You Are Paying to Build the Engine. Then They Throw You Out of the Car.

The most powerful models are released to a vetted few while everyone else pays for the privilege of waiting at the back of the line.

A few weeks ago I wrote about Anthropic switching off its most powerful AI for the whole world. I thought that was the low point. I was wrong. It just became a pattern.

This week OpenAI previewed GPT-5.6, its most capable model family yet. And once again, almost nobody gets to touch it.

What Actually Happened With GPT-5.6

Read the announcement yourself and the language is almost clinical. The preview is "not a broad self-service program." Participation is "limited to a small group of trusted partners and organizations." It is "not available to individual consumers or through public enrollment." There is "no public application or waitlist."

So who decided who is trusted? Here is the part that should stop every founder cold. In OpenAI's own words: "As part of our ongoing engagement with the U.S. government, we previewed our plans and the models' capabilities ahead of launch. Working in coordination with the government, we are starting with a limited group of trusted partners and organizations whose participation has been shared with the government before releasing the models more broadly."

In practice that meant releasing GPT-5.6 Sol, the most powerful of the new models, to roughly 20 organizations whose names were approved by the government one by one. The request came from the White House, with the Trump administration asking OpenAI to stagger the launch and sign off on partners individually. By several accounts it is the first time an American AI company has shipped a frontier model under a government-managed access list. To its credit, OpenAI said it does not believe "this kind of government access process should become the long-term default." But the precedent is already set.

Sit with it. The most powerful intelligence ever built gets handed first to a pre-approved circle, vetted in coordination with a government, while the rest of us wait for permission. This is no longer about one company or one country. It is about who gets to stand at the front of the line for the future, and who pays for the privilege of standing at the back.

The Promise Was the Opposite of This

Remember what we were all sold. The entire promise of AI was that it would flatten the playing field. A kid in a village with a laptop and an internet connection would have the same brainpower at his fingertips as a team in Silicon Valley. Intelligence, finally, democratized. The greatest equalizer in human history.

That was the dream. Build something amazing from anywhere. Compete with anyone. Geography and bank balance would stop being destiny. What we actually got is the oldest story in the world wearing a new outfit. A small group at the front. Everyone else waiting for a turn that arrives late, if it arrives at all.

Here Is the Part That Is Genuinely Unfair

Let me be precise about what stings, because it is not just exclusion. It is the arrangement underneath it. The people sitting at home tonight, paying for API keys, burning their own savings on tokens, building little tools and big dreams, are not bystanders. They are fuel.

Every API call you pay for trains the flywheel. Your usage, your feedback, your edge cases, your money, all of it feeds the machine that these labs then refine and hand, first, to a chosen few. You are paying to sharpen a knife that you are not allowed to hold first.

Funded by the world, reserved for a few. The whole planet pays into the engine; a vetted circle gets to drive first.

Think about what that means competitively. Whoever gets prior access to the frontier is permanently one step ahead. They ship faster. They automate sooner. They win contracts you cannot even bid on yet, because the tool that would let you compete is still behind a door marked "trusted partners only." By the time it reaches you, they have compounded six months of advantage on top of the model you helped pay to build.

That is not a small gap. In a field moving this fast, a few months of exclusive access is the difference between leading a market and being a footnote in it.

You cannot charge a developer in Lagos, Dhaka, Karachi, Manila or Sao Paulo for the API that funds your frontier research, and then tell that same developer they are not trusted enough to access what their money helped create.

So here is my question, and I want a real answer. Why are we paying you to build things, only to be blocked from using them? If the technology is too powerful or too sensitive to share with the whole world at once, fine. Make that argument honestly. But then do not take the whole world's money to build it. You do not get to socialize the cost across the entire planet and privatize the access to a hand-picked few.

This Is the Pattern, Not the Exception

If this were a one-off, I would not be writing a second article. But step back and the shape is unmistakable. The United States has gated every transformative technology the moment it became powerful, going back decades.

The tool changes. The gatekeeping does not. Encryption, then the chips, then the models themselves.

In the 1990s, strong encryption, which is just mathematics that protects ordinary people's privacy, was classified as a munition under the Arms Export Control Act. When Phil Zimmermann released PGP and it spread worldwide, the government opened a multi-year criminal investigation into him. For most of that decade, sharing powerful privacy tools with the rest of the world could be treated as arms trafficking.

In January 2025, the Bureau of Industry and Security sorted the entire planet into three tiers for access to advanced AI chips. A favored inner group got unrestricted access. Most of the world got rationed. Around twenty countries got nothing.

And now, in the same season, the two leading AI labs both restrict their frontier models in coordination with the government, releasing first to a vetted circle. Encryption. Then the chips that run the models. Then the models themselves. Each time, the same move: the most powerful tool of the era is held close, and the rest of the world is told to wait its turn. The technology changes. The gatekeeping does not.

The Irony Nobody Wants to Say Out Loud

Here is what makes this almost absurd. This frontier was not built by one nation in isolation. It was built on talent imported from the exact countries now waiting at the back of the line.

Independent analysis of the world's elite AI researchers found that US institutions employ the majority of them, and that this lead rests almost entirely on foreign-born talent, with about two-thirds of the top-tier AI researchers at US institutions having done their undergraduate degrees abroad. A separate study from Georgetown's CSET found that nearly 70% of America's top AI researchers were born abroad.

So the engineers come from our countries. The paying users come from our countries. And the output of all of it gets handed, first, to a closed circle most of us will never be inside. We are good enough to build it. We are good enough to fund it. We are just not good enough to use it first.

What I Am Actually Asking For

I am not anti-American. I am not anti-OpenAI or anti-Anthropic. These are extraordinary companies doing genuinely historic work, and I use their tools every day to build my own. What I am against is a future where the single most empowering technology in human history defaults to the same old hierarchy. Insiders first. Everyone else later, maybe, on terms decided without them.

If AI is going to be the equalizer it was promised to be, then access cannot keep flowing along lines of who you know, where you are, and whose approved list you happen to be on. The boy in the village was supposed to get the same shot as the company in the valley. That was the entire point. We are not asking for charity. We are asking to stop being treated as the engine that powers a car we are not allowed to ride in.

At Integral Technologists, we help business owners actually put these tools to work, wherever in the world they are sitting, because capability should not depend on a postcode. While the inner circle argues about access, the businesses that win are the ones executing fastest with the tools that are available right now. That is the work we do, through AI automation systems and a complete business setup built to run without you.

So I will ask you what I keep asking. Should the future be built by everyone and open to everyone, or built by everyone and reserved for a few? If you have ever paid for an API key with a dream of building something that matters, you are not just a customer. You are the fuel. And you deserve to know it.

Frequently Asked Questions

What is the GPT-5.6 limited preview?

In June 2026, OpenAI previewed its GPT-5.6 model family (Sol, Terra and Luna, with Sol the most powerful) to roughly 20 partners whose participation was individually approved by the US government. OpenAI states it is not a broad self-service program, it is limited to a small group of trusted partners and organizations, and there is no public application or waitlist.

Why did OpenAI restrict GPT-5.6 access?

The request came from the US government. Reporting from CNN and Axios says the White House and the Trump administration asked OpenAI to stagger the GPT-5.6 release and approve partners one by one. In its own words, OpenAI previewed its plans and the models' capabilities to the government ahead of launch and started with a limited group of trusted partners before a broader release. OpenAI also said it does not believe this kind of government access process should become the long-term default.

Can developers or the public use GPT-5.6 right now?

No. The preview is explicitly not available to individual consumers or through public enrollment, with no public application or waitlist. Access is limited to a pre-approved circle of partners. OpenAI has said a broader release is planned in the coming weeks.

Is this the same as the Anthropic Fable 5 shutdown?

It is the same pattern. Two weeks earlier, a US export control directive forced Anthropic to switch off Fable 5 and Mythos 5 worldwide, which I covered in this article. Now OpenAI is releasing its most powerful models first to a government-vetted circle. Two leading labs, the same season, the same gatekeeping in coordination with the government.

What can businesses outside the inner circle do about it?

Do not build your company on a single model or vendor you do not control. Treat frontier AI as leverage, keep working options across multiple providers, own your data, workflows and customer relationships, and focus on putting available tools to work faster than competitors. For most real businesses, execution still beats access. If you want help building that, get started here.

Sources and References

1. OpenAI, Previewing GPT-5.6 Sol: a next-generation model (official announcement).
2. OpenAI Help Center, A preview of GPT-5.6 Sol, Terra, and Luna (verbatim access limits and government coordination language).
3. TechCrunch, OpenAI limits GPT-5.6 rollout after government request, says restrictions shouldn't be the norm (June 26, 2026).
4. CNN Business, White House asks OpenAI to limit its next model release (June 25, 2026).
5. Axios, Trump administration asks OpenAI to limit release of GPT-5.6 (June 25, 2026).
6. CNBC, OpenAI limits new AI models to "trusted partners" at request of US government (June 26, 2026).
7. Anthropic, Statement on the US government directive to suspend access to Fable 5 and Mythos 5 (June 12, 2026).
8. Wikipedia, Crypto Wars (encryption classified as a munition; the PGP investigation).
9. Congressional Research Service, US Export Controls and China: Advanced Semiconductors (R48642) (the tiered AI chip rules).
10. MacroPolo, The Global AI Talent Tracker (US AI lead built on foreign-born talent).
11. CSET, Georgetown University, Voices of Innovation (nearly 70% of top US AI researchers born abroad).

Faiz Ahmed

CEO of Integral Technologists. Eight years building businesses in healthcare and medspa, now helping owners across the US and Pakistan put AI to work without being hostage to any one vendor. Connect on LinkedIn.