMF.

AI Brings High Costs, Requiring Smarter Pricing Models

Server costs are no longer the sole concern.

SaaS businesses using AI must now factor in the cost of each API call to the LLM.

This shift has driven the rise of outcome-based and usage-based pricing models, as traditional license-based or fixed pricing doesn’t safeguard margins.

With fixed pricing, some customers may cost you more than they pay, while others generate profit, making cost control unpredictable.

Pricing Is too often defined arbitrarily

Sam Altman admitted OpenAI’s pricing was set arbitrarily.

While pricing always involves uncertainty, several methods—like user interviews, willingness-to-pay surveys, data analysis, and competitor research—can help establish a reasonable range.

Though their case is unique as pioneers of this offering, it’s a reminder that leveraging data points is key to informed pricing decisions.

Setting-up pricing in your product with a flexible billing architecture

Adapting your pricing over time is inevitable, especially as you monetize new features.

To make these changes efficiently and without heavy resource involvement, a flexible pricing system is crucial.

This often means decoupling pricing logic from your code, allowing non-technical teams to implement updates.

It also simplifies the process of grandfathering existing customers into your new pricing strategy.