When you ask an AI a question, the answer is computed somewhere on a chip. The race to build those chips is one of the most expensive battles in technology today — and in 2026 it intensified sharply.
What is happening
Etched launched with a $5B valuation and over $1B in signed contracts for specialised AI inference chips. At the same time, DeepSeek announced it is designing its own custom silicon to reduce dependence on current suppliers. Demand for dedicated inference capacity has grown large enough to support an entire new wave of hardware companies.
Why "dedicated" chips?
General-purpose chips are flexible but not optimal. When you know exactly which computation will be repeated billions of times, you can build a chip that does only that — far faster and with far less energy. That is the logic behind dedicated AI silicon.
The lesson hidden in this news
Good news for children drawn to electronics and hardware: in an era when everyone talks about software, the industry's real bottleneck is hardware. A student who understands circuits, has worked with microcontrollers and knows digital logic is entering a field with rising global demand.
Conclusion
Behind every AI answer stands a chain of hardware engineering. If your child is more interested in "how does it work" than "how do I use it", our electronics and robot building programs are a natural starting point.
Get your child ready for this world
AI, robotics and programming programs at Novin Zehn — in person in Tehran and live online worldwide.
Explore programs