Finale: Toward a vast frontier ― so we decided to try building one ourselves

The long walk is finally at its end. Up to here, we've looked together at the upheaval happening in the world of semiconductors. At the last, let me look back over the whole once — and then tell you honestly why I, the writer, prepared "four courses" beyond this corner. That's how I'll close.

F.1 What have we seen?

In the preface, I said "a Cambrian explosion is happening in the world of chips." We've unraveled its contents, in order.

   The "explosions" we've seen on this walk
 ───────────────────────────────────────────────────────────────
   The kinds of brain exploded       CPU, GPU, TPU, and quantum, bio (Ch. 2)
   The language of design was opened  ISA -> the x86/ARM empires -> RISC-V (Ch. 3-5)
   The transistor went 3D             flat -> FinFET -> nanosheet (Ch. 6)
   The making opened its door         drawing without carving a stamp, ride-share manufacturing (Ch. 7)
   AI began to have a body            Physical AI, robots, cars, space (Ch. 8)
 ───────────────────────────────────────────────────────────────
   = Three walls that were once closed behind heavy funds, special facilities,
     and a limited few specialists are wobbling at the same time.

What kind to make, who can design, how to manufacture — these three walls are wobbling at the same time. That's why it was an "explosion." This is a rare period, one that comes once in decades. We are standing right in the middle of it.

F.2 But ― design became possible. Yet there are few factories to make it

Here, let me offer an honest word that pours a little water on hands-off optimism. Now, anyone can design a chip. That's because RISC-V, a mechanism usable free of charge, and the tools for design, are in place. But design alone is still just a picture on paper. And there are only a very few factories in the world that will make it into a real chip.

There was a symbolic event. A pioneering manufacturing service that had supported individuals and small teams in making chips cheaply ended its business in 2025, out of funding trouble. A place many had relied on as "where to get things made" suddenly disappeared. Fortunately, another company is moving to take up the succession, but this one incident teaches us clearly: if even one of the few factories ends its business, a great many people are left unable to get their chips made — the door to design may be open, but the making side is still that precarious.

※ This is exactly why it's worth watching. Getting carried away with "anyone can do it now, it's a rosy age" and giving up with "it's a big-company world anyway" are both inaccurate. The door really did open. But whether we can keep it open depends on us — those who use it, support it, and carry it on — that, I think, is the most honest reading of where we are.

F.3 Even so ― a vast frontier spreads out ahead

Having looked squarely at how precarious the making side is, I still want to say this: ahead of this ground, a vast frontier spreads out.

Consider it. The toll on the shared language became zero (Ch. 5). The tools that help design are available free. Even the gate of manufacturing is getting lower (Ch. 7). And now, AI, a reassuring partner, turns the trial and error of design and verification with us, at high speed. A human thinks "what do I want to make," AI moves its hands, and finally real physics passes judgment. With this new shape of development, even without vast capital, wherever you are in the world, you can take on your own chip.

There has never been an age when so many doors opened at the same time. Untrodden land spreads out before us. Precarious, but wide. Precisely because there's risk, the view for the one who steps out first is exceptional. This is a rare frontier for those who take on a challenge.

F.4 So I decided to try building one myself

From here is my story, as the writer. In the preface, I wrote that I'm "someone who covered technology and business as a journalist." So, watching this upheaval, one question wouldn't leave my head. "Has the door really opened, all the way to ordinary individuals and engineers?"

And I came to think this: getting the feeling you understand from explaining it is half-baked. Whether the door has truly opened can't be settled by arguing on paper. Build a chip with your own hands, and you'll find out. The parts that go well and the parts where you stumble — all of it, honestly, I wanted to check. And so I made "courses for finding out." I stopped being a bystander who comments, and decided to step into the frontier as a participant.

F.5 To you ― the far side of this door

This corner has been the map of the world that leads up to that "let's try." From what a chip is, to the diversification of brains, the opening of the language, the move to 3D, the shift in how chips are made, and on to the frontier of Physical AI. By now, the map inside you should have widened considerably.

If, having finished reading, a small flame has been lit deep in you — "maybe I'd like to touch this a little myself" — then I've prepared the road ahead for that. From the board (the foundation), to logic circuits (FPGA), to the guts of the silicon (tape-out), it's four courses that build up step by step. A serious challenge record, with its ups and downs. But the frontier shows its view only to those who step out.

— Thank you for walking all the way to the end.
May this map be a small support for your first step.