One Person. Four Months. About ¥1 Million. AI Has Changed the Rules of Building a Company.
Building toward launch without employees, alongside AI. A look at my actual OpenAI spending and how AI is changing what it takes to build a company.

Illustrative image. The total has been corrected, but individual entries retain the original artwork. Refer to the spending table in this article for the accurate dates and breakdown.
It has been roughly four months since I started developing ZipUpp.
Today, the platform is almost complete and very close to launch.
I have built it without employees and without a development team. I am not even a professional software engineer.
AI made this possible.
As of September 19, 2026, my total OpenAI spending related to this development has reached:
¥820,951.
There will probably be some additional costs before launch, but I expect the total to remain around ¥1 million.
When I stop and think about that, it is remarkable how much the world has changed.
Four months to build something like this
ZipUpp is not simply a website.
It is a recruiting platform designed to connect candidates and employers while integrating AI-powered screening, anonymization, information disclosure, AI interviews, and other functions into one system.
Traditionally, building a platform of this scope from scratch would likely have required multiple engineers, significant funding, and a much longer development period.
Instead, I built it alone.
With AI, I can turn an idea into a design, turn that design into code, test it, identify problems, fix them, and immediately move on to the next feature.
That cycle can now happen at extraordinary speed.
And after roughly four months, ZipUpp is nearly ready to launch.
One person. Four months. About ¥1 million.
A few years ago, this would have been difficult to imagine.
My actual OpenAI spending
Here is the detailed OpenAI payment record, totaling ¥820,951 as of September 19, 2026.
| Date | Description | Amount |
|---|---|---|
| 2026/5/8 | ChatGPT Plus | ¥0 |
| 2026/6/4 | Credit | ¥3,410 |
| 2026/6/5 | ChatGPT Pro 5x | ¥16,800 |
| 2026/6/8 | ChatGPT Pro 20x | ¥14,973 |
| 2026/7/8 | ChatGPT Pro 20x | ¥30,000 |
| 2026/8/4 | Credit | ¥27,280 |
| 2026/8/4 | Credit | ¥27,280 |
| 2026/8/5 | Credit | ¥34,100 |
| 2026/8/5 | Credit | ¥27,280 |
| 2026/8/6 | Automatic credit top-up | ¥5,988 |
| 2026/8/7 | Automatic credit top-up | ¥6,097 |
| 2026/8/7 | Automatic credit top-up | ¥6,138 |
| 2026/8/7 | Automatic credit top-up | ¥6,049 |
| 2026/8/7 | Automatic credit top-up | ¥6,015 |
| 2026/8/7 | Automatic credit top-up | ¥5,974 |
| 2026/8/8 | ChatGPT Pro 20x | ¥30,000 |
| 2026/8/8 | Automatic credit top-up | ¥6,131 |
| 2026/8/8 | Automatic credit top-up | ¥6,057 |
| 2026/8/8 | Automatic credit top-up | ¥5,981 |
| 2026/8/8 | Automatic credit top-up | ¥5,995 |
| 2026/8/8 | Automatic credit top-up | ¥6,023 |
| 2026/8/9 | Automatic credit top-up | ¥6,008 |
| 2026/8/29 | Automatic credit top-up | ¥6,820 |
| 2026/9/2 | Credit | ¥27,280 |
| 2026/9/2 | Credit | ¥27,280 |
| 2026/9/2 | Automatic credit top-up | ¥5,981 |
| 2026/9/2 | Automatic credit top-up | ¥5,981 |
| 2026/9/2 | Automatic credit top-up | ¥5,974 |
| 2026/9/3 | Credit | ¥27,280 |
| 2026/9/3 | Credit | ¥27,280 |
| 2026/9/3 | Credit | ¥27,280 |
| 2026/9/3 | Automatic credit top-up | ¥5,981 |
| 2026/9/3 | Automatic credit top-up | ¥5,995 |
| 2026/9/3 | Automatic credit top-up | ¥5,974 |
| 2026/9/3 | Automatic credit top-up | ¥5,974 |
| 2026/9/3 | Automatic credit top-up | ¥5,974 |
| 2026/9/3 | Automatic credit top-up | ¥5,974 |
| 2026/9/5 | Credit | ¥27,280 |
| 2026/9/7 | Credit | ¥15,345 |
| 2026/9/8 | ChatGPT Pro 20x | ¥30,000 |
| 2026/9/16 | Credit | ¥27,280 |
| 2026/9/17 | Credit | ¥27,280 |
| 2026/9/17 | Credit | ¥27,280 |
| 2026/9/18 | Credit | ¥27,280 |
| 2026/9/18 | Automatic credit top-up | ¥5,995 |
| 2026/9/18 | Credit | ¥27,280 |
| 2026/9/18 | Automatic credit top-up | ¥5,974 |
| 2026/9/19 | Credit | ¥119,350 |
| Total | ¥820,951 | |
The most exciting part comes next
Building the product does not guarantee success.
The real marketing begins now, and the market will ultimately decide whether ZipUpp succeeds.
But I believe the concept has the potential to significantly change how recruiting works.
And imagine what becomes possible if users begin adopting it at scale.
A platform built by one person, with AI, in roughly four months and for around ¥1 million could potentially grow into a system used by people across a massive industry.
That possibility is incredibly exciting to me.
AI isn't simply another productivity tool
Building ZipUpp has convinced me of something much bigger:
AI is changing the way businesses are built.
In the past, doing something ambitious generally required building an organization first.
You needed engineers, designers, marketers, planners, capital, and infrastructure.
AI is beginning to change that assumption.
The amount of execution power available to a single person is becoming dramatically greater.
That is why ZipUpp is also an experiment for me:
How large can a one-person company become when AI is used to its full potential?
Instead of measuring success by how many employees a company has, what happens if we maximize the productivity and capabilities of one person through AI?
How far can that company go?
That is an incredibly exciting question.
The AI revolution is no longer something waiting for us in the future.
It is already happening.
And the pace of change is likely to accelerate.
We cannot afford to miss this wave.
I want to see how far I can take ZipUpp—and how far the idea of an AI-powered one-person company can go.