July 6, 2026 July 6, 2026 Home » Architecture » Architectural Design at $20 I was watching one of the Arab economic channels. One of the analysts was confidently discussing the imminent end of cheap AI subscriptions. His exact words were close to: Brace yourselves, the $20 monthly plan won't last long.
I'll be honest with you. I don't build my professional opinions on TV comments, especially when it comes to technology or stock markets. Analysis in that field easily mixes with personal interests, with a desire to move a stock price, and with the mechanisms of creating a media cycle around a specific investment thesis. So I turned off the program and started researching the matter myself.
What I found was more interesting than the original claim.
The numbers being circulated in the tech press deserve to be looked at directly, without drama and without belittlement.
Uber consumed its entire AI budget for 2026 within four months. By March of this year, 84 percent of its engineers were using Claude Code in their daily work, and about 70 percent of the company's software code is written with the assistance of AI. The company's leadership itself admitted that this massive increase in consumption of Tokens did not produce a proportional increase in the final product value. More spending, without necessarily more production.
At NVIDIA, the vice president of deep learning publicly stated that his team's computing costs have exceeded the costs of the employees themselves. That's worth stopping to ponder. The company that sells the AI infrastructure admits that the infrastructure now costs more than the people who use it.
One company, which has been talked about in tech circles without officially revealing its name, racked up a cloud usage bill of $500 million in just one month due to a lack of consumption caps. Token management, which seems like a technical detail, turned into an executive financial crisis.
Gartner predicts that global spending on AI agent software will reach $207 billion by 2026, within an overall tech spending environment of $6.31 trillion, representing a 13.5 percent increase from the previous year. AI is the primary driver of this growth.
In contrast, a study from the Massachusetts Institute of Technology (MIT) found that AI automation is economically viable in only about 23 percent of jobs that rely on computer vision. In 77 percent of cases, a human remains the cheaper option.
These numbers do not tell a single story. They tell a complex story, and that's why it's worth knowing.
I have been practicing architecture for more than twenty years. Recently, I wrote about my longing for rendering, about the V-Ray era and the hours of waiting for a single image. Today, I can generate dozens of visual concepts in minutes. I will say plainly that AI has boosted my productivity in ways I did not expect and do not want to turn back from.
But there is a question I've started asking myself, which is entirely separate from whether the tools are useful or not.
What will happen to the economics of architectural practice if the cost of these tools increases tenfold?
If the $20 subscription became $200, and the professional access cost required by large companies reached $1,000 per user per month, what would change? The tools wouldn’t change. The outputs wouldn’t change. What will change is the cost structure for each architectural practice that built its workflow around these tools. And this cost will show up somewhere in the end. In project fees. In hiring decisions. In the competitiveness of small offices compared to large offices.
We are not Uber. We are not NVIDIA. But we are practitioners now managing large parts of our daily work through cloud infrastructure that we do not own, and at prices set by companies whose primary commitment is not architectural specialization.
There is a second issue I find quite confusing,…