Nathan Lambert

All these companies had fabs, and for most of the fabs, they threw them away or sold them off or they got rolled into something else. And now everyone relies on TSMC, right? Including Intel, their latest PC chip uses TSMC chips, right? It also uses some Intel chips, but it uses TSMC process.

All these companies had fabs, and for most of the fabs, they threw them away or sold them off or they got rolled into something else. And now everyone relies on TSMC, right? Including Intel, their latest PC chip uses TSMC chips, right? It also uses some Intel chips, but it uses TSMC process.

Yeah. So, I mean, like I mentioned, right, the cost of building a fab is so high. The R&D is so difficult. And when you look at, like, these companies that had their own vertical stack, there was an antiquated process of, like, okay, like, I'm so hyper-customized to each specific chip.

Yeah. So, I mean, like I mentioned, right, the cost of building a fab is so high. The R&D is so difficult. And when you look at, like, these companies that had their own vertical stack, there was an antiquated process of, like, okay, like, I'm so hyper-customized to each specific chip.

Yeah. So, I mean, like I mentioned, right, the cost of building a fab is so high. The R&D is so difficult. And when you look at, like, these companies that had their own vertical stack, there was an antiquated process of, like, okay, like, I'm so hyper-customized to each specific chip.

But as we've gone through the history of sort of like the last 50 years of electronics and semiconductors, A, you need more and more specialization, right? Because Moore's law has died. Denard scaling has died, i.e. chips are not getting better just for free, right? You know, from manufacturing, you have to make real architectural innovations, right?

But as we've gone through the history of sort of like the last 50 years of electronics and semiconductors, A, you need more and more specialization, right? Because Moore's law has died. Denard scaling has died, i.e. chips are not getting better just for free, right? You know, from manufacturing, you have to make real architectural innovations, right?

But as we've gone through the history of sort of like the last 50 years of electronics and semiconductors, A, you need more and more specialization, right? Because Moore's law has died. Denard scaling has died, i.e. chips are not getting better just for free, right? You know, from manufacturing, you have to make real architectural innovations, right?

Google is not just running on Intel CPUs for web serving. They have a YouTube chip, they have TPUs, they have pixel chips, they have a wide diversity of chips that, you know, generate all the economic value of Google, right? Running, you know, it's running all the services and stuff. And so, and this is just Google and you could go across any company in the industry and it's like this, right?

Google is not just running on Intel CPUs for web serving. They have a YouTube chip, they have TPUs, they have pixel chips, they have a wide diversity of chips that, you know, generate all the economic value of Google, right? Running, you know, it's running all the services and stuff. And so, and this is just Google and you could go across any company in the industry and it's like this, right?

Google is not just running on Intel CPUs for web serving. They have a YouTube chip, they have TPUs, they have pixel chips, they have a wide diversity of chips that, you know, generate all the economic value of Google, right? Running, you know, it's running all the services and stuff. And so, and this is just Google and you could go across any company in the industry and it's like this, right?

Cars contain 5,000 chips, you know, 200 different varieties of them, right? All these random things. A Tesla door handle has two chips, right? It's ridiculous. And it's a cool door handle, right? You don't think about it, but it has two really chipped penny chips in there, right?

Cars contain 5,000 chips, you know, 200 different varieties of them, right? All these random things. A Tesla door handle has two chips, right? It's ridiculous. And it's a cool door handle, right? You don't think about it, but it has two really chipped penny chips in there, right?

Cars contain 5,000 chips, you know, 200 different varieties of them, right? All these random things. A Tesla door handle has two chips, right? It's ridiculous. And it's a cool door handle, right? You don't think about it, but it has two really chipped penny chips in there, right?

Anyway, so as you have more diversity of chips, as you have more specialization required, and the cost of fabs continues to grow, you need someone who is laser focused on building the best process technology and making it as flexible as possible.

Anyway, so as you have more diversity of chips, as you have more specialization required, and the cost of fabs continues to grow, you need someone who is laser focused on building the best process technology and making it as flexible as possible.

Anyway, so as you have more diversity of chips, as you have more specialization required, and the cost of fabs continues to grow, you need someone who is laser focused on building the best process technology and making it as flexible as possible.

You know, there's more failure points, right? You know, you could have one little process related to like some sort of like chemical etch or some sort of like plasma etch or, you know, some little process that screws up. You didn't engineer it right. And now the whole company falls apart. You can't make chips. Right.

You know, there's more failure points, right? You know, you could have one little process related to like some sort of like chemical etch or some sort of like plasma etch or, you know, some little process that screws up. You didn't engineer it right. And now the whole company falls apart. You can't make chips. Right.

You know, there's more failure points, right? You know, you could have one little process related to like some sort of like chemical etch or some sort of like plasma etch or, you know, some little process that screws up. You didn't engineer it right. And now the whole company falls apart. You can't make chips. Right.