ChatGPT is reportedly back to exceeding 10% monthly growth. A new model is about to launch. Fundraising talks are heating up. And OpenAI has begun testing ads inside ChatGPT, with the expectation that advertising will make up less than half of long-term revenue.
Growth is strong and demand is accelerating. But behind the momentum sits a structural truth: advanced AI is extraordinarily expensive to build and operate.
Reports suggest OpenAI continues to run at a significant loss. Whether that is accurate or not, the economics are clear. Training frontier models, maintaining infrastructure, and serving millions of users in real time requires enormous capital. As companies move closer to IPO conversations, profitability shifts from ambition to obligation.
We are already seeing early signals of what that could mean. More subscription tiers. Tighter usage limits. Premium models reserved for paying users. Advertising introduced carefully, but deliberately.
The question is not whether AI will become commercialised. It already is.
The real question is what happens to access as the cost of intelligence rises.
Right now, AI is broadly available. Anyone can open a browser and start learning, exploring or creating. But history tells us that when things become expensive to produce, access fragments. The most powerful capabilities move behind higher price points. Performance advantages compound. And that is where inequality grows.
Those with the means will always access the strongest tools. But what about students, job seekers, career changers, small business owners and those already navigating the rising cost of living? If advanced AI becomes a premium advantage, we risk creating a two-tier system of productivity and opportunity.
This is not inevitable, but it is plausible.
So what can individuals and organisations do? Support accessible tools. Invest in digital literacy. Share knowledge. Advocate for inclusive AI policies at work. Push for transparency in how models are funded and monetised.
AI has the potential to be the great equaliser. It can lower barriers, expand knowledge and amplify human capability. But if access narrows as performance improves, it could just as easily deepen existing divides.
The future of AI should not belong only to those who can afford it.