There’s never been as big a shock 
to the semiconductor industry
as the current surge in demand from AI.
AI is driving incredible computing demand.
Easily 40-50% above our maximum 
capacity, which I’ve never seen before.
Progress in AI has been driven by 
exponentially increasing compute.
To sustain that progress, foundries like 
Intel will need to manufacture far more
leading-edge compute, both by producing more 
chips and by improving what each chip can do.
But the production of these chips is bottlenecked 
by tools that Intel buys from their suppliers.
Most famously, ASML’s EUV machines.
If they don’t have the equipment,
they can’t make the chips here.
I spent two days touring Intel’s fabs to
prepare for these interviews with its leadership.
I wanted to understand what it takes to
manufacture the world’s most advanced chips 
and how hardware is shaping the future of AI.
I’m joined by Chris Auth, who leads 
development of all the processes
that are currently under manufacturing.
The thing I want to understand in more
concrete detail is how the development 
of a new process actually happens.
If 14A is going to come out in 
three years or two years from now,
what is happening from now till then?
I want you to walk me through as much
detail as you can, everything from the level 
of “how is the yield improving?” to what
kind of telemetry you guys look at?
What kind of experiments do you guys run?
I want to understand what is going to happen over 
the next two years as concretely as possible.
14A has been in development for 
about three-plus years at this stage,
so quite a bit of pre-work has been done to 
get it to the stage that it’s at right now.
Right now, we’re getting very close 
to where we’re finalizing the process.
We’re finalizing the tool sets 
that it will be manufactured on,
and we’re working on the fine-tuning 
of each of the process steps.
There are about 2,000 steps in the 
factory that you have to go do.
You need to do really intricate work at each 
one of those to make sure it’s fine-tuned,
it can go into manufacturing, it can run 
thousands of wafers a week on this process,
and it can be repeatable every single time.
Right now, you might have a certain yield.
Walk me through how that 
increases month over month.
There’s a line running at 14A.
What happens from now till next
month that will result in the yield changing?
How do you identify where the defects are?
How do you fix that?
There’s a constant
process that we go through in identifying defects.
There are two main ways that we’ll identify them.
One is we have inline metrology.
We have tools that will look at
the wafers and say, “Ah, there’s a defect there, 
there’s a defect there, there’s a defect there.”
Then also at the end of line, either electrical 
test or we sort and see where the fails are.
We will identify failures, and then we 
will get analysis on those failures.
We’ll cross-section it.
You were in the TEM lab yesterday,
and you saw how you can get cross-sections.
You’ll get a cross-section of a defect,
and from that, you’ll be able to say, “Oh, 
that defect is between these layers, so it
must have come from this section of the flow.
Oh, we have inline metrology in that section,
which has identified defects.
Those are the defects we need to go after.”
What we’ve learned is that you don’t want 
to do too much innovation in one node,
because the more innovation that 
you do, the more complex it becomes.
While you’ll be able to get that to work, you 
need to hit a certain schedule, and that’s the
schedule that your customer ultimately wants.
Having that predictable schedule is a key piece.
You’ve got to figure out how much innovation 
you can put in in order to meet the customer’s
requirements in the timeline that they need it.
Intel learned how costly that tradeoff can
be with delays to its 10 nanometer node, 
whose original target launch date was 2016,
but which didn’t begin manufacturing until 2018.
Usually Intel has been the keeper of Moore’s Law,
a new cadence of manufacturing 
every two years, but their last
set of products slipped almost four years.
Intel today having its worst day in four months,
the stock now down more than 15%.
I also want to point out Taiwan
Semiconductor up 13% off of Intel’s bad news.
Walk me through what lessons you think Intel
has learned from 10 nanometer.
Ten nanometer was a good example
of where we had very aggressive targets.
We were doing what was called hyperscaling
at the time, so we were scaling much faster 
than what had been historically done.
In order to do that, we came up with all sorts 
of innovative ideas on how we could achieve that
without new tools, because 
the tools weren’t ready.
That put a lot of burden on the process 
and a lot of burden on the engineers.
Now, at the end of the day, 10 
nanometer is a great technology.
It has world-class yield, it has world-class 
performance, it has world-class cost,
but it took us longer to do that 
than what we would have liked.
So we’ve tried to take that learning into 
future technology nodes and say, “Hey, we
need to limit the amount of innovation we put in 
there to make sure that we keep it on schedule.”
What do you think was the main technology that 
should’ve been saved for a future version?
EUV was probably the big thing.
When we started to define 10 nanometer,
the conclusion was EUV was not ready.
EUV is extreme ultraviolet lithography.
It prints chip patterns with light 
at a wavelength of 13.5 nanometers,
about 14 times shorter than that emitted by 
previous tools that chipmakers had relied on.
The shorter the wavelength, the finer the 
features that a single exposure can print.
But Intel decided this technology 
wasn’t going to be ready in time.
TSMC transitioned to EUV much sooner.
It put EUV into volume manufacturing in 2019 and 
ramped it up further on its 5nm node in 2020.
Intel’s first EUV process did not reach products 
until 2023, so TSMC had about four more years
of experience with these machines.
Yesterday, we were at the research
lab getting a preview of 2D materials 
and CFET and ruthenium interconnects.
I’m curious about how you decide whether that 
technology is mature enough for the next node.
One thing that did also come out of 10 nanometer 
is we started to define a set of maturity metrics.
We use those maturity metrics in a process 
which we call phase gate, where we would say,
“To go from pathfinding into development, 
these are the criteria that we need to go hit.”
There are yield criteria, defect criteria, 
performance criteria, and we need to hit those
in order to move to the next set of the flow.
If there’s something that doesn’t make that,
then it misses that, and it 
goes on to the next technology.
We focus on the items that are ready and 
we can run and put into a process, because
one of the key things is staying on schedule.
Staying on schedule is hard partly because new
transistor designs take decades to mature.
Modern chip transistors evolved from the
planar MOS transistor, demonstrated in 1960.
The history of transistors since has been
the progression from two dimensions into three.
Each development that Chris goes on to describe,
from FinFET to gate-all-around, 
is a step in that progression.
So it’ll help to first look at a basic transistor 
and how its parts have changed over time.
A transistor is basically a switch.
When it’s turned on, it allows current
to flow between its two ends, which 
are called the source and the drain.
A piece called a gate sits on top of the 
silicon between the source and the drain.
Applying a positive voltage to the gate 
turns the switch on, because the positive
voltage pulls negatively charged electrons 
in the silicon beneath up toward the gate.
However, a thin layer of insulation stops 
them from actually reaching the gate.
So the electrons collect just beneath that 
insulation and form a thin conducting path
from the source to the drain.
This is called the channel.
Remove the voltage and the electrons disperse, 
the channel disappears, and the switch turns off.
As transistors shrank, the source 
and the drain got so close together
that current could flow through the channel 
even when the switch was supposed to be off.
We needed a better way to control the channel 
and prevent current from leaking through.
The fix was to give the gate a tighter grip 
on the channel by touching it from more sides.
In the original planar transistor, the 
gate only touches the channel from the top.
The FinFET transistor squeezes the channel 
into a thin vertical fin of silicon.
The gate drapes over the fin, so it 
controls the channel from three sides.
The next progression from there 
was to cover not only the top and
the sides of the channel, but also the bottom.
This technology is called gate-all-around, and
it slices the channel into thin ribbons, each of 
which gets completely wrapped around by the gate.
But getting gate-all-around to work required 
new materials and new manufacturing steps.
To build these floating ribbons of silicon, 
you first sandwich them between alternating
layers of a temporary placeholder material, which 
holds the space where the gate will eventually go.
We use a placeholder because the materials in 
the real gate wouldn’t survive the processing
that comes after, which involves baking 
the wafer at around 1,000 degrees Celsius.
Once these hot steps are 
done, chipmakers etch away the
placeholder, leaving the ribbons suspended.
Then they fill the gaps around every ribbon
with a thin coat of a special insulator called a 
high-k material, whose purpose is to let the gate
pull more strongly on the channel.
And then on top of this material,
they put the metal of the gate itself.
FinFET was demonstrated in a lab in 1989
and reached volume production at Intel in 2011.
A gate-all-around transistor had been demonstrated
a year earlier, in 1988, but Samsung didn’t 
begin producing gate-all-around chips until 2022.
So these ideas, FinFET and gate-all-around, 
were developed around the same time,
but they reached production a decade apart.
I wanted to ask Chris why it took so long to
get to this current gate-all-around technology.
Maybe you can walk me through the process of
getting into high-volume manufacturing.
I want to understand concretely,
once you have an idea that, “Hey, we’re going 
to make the channel smaller, we’re going to
surround it…” All the other problems of material 
science, figuring out how to not destroy the gate,
how do they actually get figured out?
Why does it take so long?
Gate-all-around’s a good example, because 
back in the ’90s we knew of this idea.
But we knew there were going to be 
these problems on how you put this
into high-volume manufacturing.
At the same time, there are other
things like high-K metal gate.
There’s strained silicon,
there’s FinFETs that are out there.
It’s a question of when’s the right
time to go and tackle those?
Well, you don’t want to
tackle something before you have to.
If you have these other things that can drive
performance, you would go and do that.
So we went through this path.
We did strained silicon, high-K metal gate, 
we did FinFETs, and then it became, “Well,
we need to do things further.
How do we get there?”
Well, gate-all-around’s really 
the ultimate transistor out there.
Maybe it’s time to go and start tackling 
some of those issues that you have.
You’re right, now you have to look at how do you 
put high-K metal gate in with gate-all-around.
It’s that innovation that takes time.
Sometimes you need new materials.
Sometimes you need a different tool 
in order to be able to do some of the
things that you need to go do, and that 
requires invention from the tool vendors.
You have to go to the tool vendor and 
say, “Hey, I need a tool that can do XYZ.
I need a gas that can deposit this 
particular material that goes into that.
I need a heating chuck that 
has this characteristic to it.”
All those things take time.
Some of it’s iterative,
where you have to run a gate-all-around 
transistor and say, “Oh, well, that didn’t work.
Let me figure out what was 
wrong with my hypothesis and
try it again with some different structure.”
How has the nature of R&D changed from the era of
just making planar transistors smaller and smaller 
and getting the benefits of Dennard scaling,
to the world we’re in today?
Is it just a totally different
kind of R&D process?
Previously it was,
“Let’s make things smaller, smaller, smaller.”
Now you look at all these other facets
of performance at the same time, and 
improvements in size don’t come as easily.
Walk me through whether there was some crucial 
change in the R&D process at that point.
If you look at the ’80s and ’90s, yes, it was 
driven really by Dennard scaling, lithography.
We just made things smaller, transistors 
got faster, they got more reliable,
and it was a very predictable cadence.
When we started to hit the 2000s, we saw,
“That’s not good enough anymore.
We need to do something different.”
To me, it became this golden age of innovation.
For each technology node,
how are you going to provide more scaling?
How are you going to provide more performance?
It drove things like strained silicon.
It drove high-K metal gate.
It drove FinFETs. It drove gate-all-around.
It drove backside power.
These are things that in the ’80s 
and ’90s you would be like, “Oh no,
we’re not touching any of that,” because we 
just continued to have the same process flow.
We just scaled it, scaled it, scaled it.
But for somebody like me who’s a
transistor junkie, it’s awesome, 
because now the playbook’s open.
You can do all sorts of things out there.
You can bring in new materials.
Just look at the number of materials we used 
to use in the ’90s versus the materials today.
It’s probably gone up 10X.
There’s a lot of innovation.
There’s atomic layer deposition.
You’re controlling this stuff atom by atom.
So it’s changed a lot how 
you approach the process.
Before, it was, “Try to not change anything.”
Now it’s, “Take the sledgehammer and start tearing
the thing up and come up with some new idea.”
For 14A, Intel is using a new kind
of lithography machine: High-NA EUV.
It uses the same extreme ultraviolet light
as today’s EUV machines, but changes the optics.
NA, or numerical aperture, is a measure of how
wide a cone of light the machine’s 
mirrors can focus onto the wafer.
The wider the cone you can focus, the 
smaller the spot that your light can carve.
Standard EUV machines have an NA of 0.33.
High-NA widens it to 0.55.
So High-NA can print features about 
1.7 times smaller in a single pass.
When features are smaller than the lithography 
machine can print, chipmakers have to split
the pattern across several passes, each 
printing a different set of features.
The most tightly packed patterns can 
require two, three, or even four passes.
Each additional pass takes up valuable time 
on a machine that costs around $200 million,
and it introduces another opportunity for defects.
High-NA machines cost about twice as much,
but they can reduce the number 
of passes that are needed.
So the question which determines 
whether they justify their higher
price is how much more time they save and 
how much they reduce the risk of defects.
Now you guys are using High-NA EUV 
tools for 14A, whereas TSMC has said
that they don’t make economic sense yet for them.
How do you decide whether the next litho tool is
the one that is ready to adopt or not?
High-NA’s a good example of where,
because of the architecture of our technology 
versus TSMC’s, we make different conclusions
off of the same tool set.
We use backside power.
That means that you actually alleviate some of 
the pitch requirements you have on the front side.
That allows you to print 
that directly with High-NA.
TSMC has to run tighter pitches 
because they don’t have backside power,
and they can’t do that with High-NA, so 
they have to use double patterning EUV.
So for them, it’s not very helpful.
For us, it is very helpful.
Now at 14A, we’ve left it as an option, and 
it’s something that we’ll have a discussion
with the customer on, because there’s some 
value proposition to it, and there’s some
complexity that it adds.
Chris is referring to
a feature called backside power.
A chip has two different kinds of wiring.
Power lines deliver electricity to the 
transistors, and they need to be thick.
Signal lines let the 
transistors talk to each other.
They’re much thinner, but they need to be spaced 
apart so that the signal from one wire doesn’t
interfere with the signal from another.
In the past, all of these power lines
and signal lines had to be crammed 
together on top of the transistors.
Intel’s newest process moves these power lines 
below the transistors, to the back of the wafer.
This frees up space on top, so the signal 
lines don’t have to be packed as tightly.
This change makes High-NA worth it for Intel.
Without backside power, the wiring is so
tight that it would require multiple 
passes to print, even with High-NA.
With backside power, a single 
High-NA pass is enough.
Outside companies can now commission Intel 
to manufacture chips that they’ve designed.
They follow Intel’s process design kit, or PDK: 
the menu of what a given process can build,
and the rules for building it.
Do you need to lock that into
your PDK right now for customers who 
are planning on developing on 14A?
It’s an option in the PDK today, and we’re 
handling it as a customer-by-customer
discussion as to whether or not they 
want to go down that path or not.
For 18A, there was public reporting that 
there were delays in or issues with the PDK,
such that some instructions didn’t work, 
and it was difficult for people to use
your foundry to build their chips.
Tell me about the process of Intel
becoming a foundry and then having a 
PDK that other customers can use to make
whatever chip they’re making as well.
There’s definitely a learning curve to
how you become a foundry.
That’s something that our
competitors have gone through as well.
PDK is a good example of that.
For decades, we’ve been an IDM.
We’ve gone to our product teams
and shown them our transistor data, and we 
share that data back and forth, and we haven’t
had a formal PDK-type process.
So that’s a new process that
we’ve had to implement on 18A.
18A started out not having a PDK.
There’s definitely a lot of feedback we’ve 
gotten from the customers as to what they expect
for quality and maturity of the PDK and what 
sort of information they want to see in there.
We’ve tried to incorporate that into 
making sure that we have a PDK that
meets our customers’ requirements.
That’s where we’ve ended up.
We’re starting to get PDKs that have this, 
and 14A will have that from the get-go.
It has very high quality PDKs.
Again, there are maturity metrics for
what is a PDK 0.5 versus a 0.9 versus a 1.0.
We’re constantly going through and taking
customer feedback and incorporating 
that into the quality of our PDK.
For most of its history, Intel’s 
fabs made only Intel’s chips.
That could be part of what 
allowed TSMC to gain an edge.
Serving many customers gave TSMC more wafers 
to learn from, more revenue to reinvest,
and more pressure to hit deadlines.
Basically, all enterprise customers
have a shared set of product requirements 
that you have to meet if you want to even
start talking about doing business with them.
These are things like SSO, so employees can
log into your app with their company accounts, or 
directory sync, so that whenever someone leaves,
their access gets shut off automatically.
These features are harder to get right than
you think, and the risk-reward for building 
them in-house doesn’t really make sense:
a slightly better version of SSO 
probably won’t accelerate your growth,
but a broken one will definitely stunt it.
WorkOS gives you access to battle-tested versions
of all the most common enterprise features—things 
like SSO, SCIM, RBAC, audit logs, and more.
WorkOS also has versions of these features that 
are purpose-built for agents, so they can log in,
grab the right permissions, and 
take actions within your product.
All of these features can be plugged in 
with just a few lines of code, so you can
add them and get started selling to enterprises.
OpenAI, Anthropic, and Thinking Machines—as well
as more than 3,000 other companies—all rely on 
WorkOS to meet their enterprise requirements.
If you want to join them, go to 
workos.com and tell them that I sent you.
If a butterfly flapped its wings in the ’60s, 
could you imagine a world where gate-all-around
is not even a word people have heard of, an idea 
people are familiar with, but people are thinking
of an extremely different branch of the tech tree?
How far down the tech tree do you think
things could be totally different?
Could it be that we’re not even using copper?
If you look at, say, transistors and 
interconnects, where we’ve gotten to,
as far as the transistor and gate-all-around 
and it being the ideal transistor structure,
was somewhat inevitable.
As I said, in the ’90s we were
already talking about these types of things.
Backside power, that one’s pretty out there.
Even in the ’90s, the idea that you’re going to 
strip off the entire silicon substrate and put
lines on the backside, that one is probably 
something that’s a little bit out there.
Copper, of course copper wires are the 
best, so I don’t know how you would end
up somewhere else besides copper.
But you guys are exploring other—
For some of the very tightest pitches, it’s 
inevitable that you will replace copper.
That’s a good example of where you will have 
another material that will come in there.
But for the larger lines, I think 
copper is about where it’s at.
Once you start to do gate-all-around 
transistors and you do backside power,
now you start to enter a very different regime.
Now you see it’s not, “Hey, there’s just some 
substrate and you stack everything on top.”
Now you’ve got a transistor which is floating 
in interconnect, and you’re like, “Okay, well,
why don’t I have transistors, interconnect, a 
little more transistors, more interconnect?”
This whole 3D stacking that we now do with just 
discrete chips that we stack on top of each other.
Now you’re like, “Why can’t I just 
build an entire structure like this?”
You see it with memory, NAND, DRAM.
They have some incredible 3D structures.
NAND already does it, DRAM talks about 
it, but I think that DRAM inevitably
is going to go there too.
A lot of people say, “Oh no,
we’re not going to do this because we 
can’t do it in the next five years.”
Well, come on, let’s put a 
larger timescale on this.
Let’s put 10 years, 20 years.
In 2045, do you think you could do it?
Now you start to say, “Oh yeah, 
I could see us having some pretty
exotic structures at that point.”
We limit ourselves by what we see
as possible in the next four years.
As I’ve always said, when people say,
“How do you see technology scaling going?”, 
we have visibility for about four years.
Then you get out to the four years, and 
you go, “Yeah, I think there’s something
we can do another four years out of that.”
And you get to that four years, and you’re like,
“Yeah, I think you can do something.”
But if you start to talk about 10 years,
it becomes hard to say what it’s going to be.
Chris’s idea of building upward could start
with the transistors themselves.
Chips use two types of transistor,
called n-type and p-type, which work the 
same way, but with the charges reversed.
Nearly every circuit is built by pairing them up.
Today, chipmakers build them side by side.
The next step is to stack them directly on 
top of each other in order to save space.
In principle, that fits twice as 
many transistors into the same area.
This design is called a CFET, short for 
complementary field-effect transistor.
But every extra layer adds manufacturing steps 
and cost, and also leads to heat building
up that doesn’t have a clear way to escape.
Because logic consumes more power than memory, you
can’t do stacks of DRAM that make HBM or 3D NAND 
as easily with logic, because of the heat issue.
Is that the fundamental issue over the long run?
How you—
Thermals. There are three things you have when 
you try to say, “Hey, what limits compute?”
It’s how fast can you do the switch, 
so the frequency of the device.
How much power does it take in order to do 
that switch, so the power efficiency of it.
The third one is, how do you dissipate the 
heat that was generated by doing that switch?
That’s the thermals piece of it.
These are not new problems.
You can go back 20 years and you can 
see articles on these three things.
The question is, thermals… how do you 
deal with that, particularly when you
start to talk about these 3D-type devices?
There is a need for that sort of system-level
analysis and simulation tools.
I think we’re in its infancy
because we’ve just started to look at this.
It is a limiter, and it will be a limiter, but
again, these are problems that we can go solve.
The problems respond to work, but engineering time
is finite, and so are the wafers for experiments.
Deciding where they go across thousands
of problems is itself one of the 
hardest parts of bringing up a node.
I sat in on one of Intel’s Hot Topics 
meetings, where those decisions get made.
I’m here with Kevin Fischer, who was laying down 
the law this morning at the Hot Topics meeting,
where you guys were discussing how you’re 
improving 18A and 18A-P technologies.
People would come to you with different 
experiments they were running,
different ideas they had for improving 
yield and performance, and then you and
your integration engineers would tell them 
what data you needed, how they should proceed,
how you think they could move things faster.
I want people to have a sense of
the culture at these meetings.
One thing I found striking is that this
is bottlenecking your most advanced node.
You want progress to happen fast,
so you kind of lay down the hammer.
I would say the average engineer doesn’t
understand the stakes or the urgency of how 
this is going to impact Intel’s bottom line,
how it’s going to impact our customer.
They just have no visibility into those issues.
They will typically have a certain set 
of criteria that they want to go capture.
They’ll hit boundaries of, “I asked for 
something and someone said no, and I’m done.
I guess I can’t have this thing that I asked for.”
One thing you learn in this industry is that
you will never have a complete data set.
If you want to wait for a complete data set,
you’ve waited too long.
It’s already too late.
A lot of what I end up doing in those 
meetings is deciding how much risk we’re
going to take on a certain decision 
or not take on a certain decision.
You need to say, “Look, do we go now?
Do we start risking production?”
Because you’re risking the bottom line.
You’re risking revenue.
I’m worried you’re going to put me in 
the light of being an ogre because I’m
putting the hammer down on people.
That’s a great reputation to have.
It’s really conveying urgency.
Every day counts.
Can you walk me through a 
particular issue that might come up?
One of the examples you saw 
this morning is epi nodules.
Epi is short for epitaxy, which means growing 
new material on top of an existing one.
Chipmakers flow gases over a heated wafer, 
and the gases leave atoms on its surface.
The crystal structure of the material 
underneath acts as a template.
Each new atom lines up with the atoms below it, 
and the new layer continues the same crystal.
Chipmakers use epitaxy at 
several stages of making a chip.
In the step Kevin describes, they use it 
to grow the source and the drain as small
crystals on either side of the gate.
During the deposition of our epitaxy,
we get abnormal growth, a 
little thing like a tumor.
It’s growing super fast.
That causes problems,
because one epi will short to the wrong 
epi and you get transistors not working.
Then you’ve got to narrow it down 
to how you’re going to fix it.
For the past decade, Intel’s chips have 
had this feature called Processor Trace.
When you turn it on, it starts recording all 
the code your program runs, and it captures
enough detail that you can go back later 
and reconstruct every single function call.
Most companies don’t need that level of detail.
But Jane Street does: they need to investigate
specific, fleeting moments in 
their trading systems—things like
a delayed response to a network packet.
Those events tend to be too brief for
most debugging tools to catch.
So when one of Jane Street’s
engineers noticed Processor Trace, 
he started playing around with it.
And then he and a summer intern turned that 
exploration into a new tool called magic-trace.
Magic-trace leverages the output from 
Processor Trace to give Jane Street a
history of every function call leading 
up to a specific moment of interest.
This project was completely bottom-up—it 
started just as something that one
engineer thought would be valuable.
And now, traders and developers across
Jane Street use magic-trace every single day.
If you want to work somewhere where you’re trusted
to pursue the ideas that you think matter most, 
Jane Street is hiring researchers and engineers.
Go to janestreet.com/dwarkesh to learn more.
One of the pieces that you observed this morning
was a special clean that we were looking 
at, which would eliminate the nodule.
The nodule would deposit, 
and then we’d get rid of it.
Some of the other discussion was 
about the source of the nodule.
Get rid of the source of the nodule.
There’s some issue. How do you
decide what the source of it is?
For the most part, we work highly parallel.
How do you decide?
It’s not always evident
what the easiest way to fix the problem 
is, so you work on all the potential fixes.
Sometimes it’s not the one you expect first.
Once we’ve identified it as a primary problem
for the technology, that’s where this 
Hot Topics meeting gets involved.
That’s how we align the resources 
into those engineering teams so that
we can accelerate the work.
You said that there’s some
fast-track process where wafers can move through 
faster if they’re bottlenecking some experiment.
We have a tiered system of what 
we call hot boxes in the factory.
Every grade of hot box goes a little bit faster.
It’s just how the factory prioritizes that
particular lot, that block of silicon 
that’s running through the factory.
Our highest grade of hot box, 
HotBox 1, races through the factory.
It gets priority at every single tool, 
but it’s a drag on the entire factory.
If you had a factory person in here, 
they don’t like playing that card,
because you end up using the tools 
inefficiently so that you can go super fast.
Think of it as a highway, and you set 
aside that one lane for the carpool,
which everyone could use, but you’re 
saying, “Okay, it’s just the carpool.”
That’s our HotBox 1.
Everyone clear out of
that lane except for that one experiment, and 
that one experiment races down the highway.
Kevin’s job is deciding what goes onto the 
production line now, but the day before, I
went to Intel’s research lab with Jack Kavalieros 
to see ideas that aren’t ready for the node yet.
We started with wiring.
A transistor is only useful once it’s connected
to others, and on a modern chip that takes 
more than a dozen layers of metal wire stacked
above the transistors, called interconnect.
Those wires are made of copper, and as they
shrink, copper starts to run into trouble.
Intel is exploring a replacement: ruthenium.
Ruthenium has higher resistivity, so why are 
you guys exploring ruthenium for the future?
Interestingly enough, as you scale the dimensions 
of those copper wires, you run into challenges.
For example, copper typically requires a liner in 
order to deposit the copper to fill those lines.
Ruthenium is very different in that regard 
because it’s a material that we can actually etch.
As such, by etching it, we don’t need that liner.
Now you scale those very fine dimensions,
and if you can trade off that resistive 
liner with no liner, ruthenium starts to win.
There is a crossover point where 
ruthenium will have much better
conductivity, better resistivity than copper.
That’s roughly in the 20-nanometer pitch range,
which means, if you can imagine, half of the 
pitch is your metal line, and the other half is
the dielectric that isolates them.
So what is the liner?
Think of it almost like a glue layer or transition 
layer that we need in order to deposit the copper
and have it stick, and have it make that 
connection to the copper lines below it.
These are all interlocking copper 
lines that interconnect to each other.
But every time you deposit copper, 
you’ve always had a liner underneath it.
In the past, that liner 
may have been a significant
portion of the entirety of that conduction path.
But over time, we’ve scaled it, made it thinner,
and so copper comes back into play as a contender.
One thing I’d be curious to know is how you
decide that a technology is mature and 
ready for high-volume manufacturing.
With 10 nanometer, you were mentioning 
you guys moved on from copper to cobalt,
and there were potential issues with that.
How do you decide that this
technology is going to work at scale?
The integration of these materials into such
a complex system as the transistor ecosystem 
involves many, many different components.
The scalability is one part, but 
its ability to integrate is another.
At the time we made a decision 
for cobalt, it was the forerunner.
The results were stronger at the time.
But as we work with our ecosystem externally,
the vendor ecosystem that provides these materials 
and the tools that we integrate them with, we find
solutions down the road that may bring copper 
back in line as being the stronger of the two.
That’s exactly what happens.
Every technology that we go through,
we will evaluate multiple options.
As we get closer to our time horizon
for delivering a manufacturing-worthy 
product, we have to make a decision.
The leading option at the time was cobalt.
Now it’s copper again,
and perhaps in the future it’ll be ruthenium.
I’d love to get your thoughts on CFET, basically
stacking the transistors on top of each other.
What is progress like there?
Can you keep making more and more layers?
Is it just two?
Can you go three, four?
How high can you get?
That evaluation of CFET is currently underway.
Not just by us, by our competition.
Labs externally are looking at it.
It is the ultimate in stackability.
We’re taking transistors that had 
to be compressed more and more,
and we’re relaxing that in some ways because 
we’re going to stack them on top of each other.
But the act of stacking does not 
mean that they’re not interconnected.
What used to be connections on 
the top or even on the bottom,
as we’ve seen from our press release, now has 
to go vertically in between the transistors.
That’s a new challenge because the aspect 
ratio of creating trenches and filling those
trenches becomes more and more difficult.
It’s not as straightforward as stacking the
transistors, because you still have 
to think about interconnecting them.
Is the issue there the etch to get that deep hole?
Aspect ratio for the etches is a big part of it.
We’re working with our vendors to evolve on 
that and to improve that, but it’s also what
you fill those very tiny spaces with.
Because filling can sometimes leave
air gaps or holes that are never 
really filled, and that’s not ideal.
I assume there’s a version of 
CFET that works inside this lab.
What is the kind of wafer or 
process that is working right now?
Then what’s the next step?
What’s the next step after that,
until it’s inside the full high-volume process?
Part of the process of evolving these
technologies, especially something more complex 
like an architecture — not just a basic chemistry
or a material, but a full architecture like CFET 
— is that we have to make short-loop vehicles
that can give us faster turnaround.
The normal manufacturing process can
take a month or longer to complete the 
thousands of steps that it goes through.
Our goal here is to take unit cells 
of that process and qualify them.
We take advantage of short-loop vehicles 
that can give us electrical parameters,
chemical parameters that can resolve some 
of those challenges, those speed bumps.
Once we’ve resolved them, then we combine 
them into a more complex test vehicle.
What you would find in here 
is something very simple.
You might be making a diode, you might be making a 
capacitor, which are components of the transistor.
As those evolve and become stronger, better 
performing, you combine them to make the next
stage, which might be the transistor itself.
In the CFET example, you might not choose to
do four ribbons on top of four ribbons.
You might just choose to do one on one,
because naturally the aspect ratio 
of dealing with something that
much taller becomes more and more challenging.
So we would have a simplified version of a CFET,
which has only one wire for the 
NMOS and one wire for the PMOS.
That simplifies our job, but it still 
helps us evaluate the stacking of devices
and how they will interconnect.
But are the feature sizes here as
small as they will be eventually?
Or are the feature sizes bigger?
Oh, absolutely. In some cases, at least as small.
For example, you’ve heard of 2D materials,
these transition layer 
dichalcogenides, as they’re called.
They’re very, very thin materials.
They’re essentially one monolayer
of atoms that we deal with.
Those depositions can happen in this lab.
We can evaluate those films in terms of 
conductivity, our ability to contact them,
our ability to modulate that conductivity, 
which is what the transistor does.
So yes, we do have the ability to not only 
deposit films that are very small and thin to the
atomic scale, but also print very small features.
We have EUV technology, or E-beam technology, that
can allow us to print very small features similar 
to what we will eventually make in production.
I guess I don’t understand intuitively, 
lacking the context in this process,
if you have the features that are as 
small for a working CFET right now.
Why will it take 10 years 
for it to be in my phone?
I understand that there’s a process, 
but why is it such a long process?
The answer is slightly different when 
you’re dealing with an architecture
like CFET versus a novel material.
Materials have, of course, the
challenges of being deposited, etched, removed.
Their compatibility with other layers above and
below, all those things have to be researched 
first before you put them in a production line.
An architecture may seem like a simple evolution.
“Oh, you’re making nanoribbon
transistors next to each other.
What’s wrong with just stacking them?”
But the act of stacking them, of course, increases 
the aspect ratio for etches, depositions, cleans,
and the technology may simply not be ready yet.
But if you’ve already made it here, there’s a
version of the technology that’s ready, right?
Correct. We’ve made a simplified version that
may not be an advantage over our current 
side-by-side technology of nanoribbons,
but we need to evolve it to make it maybe 
two nanoribbons, or to add the right type of
metals and dielectrics, in order to extract 
the maximum performance out of that device.
Got it. So it’s like a long tail of—
There’s a thousand different steps that
all have to come together in a very unique way 
to make the performance compatible or actually
supersede the prior technology.
Where do the ideas tend to fail?
You have 20 ideas like 2D materials.
What is the way in which you decide
this is actually not worth pursuing further?
Sometimes we discover fundamentals about the
material that will be incompatible with 
the rest of the processing that we do.
The first level of experiments may 
simply be very simple depositions.
You say, “Okay, I can deposit the film, 
I can look at it, I can measure it.
But the moment I try to make a 
contact to it, nothing sticks.”
It could be something as 
fundamental as the adhesion.
The next step in the process simply 
has no way to adhere to the material.
So you may rule out a material because 
of something as simple as that.
More complex challenges might be 
that the subsequent processing may
have gaseous interactions with downstream steps.
For example, I’ve deposited the film, I’ve capped
it with other materials, but they’re not hermetic.
So my third or fourth or fifth step in the
process may actually still interact 
with it and change its properties.
That makes it unstable, and now we have to 
find better hermetic sealing for that process.
If we don’t find a way to control 
it, then it’s a dead end for us.
Stacking transistors is still in the research lab.
Stacking whole chips is already in production,
and it tackles a different problem: how 
separate chips communicate with each other.
Modern computers are divided into multiple 
parts, each optimized for a different task.
For example, Nvidia’s Blackwell data-center GPU 
joins two huge processor chips and surrounds them
with eight stacks of high-bandwidth memory.
These chips have to constantly send
data back and forth with each other.
And in AI, the speed of that traffic,
not the raw computation itself, is often 
what limits how fast the workload finishes.
Imagine two chips side by side, 
communicating through their facing edges.
If you double each chip’s width and length, 
you’ll have four times the area for circuitry,
but only twice the edge length for connections.
This is the area-versus-perimeter problem.
As transistors become denser, a chip’s 
ability to do computations increases
faster than its ability to move data around.
The industry has known about this problem
for years, so multiple technologies 
have been invented to address it.
The completed circuits must undergo two 
last operations: bonding and packaging.
Once a chip is made, it still has to be connected 
to other chips and to the rest of the computer.
That step is called packaging, and it has 
changed a lot in the last few decades.
At a high level, you can think of the 
evolution of packaging like this: initially,
chips were laid out side by side.
Wires connecting them went out
the side of one chip to another.
To allow for more communication
density, the industry introduced a 
shared piece of silicon, called an
interposer, for multiple chips to sit on.
A chip can talk to other chips through the
wiring that runs directly through the interposer.
This type of packaging is often called “2.5D”.
Another way to increase connection density 
is to bond chips directly to one another.
This allows for shorter connections between chips, 
but the early versions of this idea used solder
bumps, which are bulky and increased resistance.
It would be better to skip the solder and fuse
the signal wire connections of the 
two chips directly to each other.
This is called hybrid bonding, and 
it is one of the technologies the
semiconductor industry is investing in the most.
Hybrid bonding works by aligning the copper
interconnects of one chip with those of another.
Then these chips are pushed together and heated
so their interconnects expand and 
bridge the gap between the chips.
This process makes the insulating 
surfaces around the copper bond as well.
And so by removing the solder bumps, 
you end up with much denser connections.
But the reason this technology is only now 
entering volume manufacturing is because
it took a while to iron out the alignment 
and defect challenges that it introduced.
In the packaging lab, I asked Tyler Osborn, 
a director in Intel’s advanced packaging
technology development group, 
how they put this into practice.
Intel has spent over 50 years 
optimizing its CPUs for speed.
But lots of these optimizations—like 
caching and branch prediction and
speculative execution—reduce predictability.
The exact same instructions can take four
cycles to execute one time and 300 the next.
When you stack billions of instructions together
and execute them across multiple threads and 
cores, these non-determinisms compound into
all sorts of hard-to-reproduce bugs.
If you work on serious systems,
you’ve definitely experienced this.
Antithesis solved this by doing something a bit
crazy: they built a 100% deterministic computer.
They identified every single door in Intel’s x86
architecture where non-determinism could 
leak in—network packets, disk reads,
temperature sensors, even time itself.
And they either welded each of these
shut or slotted in a controlled simulation.
This deterministic computer is just a foundation.
On top of it, Antithesis built a 
testing platform that compresses
years of production hardening into minutes.
It runs your software through thousands of
parallel worlds, injecting faults 
and intelligently steering each one
towards a one-in-a-billion failure that 
you’d never be able to test yourself.
This determinism is what’s behind all the 
other insane Antithesis features that I’ve
mentioned over the past few episodes.
It’s what lets you reproduce any bug
and time travel to any point in the test run.
Most importantly, it’s what lets you—or your
agents—trace every failure back to its root cause.
Go to antithesis.com/dwarkesh to learn more.
Two different ways you could do hybrid 
bonding: die to die and die to wafer.
Maybe you can talk about the advantages of both 
and which one actually practically works better.
Sure. Around Intel, what I think 
you’re referring to as die to die,
we typically call wafer to wafer.
We actually take a full wafer and
another wafer and bond them together.
That’s an approach that’s really popular
right now in the CMOS image sensor industry.
In fact, it basically dominates it.
It’s being looked at for other 
things like memory stacking.
People have published ideas 
about all kinds of stuff.
One of the challenges with that, though, is that 
you need dies that perfectly match each other, and
are perfectly left-handed and right-handed, so you 
have to be able to bond exactly one to the other.
One-to-one die size is a challenge.
If they’re not the exact same design
device on top and bottom, that’s a problem.
With wafer to wafer, it’s challenging to make
sure that you have wafers that match 
really well in terms of the sort.
So sort good array here 
versus sort good array here.
If they match exactly — good die and 
bad die all match — I can get 100% yield
of that operation of the good dies.
But the chances of that are very low.
There’s going to be some randomization of 
the two, and whatever that randomization is,
will be loss of good bonding to bad.
Whereas with chip to wafer,
I can dictate the terms.
I can singulate and have known good dies
that I’ve tested at the chip level, and then only 
bond known good to a known good base position.
Do you think then chip to wafer is the way 
that you’ll do hybrid bonding in the future?
I think we’re going to do both.
We definitely have the capability to do both.
It’s really customer and architecture 
driven, but certainly chip to wafer in
the logic-to-logic space makes a lot of sense.
You were explaining that going from advanced
packaging to hybrid bonding, the 
size of the pitches is decreasing.
How much further could that go?
That’s a wonderful question.
Historically, like we were saying, 
15 years ago first-level interconnect
was 130-150 micrometer pitch.
Today in HVM, we’re doing 25
micrometer pitch in Foveros.
Here we’re doing sub-10,
but there’s already been literature out there.
In fact, the biggest conference every year
in packaging is the IEEE ECTC 
conference that happens in May.
There was a paper there this year 
showing reliability data on a
0.45 micron pitch hybrid bonded device.
So you can imagine that’s a whole other
order of magnitude in density again 
from where we are right now in HVM.
That scalability has already been proven.
So I see a long runway there.
Within interconnect, within a single 
wafer, how far separated are they usually?
In the back end of the fab, you can imagine the 
few layers below packaging are typically around
one micron to a half micron in size.
So if you can push packaging down
to 0.5 microns or below, you can almost think 
about taking layers off the top of the die and
connecting a lower layer within the silicon 
to another lower layer within the silicon.
That’s really where the beauty 
of hybrid bonding happens.
You could think someday about even having 
reduced interconnect layers on both sides
that we can bond together and have less total 
IR drop between the two transistor layers.
It’s really cool to think about where that 
could go once we get into those regimes.
Actually, I am very curious about…
For AI, it makes sense why hybrid
bonding is so powerful, because then you 
can get higher bandwidth to the memory.
But you were saying there are 
issues with heat and everything.
What is the advantage for non-AI 
applications of hybrid bonding?
Excellent question. There are companies 
out there that are looking at it for
non-AI applications, even in client devices.
One of the big things I think it can help
with there is, number one, obviously the 
same raw interconnect density, especially
between something like a CPU and a GPU.
But the other one is around power.
Like we talked about before, as you 
get to finer and finer interconnects,
you can reduce the total power consumption, 
in picojoules per bit, for that interface.
With hybrid bonding in those applications, 
you could think about that as another way
that packaging can help reduce 
the total system power budget.
That gives you lots of benefits: lower 
power, less heat, for the same benefit
in terms of actual electrical performance.
So far, this has been about making better chips.
The other half of the question is how many 
chips the industry can make if AI labs are
willing to pay for far more of them.
I’m getting a chance to talk to
Pushkar Ranade, who is the CTO of Intel.
I’m very interested in understanding how
the fact that the entire industry has to 
rely on the same suppliers bottlenecks
your ability to ramp up production.
For this EUV tool, ASML produces
fewer than 100 of them a year.
They don’t plan to produce more
than 100 a year, or significantly 
more, even by the end of the decade.
That has to be split between you, between 
TSMC, between all the memory makers who
are ramping up production, like SK Hynix.
How does that limit your ability to scale
up to all this demand?
It is a limiter.
We have to make the most of the 
capacity that we already have installed.
We’ll try to increase the output 
from our existing fleet of tools.
But EUV or ASML is not the only bottleneck.
It is definitely a bigger bottleneck than
some of the other tools, but 
to a large extent, we’re seeing
all equipment seeing tremendous supply shortage.
Even though there are multiple providers of other
tools, once you lock into a process flow 
— let’s say 18A, which is a mature process
now — we need a specific tool.
We can’t just go and get a
different tool for the same process.
So we’re locked into that one tool,
one supplier, and we rely on that.
What lessons does this give for
14A in terms of how you structure it?
By 2028, it’s going to be such a crazy world.
How do you structure things in such a 
way that you have flexibility in ramping?
The way to do it — and we’ve been doing 
it, and generally everybody does it — is
you try to maintain as much fungibility as 
possible going from one node to the next node.
Going from 18A to 14A, maybe you would 
say 80% of the tool set can be reused.
Then, of course, 14A would require some new 
capabilities because it’s a denser geometry,
so it would require some unique tools.
Do you have a sense of what the ultimate
bottleneck is on how many wafers you guys could 
produce a month, even if you were to decide
to invest infinity money into doing CapEx?
Given the existing constraints, let’s say we had
to maximize the output from a given footprint of 
tools, then I think it’s incremental, because the
process today is already yielding pretty high.
If your yield is already pretty high,
getting every 1% extra yield is extremely hard 
because it’s already a finely tuned process.
Are there enough tools in the world so 
you could 10X the amount of wafers you’re
producing at the leading edge?
Not 10X, but probably 2X, yes.
That’s quite interesting, because 
the AI industry for a decade has
been 3Xing compute year over year.
You’re saying that just can’t continue.
Unless the entire supply chain — starting 
from the equipment suppliers, their suppliers,
their sub-suppliers — unless everybody 
produces more, you can’t increase the output.
Just to give the audience a sense of how big 
the AI demand is and how well it monetizes the
products you guys are producing: you can crunch 
the numbers, and the numbers seem to suggest
that just a couple billion dollars of CapEx on 
Intel’s account results in enough chips — or the
logic for enough chips — to produce, say, $100 
billion of revenue for the AI model companies.
Of course, there are so many 
other players within that stack.
There’s also memory and other things.
But it’s crazy that a couple billion
dollars of CapEx, which can be amortized over 
many different years, produces $100 billion
of revenue every single year.
That seems to suggest that at
some point that should percolate down.
At some point, the AI labs could just say,
“We’ll just buy everything.
Produce as much as you can.
We’ll guarantee future income streams.”
There are early signs that could
actually be happening already.
The demand signals are just so high.
But I think you were alluding to how the value is 
accumulating at the top of the technology stack.
What is essentially happening is we moved from 
a fully vertically integrated industry 30-40
years ago, maybe longer, to a very horizontal 
industry, where you have foundries, you have
equipment makers, you have their suppliers.
Today, every hyperscaler is designing their
own chips, and the scale of 
the hyperscalers is so large.
Some of the model companies, for 
example, their demand is so high
they could easily consume an entire fab’s worth 
of silicon on their own, or multiple fabs’ worth,
and they definitely have the capital to fund that.
So in some sense, we are moving from a horizontal
world again to a pseudo-vertical world.
It’s not monolithically vertically integrated,
but it’s kind of virtually vertically integrated.
What is the nature of economies of
scale in this business?
If you look at how many
wafers at the leading edge TSMC versus Intel 
is shipping, it’s upwards of 3X for TSMC.
What kinds of advantages do you get 
at this point from economies of scale?
Or is it that once you reach a certain 
point, the economies of scale get weaker?
That’s a very important question.
This entire business just
relies on economies of scale.
If you look at all the advanced manufacturing
companies out there, in semiconductor foundries 
the scale economics is probably the purest.
The reason for that is the fixed 
cost is just so incredibly high.
Once you build a fab, like 80% 
of your cost structure is fixed.
What that means is every extra percentage of 
utilization is adding far more in revenue.
So it’s extremely important.
We’ll fight for every extra
percent of fab utilization, but also 
every extra percentage of fab output.
Let’s say I’m making some wafers 
today and the yield is 60%.
If I improve the process and the die yield 
becomes 65% or 70%, it’s not a linear advantage.
I’m getting way more output 
for the given footprint.
So there’s always room to 
improve the economies of scale.
That’s one of the reasons why you’re seeing 
so much consolidation in this business.
There are only three foundries left at 
the advanced node for that exact reason.
Along the way, many companies 
just stopped manufacturing.
I don’t know if you’ve seen…
There’s a chart where 40
years ago there were like 20 foundries.
Over time the number of foundries kept
diminishing until today there are only three.
What is then the concrete master
plan to catch up to TSMC?
Even though TSMC is ahead,
there’s so much overflow demand at their 
leading edge, where N3 is now majority AI
accelerators and maybe majority Nvidia alone.
There’s all this demand for CPUs, for other
things, which are also, by the way, needed for AI.
There’s this humongous influx
in demand for CPUs because of AI.
How do you see this opportunity right now,
where Intel’s a few years behind, 
while TSMC is just way oversaturated?
Catching up to TSMC has multiple facets.
The catch-up that we need to do is in scale.
Like you said, they’re just a much larger 
wafer manufacturer today than we are,
and even bigger than what Samsung is.
Catching up in scale will be a gradual process.
If you meant catching up in terms of just 
the technology, I think we’re already there.
We might be a few percent off on 
performance or power or area density.
But this sort of benchmarking, where one 
foundry is ahead or one foundry is behind,
makes sense when you’re a whole generation 
behind or two generations behind or ahead.
That’s when it makes a meaningful 
difference in your die output or your
die area or functionality like power.
But if you’re within a few percent,
you’re splitting hairs at that point in 
time, because it’s not very unidimensional.
It’s not just about density, it’s 
not just about raw performance.
Different applications require different vectors.
You can be off by a few percent,
but you need to be on time.
Your execution needs to be
such that your yields are very high to 
support very large and very rapid ramps.
That actually is more important than catching 
up on a few percent in performance or area.
I was looking at some SemiAnalysis 
numbers that seem to suggest that in
terms of transistor density, 14A — which will 
come out in 2028 — would be comparable to N2.
First of all, I would dispute your benchmarking.
But like I said, it’s important to be generally 
in the right ballpark and be very predictable.
Also, it’s not unidimensional.
We don’t build one type of
transistor, let’s say 14A or 18A.
You build a few different types of transistors.
There’s a low-leakage transistor, there’s a 
high-performance, high drive current transistor.
You’re building SRAMs. There’s a high-density 
SRAM, there’s a high-performance SRAM.
There are many other functionalities 
like analog circuits, RF circuits.
It’s the sum of all of those that 
makes a technology competitive.
I guess that makes the CapEx you 
do right now even more important.
TSMC is on track to do around 
$50 billion in CapEx this year.
You’re on track to do around $20 billion.
If the tools you’re buying right now will
determine who can service all this crazy AI 
demand that’s coming up, it makes the difference
in CapEx right now even more important, 
or for the US versus other countries.
Yes, very much. It’s a little bit of an 
unfair comparison between us and TSMC today,
just because today they’re a foundry for basically 
everybody and we’re largely a foundry for Intel.
We’ll become a foundry for others.
In due course, our CapEx will also grow.
Intel started as a DRAM business in the late ’60s.
You guys were producing DRAM, which you sold off.
A couple years ago, you sold 
off your flash business.
Thoughts on that? For context 
for the audience, these are
now multi-trillion-dollar industries because 
they’ve just exploded as a result of AI demand.
Clearly, in hindsight, probably not the smartest 
time to sell off a flash memory business.
There were good reasons why 
we did that when we did.
Clearly, nobody could see this enormous demand.
But today, the world will eat up as much compute
as can possibly be made, and it’ll eat up 
as much memory as can possibly be made.
I would argue that the supply crunch is larger 
for memory today than it is for compute.
Can I summarize what seems to me 
the takeaways from your projections?
You can tell me if that’s correct.
That marginal production can increase 
actually very fast — on the scale of 12
months — but the amount of increase up till 
the end of the decade is extremely limited.
Production could be higher next year than 
what people are forecasting, but production
in 2030 is not going to be significantly 
higher than what people are forecasting.
Unless the entire supply chain 
ramps up production very quickly.
We’re already in the middle of ’26, so there’s 
not that many years until the end of the decade.
Given the lead times…
Let’s say you place an
enormous order for tools today, and you’re 
looking at 18 months to get those tools.
You’re already into basically ’28.
From that point on, if you have to
build a new shell, build a new fab, you’re 
basically looking at the end of the decade.
I think it’ll be interesting to discuss the 
partnership you guys have with Elon on Terafab.
What’s going on there?
Well, Elon’s publicly talked
about his desire to build a fab.
We either build the Terafab or
we don’t have the chips.
And we need the chips,
so we’re going to build the Terafab.
Sort of the same thinking.
He’s seeing just an enormous increase in 
demand for silicon, for compute and for memory.
I think he came to the conclusion that his own 
businesses are going to require a lot of silicon.
He thought maybe he should 
just do it all on his own.
But unlike building a car, for example, 
this is such an intricate and complicated
process that we thought we could work together 
to help accelerate his bringing up his fab.
But because of the bottleneck at the 
tool level, it seems like the net
amount of silicon in the world won’t increase.
It’s just that maybe if somebody were to start
a new fab, they would get access to guaranteed 
supply in a way that they wouldn’t have if they’re
working with another person’s foundry.
It’s not just tools.
A fab does require a lot of power, a 
lot of water, a lot of consumables.
There are many other input variables as well.
But those can be scaled in the
way that tools cannot.
Some more than others, but yes.
In one sense, yes, we need a lot more equipment.
But it is true, we also just need a lot more raw
fab space, raw clean room capacity.
If you think of a foundry like Intel,
which has to build these expensive facilities and 
buy expensive tools, I wonder what you think the
economics of that fundamentally could look 
like, whether it could get more economical,
whether it could scale faster.
Like I described earlier,
this is an extremely capital-heavy business.
If you just plot how much it costs to build
a fab over the last 20, 30, 40 years, 
I think it’s doubling every 4-5 years.
Even inside the fab, if you look at the 
cost to do lithography, which is the most
expensive module — just patterning — that 
itself is doubling at more than that rate.
Today, a High-NA EUV scanner and all its 
associated equipment costs like $400 million.
So what does that mean?
If that rate were to continue,
if you just draw the line, does that mean 
four years from now the cost of patterning
is going to be like $800 million?
It doesn’t really make sense.
A very interesting thought exercise is 
to ask, are there things that could be
just very radically different that we could do 
today that we perhaps couldn’t do 20 years ago?
This is a very risk-averse industry for 
very good reason, for exactly that reason.
You put in tens of billions of dollars, and 
you really don’t want to take any risks.
You want to keep doing things the way 
they’ve been done for literally decades.
This is not just true of 
semiconductors, but everything really.
If you’ve just kept on doing that for decade 
upon decade, you just become indoctrinated.
There are engineers who probably work at Intel who 
are new, who come in and they’re taught, “This is
how the fab works, this is the process flow.”
They just take that as gospel.
That’s just how it is done.
But is that really the best
way to build a chip today?
It’s a really valid question.
I’m very curious what you’re hinting at, 
because it could be anything from figuring
out ways to use fewer litho passes and economize 
on this expensive component, to maybe we need to
rethink the transistor.
What is the level of—
All of the above.
Transistor architecture is evolving.
Just the silicon transistor has continued to 
evolve over 30, 40 years, and that will continue.
You’ve seen the evolution 
from the planar transistor
to FinFETs to now gate-all-around transistors.
There’ll be more. That roadmap exists, it’s pretty
well defined, with a lot of people working on it.
But there are things like…
Do you really need a clean room today?
Do you need the infrastructure to be as
sophisticated, as complex as it is today?
Surely one can run an experiment where you
have people eating hamburgers and smoking cigars 
in the research fab and see what happens to yield.
That would be a great experiment to run, but 
no incumbent — Intel or TSMC or Samsung — would
want to willingly do that.
That’s what I meant by this.
It’s a very risk-averse business, for good reason.
Every one day lost in productivity is…
I don’t know how many dollars.
But you save so much money in
not having to build clean rooms in the future.
But it’s really the risk, the fear of the unknown.
Nobody is going to willingly take that risk.
That’s why I think sometimes it’s good to just
start with a clean slate and rethink everything.
The same thing applies to chip design, by the way.
Many of the ways we do CPU design 
were probably locked in 30 years ago.
Does that mean it’s smart 
to still do it the same way?
Probably not. But this is a very complex problem.
In chip design, I guess the capacity to take
risk is maybe higher.
This one, less so.
But given where the economics is today, I 
think it’s a very valid question to ask.
Maybe there’s a smarter way to do this.
To close out, maybe we’ll start off at
the same question we began the interview with, 
which is, how does Intel respond to this crazy AI
demand that’s incoming and has already arrived?
It sounds like what Intel ought to do — if the
ultimate bottleneck is that you guys already 
have these foundries, you already have the
ability to expand these foundries, but what you 
will eventually lack is the tools from upstream
in the supply chain — is to just raise 
as much capital as possible right now.
The same way that Nvidia has locked down memory 
and other components for reportedly hundreds
of billions of dollars worth of commitments that 
it’s made, Intel should maybe do the same thing:
make all these commitments so that over the 
coming years, the US is in good position to
have the bulk of semiconductor manufacturing 
at the most advanced levels happening here.
You’re suggesting we should be 
buying a lot more tools, right?
Yeah, because it sounds like otherwise in 
a couple years you’d just be in a position
where you’re like, “Oh man, we could be 
making trillions of dollars right now,
but all the tools are in Taiwan.”
Yes, that’s a very valid question.
Had we known what we know today, we 
would have invested more like a year ago.
Knowing what we know now, I 
think we’ll be investing more.