An AI chip with 208 billion transistors needs more than a clever design. Factories must be able to make all those tiny switches. That is where a familiar Dutch name appears: Philips helped both ASML and TSMC get started.
In the clip, Nvidia chief Jensen Huang highlights that industrial family history. The connection runs from Dutch chipmaking equipment through Taiwanese manufacturing to the computing power used to train and run AI systems.
In 1984, Philips and ASM International jointly founded ASM Lithography, later ASML. It started in a leaky shed in Eindhoven, building on research dating to the 1970s.
Think of ASML’s equipment as an exceptionally precise projector. Its lithography machines image patterns onto a light-sensitive silicon wafer. Layer by layer, chip structures take shape. The challenge is repeating that at an extraordinarily small scale in exactly the right place.
Three years later, Philips also helped establish TSMC, alongside Taiwan’s government and private investors. Morris Chang shaped its distinctive model: a factory making chips for other companies.
Designers no longer had to build their own chip factory at the same time. TSMC could focus on manufacturing while customers focused on design. That division of work became an important engine of the chip industry’s growth.
Nvidia Blackwell is a concrete example. Nvidia designs the AI processor and has it made with a custom TSMC process. Two silicon dies are linked by a very fast connection and operate as one GPU. Together they contain 208 billion transistors.
The roles come together: Nvidia designs the computing engine, TSMC manufactures the chip and ASML supplies important production machines. A large network of other specialist suppliers is also essential.
Philips does not make those Nvidia chips itself, but helped build two crucial links in their history. That is what makes the story remarkable: decisions and research from the 1980s still contribute to the AI appearing on our screens.
