Introduction

The investment world today is typically divided into siloed segments such as "Tech" versus "Utilities" or "Software" versus "Hardware." But what we are seeing here is the emergence of a new infrastructure that makes these terms irrelevant. At the heart of this revolution are two giants: Nvidia and Tesla. But these are not just chipmakers and carmakers. They are makers of the infrastructure of the next generation.

NVIDIA Bets on Becoming the Operating System of the AI Era

NVIDIA is no longer just a maker of computer chips. According to their CEO, Jensen Huang, they have a much bigger ambition. They want to become the foundation upon which the entire world is built to create and run artificial intelligence. Just like Windows was the foundation upon which the world built personal computers, Android has become the foundation upon which the world has built smartphones.

This theme is front and centre at NVIDIA’s GTC 2026 conference, which runs from March 16 to 20, 2026. Instead of being a typical hardware conference, this year’s conference is all about announcing something called the “Seven-Layer AI Stack.” That’s like a complete industrial operating system for AI. Everything you need to create, train, and run AI. But the underlying story at GTC is CUDA.

First announced back in 2006, CUDA is NVIDIA's proprietary programming environment for harnessing the power of the GPU for AI and scientific computing. It has been two decades since its introduction, and it is now the language that the AI world thinks in.

Four million developers worldwide are estimated to be writing in CUDA today. Every major AI model, from ChatGPT to Gemini, was developed on top of it. That is not a software product. That is a switching cost that has been baked into an entire new generation of software developers. Changing it out would take years, an effort that not a single company in the world ever wants to undertake, especially when the existing product works as well as it does.

But here’s the interesting part. NVIDIA is not trying to create this entire ecosystem alone. They are trying to become the foundation upon which others create. And they are succeeding. Big Companies like Siemens, Palantir, OpenAI, and many more are partnering with NVIDIA because they need reliable and powerful AI infrastructure. Each of these partnerships has CUDA at its core, which means that each of these partnerships is also a greater commitment to staying within NVIDIA's ecosystem. And Nvidia is the pinnacle of that.

Terafab: Tesla’s Master Engine Driving the Future of Energy and Robotics

Terafab, a fully vertically integrated semiconductor factory that will begin operation on March 21, 2026, could prove to be one of the biggest announcements in Tesla's illustrious history. Announced in a tweet by Elon Musk on X, Terafab will begin manufacturing the high-end AI chips that Tesla can no longer source from outside manufacturers like TSMC and Samsung. This is not a supply chain fix. This is a strategic move to declare independence.  

The Energy Intelligence Layer

The timing of Tesla's acquisition of a UK utility licence, announced the same week as Terafab, was not coincidental. In a world where the Strait of Hormuz crisis has placed energy security at the top of every country's agenda, Tesla's plans to provide global utilities could not have come at a more opportune time. Megafactories in Lathrop and Shanghai already churn out the hardware necessary to provide clean energy storage. Terafab will provide the intelligence to run it. The bespoke chips made at Terafab will power Autobidder, Tesla's high-speed AI-based energy trading and balancing service, at a computational rate that cannot be achieved with off-the-shelf chips.

Conclusion

NVIDIA has locked down the compute layer. Tesla is now racing to lock down the energy layer. Each is building essential infrastructure that the AI-driven economy will simply not be able to function without, and once they have been built, they will be impossible to unseat.