How Computing Power, Regulation, and Reality Challenge the AI Giant
Part 2 of the series “OpenAI Becomes the Operating System of the Future” (click here for Part 1)
1. The Value of the Future: $500 Billion in Trust
Three years ago, OpenAI was a research-driven lab with a visionary chatbot.
Today, according to current estimates, it is the most valuable private technology company in the world, with a valuation of approximately $500 billion.
This figure is more than a financial metric. It is a vote of confidence in an idea: that OpenAI is not merely offering software, but laying the foundation for a new digital economy—with ChatGPT as the interface and AgentKit as the automation layer.
However, those who venture this far forward need more than capital. They also need hardware—on an unprecedented scale.
2. The New Gold Rush Is Called Compute
In the AI economy, the bottleneck is no longer the dataset, but computing power.
OpenAI’s models consume so much energy and GPU capacity that entire data centers must be redesigned to meet demand.
The answer: strategic alliances
- AMD Deal: In summer 2025, OpenAI and AMD agreed to a record deal: 6 gigawatts of GPU power, delivered over several years.
In return, OpenAI receives option rights on up to 10% of AMD’s equity value—a clever exchange of future for capacity.
Goal: to become more independent from Nvidia, which has dominated the GPU market until now. - Stargate Initiative: Together with Oracle and SoftBank, OpenAI is building its own computing infrastructure, with locations in the US and Asia.
Investments of up to $500 billion are planned by 2026 to scale computing power to over 10 gigawatts.
For comparison: this is roughly equivalent to the electricity consumption of a smaller European country.
These projects mark the beginning of a new competition: the compute war.
It is no longer just the quality of a model that matters, but who has enough chips and energy to run it.
3. From Cloud to Body: OpenAI’s Hardware Ambition
And while OpenAI is building data centers in the background, it is already thinking about the frontend—about what stands between humans and AI.
Together with Jony Ive, the legendary former Apple designer, OpenAI is working on an “always-on” device without a display. A smart companion that listens, speaks, understands—and is permanently present.
Much remains unclear: What does the interface look like when there is no screen?
How do you protect privacy in a device that is always listening?
And are people ready to give an AI assistant so much space in their daily lives?
Yet the idea is fascinating: The smartphone as a window to the world is replaced by an AI that lives in the world—in earbuds, glasses, microsystems. OpenAI wants to reinvent the interface, not just the software.
4. Sora 2: When AI Learns to See
In parallel, OpenAI demonstrates with Sora 2 where the technical journey is heading: The model can generate highly realistic videos from text, sketches, or audio—with physically accurate movement, light, shadow, and emotion. What was still experimental in 2024 often appears deceptively real in 2025.
This opens up creative possibilities—but also ethical abysses.
Because Sora 2 can not only create fantasy scenes, but also deceptively realistic clips of real people. And this leads to an old, unresolved problem: Who owns a person’s likeness in the AI world?
OpenAI is attempting to respond with opt-out mechanisms for rights holders. However, critics call this a reversal of the protection principle: Instead of AI companies having to ask for permission, artists and studios must now actively object.
In doing so, OpenAI is entering legal uncharted territory—and symbolically represents the gray zones of the generative era.
5. Europe Applies the Data Protection Brake
While OpenAI expands globally, Europe is watching the rapid development with growing skepticism.
Not only because the EU AI Act is about to be implemented, but also because the digital everyday experience urgently needs to be modernized.
One example: the cookie reform. After years of banner overload, the EU is planning a radical simplification: Instead of clicking “Accept” on every website, users will in future be able to set centrally in the browser which data they want to share. This reduces friction and strengthens user autonomy—a signal that regulation can also be UX-friendly.
For OpenAI and other AI platforms, this means: Transparency, traceability, and consent become part of the product, not just a legal obligation.
6. The Balance Between Power and Responsibility
OpenAI is at a turning point.
With each new feature, not only reach grows, but also responsibility.
Because those who design the interface between humans and machines ultimately define how we experience technology—and whom we trust.
OpenAI’s strength lies in speed and vision. The danger lies in the same speed—and in the question of whether society, law, and infrastructure can keep pace.
The next twelve months will determine whether OpenAI fulfills the promise of an “AI operating system for everyone”—or whether it fails under its own gravity.
Conclusion: The Gatekeeper of the Next Era
OpenAI wants to be more than an AI provider.
It is building a platform that is simultaneously tool, marketplace, and infrastructure.
ChatGPT as the control center, AgentKit as the engine, AMD as the energy source—and perhaps soon an AI device in your pocket.
What Apple once was for the touchscreen, OpenAI could become for conversation:
The operating system of our language.
