Marian Sorca’s Prediction: AI Pricing for Individuals Will Rise Fast

I think AI pricing for individuals will rise fast over the next 6 to 12 months.

Possibly by 100%.

That may sound aggressive.

But when I look at the current market, the cost of running advanced AI systems, the pressure to become profitable, and the growing competition between major labs, I think today’s entry pricing may not last.

Why I Am Thinking About This Now

Part of what pushed me to think more seriously about this was OpenAI saying goodbye to the Sora app.

Sora became one of the consumer-facing examples of how impressive modern AI had become. It was the kind of product that turned heads fast, especially when viral videos started spreading across social media.

But viral attention and a cool demo are not the same thing as a sustainable business.

When a company pulls back from a product like that, I think it is worth asking a bigger question:

How long can AI companies keep offering very powerful tools to individuals at relatively low monthly prices?

OpenAI’s official pricing currently shows ChatGPT Go at $8/month, Plus at $20/month, and Pro at $200/month. (OpenAI)

That middle tier - especially Plus at $20/month - increasingly feels underpriced to me for what users now get from these systems. (OpenAI)

AI Is Expensive in a Way Many People Still Underestimate

A lot of people still think about AI like software in the traditional SaaS sense.

You build the product, you scale infrastructure, and you improve margins over time.

That logic still matters, but frontier AI is different.

These products are expensive to train, expensive to serve, expensive to improve, and expensive to support at a quality level that keeps users impressed.

And unlike many traditional apps, the cost does not disappear once the product ships.

The usage itself is part of the cost structure.

The more people rely on AI, the more compute gets consumed.

The more advanced the models become, the more the economics get pushed.

That is why I think the current consumer pricing environment may be temporary.

Why $20/Month May Look Cheap in Hindsight

If a product becomes central to your work and personal life, the price ceiling changes.

That is the key point.

AI is no longer just a curiosity.

For many people, it is already becoming part of how they write, research, code, plan, analyze, study, create, and make decisions.

Once a tool moves from “nice to have” to “must have,” the company behind it gains more room to rethink pricing.

That does not automatically mean unlimited pricing power.

Competition still matters a lot.

But it does mean that users may tolerate more than they did a year ago, especially if AI continues to save meaningful time every week.

So when I look at $20/month for ChatGPT Plus, I do not see a stable long-term number.

I see something that may later look like an early-market price. (OpenAI)

Competition Makes the Situation More Urgent, Not Less

A year ago, ChatGPT had a stronger aura of dominance.

It felt further ahead.

Its technical lead looked more obvious.

Its consumer position looked more secure.

Today, that picture looks different.

Gemini has become far more serious.

Claude has become much more serious.

And the broader AI market now feels more competitive than many people expected.

That matters because weaker relative dominance makes low-margin consumer pricing harder to justify.

If a company had an overwhelming moat, it could potentially keep subsidizing aggressive pricing for longer.

But if real alternatives are improving fast, then the path to profitability becomes more important.

In my view, that makes price increases more likely, not less.

Anthropic Is One of the Signals I Watch Closely

One of the biggest surprises to me in the last several months has been Anthropic.

Its product quality has become impossible to ignore, especially for more serious users.

And Anthropic itself says its run-rate revenue reached $14 billion, with revenue growing more than 10x annually in each of the past three years. Anthropic also framed that momentum around strong enterprise adoption. (Anthropic)

That part matters a lot.

Why?

Because B2B economics are usually stronger than consumer economics.

Businesses can justify larger contracts.

They buy for teams.

They buy for workflows.

They buy for productivity gains that can be measured financially.

Individual consumers are different.

They are more price-sensitive, even when they depend heavily on the product.

So if leading AI companies see stronger economics on the business side, that creates even more pressure to rethink how aggressively they price for individuals.

The Consumer AI Pricing Model May Change

I do not think the only possible outcome is a simple price increase.

Another possible direction is usage-based pricing.

That would make sense too.

Some users ask a few light questions per day.

Others use advanced models for writing, coding, analysis, image generation, file work, and deep research for hours.

Those are very different cost profiles.

At some point, a flat monthly price may stop making sense for the most powerful individual plans.

So I see two realistic possibilities:

1. Entry-level paid plans get more expensive.

or

2. The market moves toward a stronger usage-based model for advanced access.

Either way, I do not expect the current consumer AI pricing structure to remain untouched for long.

My Core Prediction

My prediction is simple:

Within the next 6 to 12 months, the starting price for meaningful AI access for individuals will increase significantly.

Possibly by 100%.

And even if that exact number does not happen, I still think the general direction is clear.

The market is telling us that frontier AI is valuable.

The infrastructure behind it is expensive.

The competition is real.

And the pressure to build sustainable businesses is only going to increase.

And Yes, People Will Still Pay

That is the final point.

I think people will still pay.

Why?

Because AI is becoming part of daily life.

Not for everyone in the same way, and not at the same speed.

But clearly enough that pricing power is starting to change.

When a tool becomes embedded in your work, your thinking, your planning, your output, and your ability to move faster, you do not evaluate it the same way you evaluate a random app subscription.

You evaluate it more like digital infrastructure.

That is why I think the era of underpriced consumer AI may not last much longer.

And that is why I believe the next 6 to 12 months could bring a meaningful shift in how AI for individuals is priced.