Short answer: OpenAI has filed confidential paperwork for an IPO, and according to the Financial Times it is rebuilding ChatGPT into a "superapp" that puts Codex, its AI agent, in front of the hundreds of millions of people who already open ChatGPT every week. Today Codex takes effort to use: you download it, set up a sandbox, connect your tools. That effort is exactly why the owners using it now have an edge. The superapp removes it. If you run a small business, the time to build a real head start with Codex is now, while it is still something you have to seek out, because once it is one tap inside ChatGPT, knowing how to use it stops being what sets you apart. It also matters defensively: as agents reach everyone, some of what you sell gets commoditized, and you only see which of your offers are exposed once you are using the tools yourself.
There were two pieces of OpenAI news this week, and almost everyone is paying attention to the wrong one.
Two stories, not one
The first is the headline. OpenAI submitted confidential paperwork to begin going public, a draft registration the SEC keeps private until the company is ready to show its hand. It was last valued at around 850 billion dollars and has not committed to a date, saying only that there are things it would rather finish as a private company first (CNBC). Interesting if you follow markets. Not something that changes your Tuesday.
The second is the one that does. To get ready for the public markets, OpenAI is reportedly rebuilding ChatGPT itself. The Financial Times describes a coming overhaul that turns ChatGPT into a "superapp," one place that bundles its AI agent Codex, agents that carry out multi-step tasks like booking and scheduling, image generation, and tools from outside partners, with the first version expected in the next few weeks (Fortune, TechCrunch). The pitch is a shift from an app that answers your questions to an app that does your work.
Why the superapp matters more to you than the valuation
Here is the part to sit with. Right now, using Codex takes intent. You have to know it exists, download it, set up the sandbox, and connect your tools. That friction is small, but it is real, and it is the single reason a non-technical owner can still be ahead today. Most people have not crossed it. OpenAI has said Codex passed 5 million weekly users (OpenAI). ChatGPT has hundreds of millions. The superapp is the bridge between those two numbers.
When Codex lives inside the app your customers, your competitors, and their teenagers already open every day, the friction is gone. Adoption stops being a decision and becomes a default. That is exactly what OpenAI wants heading into an IPO. For you, it means the quiet head start you can build right now has an expiration date on it.
The quieter risk: what gets commoditized
There is a second effect, and it is the one you feel in your margins. Every service you sell has a price because it takes skill, time, or access your customer does not have. As agents reach everyone, some of that gap closes. Work you charge for today, a routine analysis, a setup, a first draft, becomes something an agent does in minutes, either for your customer directly or for the competitor who figured it out first. That is commoditization up close: the same deliverable, a fraction of the price.
The uncomfortable part is that you cannot tell which of your offers are exposed from the outside. It does not become obvious until you are using the tools yourself and can see what they actually do well. The owner who is hands-on knows which parts of the business to defend, which to reprice, and which to retire before the market decides for them. The owner who waits finds out when a client says they tried it themselves. This is the real reason to immerse yourself now: not the shiny tool, but seeing what is changing in time to move first.
It does not take the whole market
Here is what makes waiting genuinely dangerous: the threat is not everyone adopting at once. It is the few who adopt well. All it takes is a handful of competitors who learn this and apply it better than you, and the playing field stops being level. They quote faster, follow up sooner, deliver more for the same price, and carry more clients without adding headcount. That is leverage, and it compounds quietly while the rest of the market tells itself there is still time. You will not read about it in a headline. You will feel it when a deal you used to win goes to someone who simply moved first.
The trap: "I'll just pick it up when it's built into ChatGPT"
This is the reasonable-sounding mistake, so let me be blunt about it. When Codex is one tap inside ChatGPT, everyone will have it. A tool everyone can open is not an advantage. What you did with it for the months before that is.
You can make a tool easier to open. You cannot make the learning curve shorter. Knowing which job to hand an agent, how to brief it, where it needs a guardrail, what to check before you trust it: that is reps, and reps take calendar time nobody can compress for you. The owner who started this quarter will not be one tap behind the owner who waited. They will be a hundred finished jobs of judgment ahead. I made that case in full in why non-technical owners are early, not late. The superapp is the deadline that argument was missing.
While your competitor waits for the easy version, you spend the next ninety days teaching Codex your business: your clients, your numbers, the way you write, the tasks you are tired of. When the superapp lands and they finally open it, they start at zero. You start with an agent that already knows how you work.
What moving now actually looks like
You do not need a strategy. You need one finished task. I have helped more than ten established businesses put Codex to work, and the ones who get traction never set out to "learn Codex." They gave it one job and watched. Here is the order that works:
- Download the Codex desktop app. It has to live on your computer to do real work, not in a browser tab.
- Set up the sandbox and permissions correctly. This is the step everyone rushes, and the one that decides whether Codex can actually operate or stalls at every turn. Here is exactly how.
- Connect one tool you already live in. Your CRM, your notetaker, your inbox. Start with what a plugin really is, then install your first one.
- Give it one real job you already do. A weekly report, a data cleanup, a first-draft proposal. Hand it over and watch it work.
If you want the plain-English version of what Codex even is before you start, I wrote What is Codex for exactly that.
Wait for the superapp, or start this quarter
| Wait for the superapp | Start this quarter | |
|---|---|---|
| Effort to access Codex | None, it is built into ChatGPT | Small: download, sandbox, connect one tool |
| Who else is doing it | Everyone, all at once | Almost no one in your market yet |
| What it is worth as an edge | Nothing, access is universal | Real, because few have the reps |
| Where you stand when it goes mainstream | Day one of the learning curve | Months of judgment ahead |
Common questions
Did OpenAI file for an IPO?
Yes. In June 2026 OpenAI confirmed it had submitted confidential paperwork, a draft S-1 registration, to the SEC, the first formal step toward going public. The company was last valued at around 850 billion dollars and said the timing is not decided yet. A confidential filing means the details stay private until OpenAI chooses to make them public.
What is the OpenAI superapp?
According to the Financial Times, OpenAI is rebuilding ChatGPT into a superapp: a single app that bundles its AI agent Codex, agents that complete multi-step tasks like booking and scheduling, image generation, and tools from outside partners, with the rollout expected in the coming weeks. The shift is from an app that answers your questions to an app that does your work.
How does the OpenAI IPO affect my small business?
Not through the stock, through adoption. To get ready for the public markets, OpenAI is pushing Codex, its AI agent, in front of the hundreds of millions of people who already use ChatGPT every week. As agents reach everyone, some of what you sell gets commoditized: work a client used to pay you for, an agent can now do for them in minutes. You cannot tell which of your offers are exposed until you are using the tools yourself and can see what is changing. The owners who immerse themselves now are the ones who reposition before the market reprices them.
Should I wait for the superapp instead of learning Codex now?
No. When Codex is one tap inside ChatGPT, everyone will have access to it, so having access will not set you apart. What sets you apart is the months of real reps you put in first. You can make a tool easier to open. You cannot shortcut the learning curve. Waiting trades your only head start for convenience.
What should a small business owner do right now?
Download the Codex desktop app, set up the sandbox and permissions correctly, connect one core tool you already use, and give it one real job you already do, such as a weekly report or client follow-up. One finished task teaches you more than a month of reading about it.
The superapp will make Codex easy for everyone. The head start is only available to the people who start before it does.
Ready to claim the window? Download Codex, set up the sandbox, then see why you are early, not late.