AI: Product or Foundational Technology?

G. Dan Hutcheson

The Chip Insider®

Summary:

AI: Product or Foundational Technology? As you know, McKinsey has talked up how huge the AI market will be in the future. They project it could be on a scale of trillions of dollars with the high end matching the GDP of today’s Germany ($4.4T), which is the third largest economy in the world… I did recently hear a senior McKinsey analyst admit to having a hard time seeing how it is monetized.

Beyond the hype cycle: I doubt that AI will ever be directly monetized on this scale in the sense that there will be a singular AI consumer product like the PC or smartphone. That’s because AI is not a product. AI is a foundational technology… Foundational technologies are things like math, Boolean logic, or chemistry … In other words, it does not matter if AI is monetized. The money will come from the products AI makes better, which are then monetized downstream… AI is already being used to create new products in areas such as EDA, …

AI is extending the horizon of market growth of computing beyond the deterministic math and Boolean logic that our markets are mostly based on today. AI is not deterministic math. AI is probabilistic math. Far more complex than the deterministic math problems computers solve day-to-day, AI’s probabilistic math has been limited by lack of compute power. But the steady climb of Moore’s Law to increase device complexity has solved this…

So as you can see the probabilistic math of AI is incredibly computationally complex and resource rich. And that’s a huge opportunity for everyone in the semiconductor industry. Who cares if AI is monetized? You should not care, because AI will drive greater monetization of semiconductors going forward.

Dan’s Book Shelf: The World According to China by Elizabeth C. Economy is a frightful look at how Xi Jinping is attempting to displace the rules-based Westphalia system of order with a China Order similar to the Qin-Han world order (Tianxia: all under heaven) that China’s dynasties used to run the country for millennia…

Maxims: Customers buy solutions, but pay for products: A product without a solution is a loser – one cannot be successful without the other. Customers only pay for products because products are the proof of the solution. For example, they will pay for a lithography tool, but they are really buying the ability to expose wafers. They will pay for an inspection tool. They are buying the embedded software that helps them improve yield by finding killer defects. You will pay for a car, but you are really paying for a set of trips that you have control over when you can leave, arrive, and where you go along the way…

“I've always been more interested in the future than in the past” — Grace Hopper

If you want to read the full article, click here.