In July 2025, a startup called  Varda Space Industries raised  

$187 million to make drugs in orbit.

They are not the only startup trying to make  stuff in space. Starcloud - formerly called  

Lumen Orbit - raised $21 million to try and  assemble a 5-gigawatt data center in space.

And tech giants Elon and Bezos have been talking a  great deal about putting AI data centers in space.

A few people on the internet have riffed  off those trends and ideas to also ask,  

“Why not do semiconductors in space too?”

Bezos himself said all the way back in 2016:  "We can build gigantic chip factories in space".

Sure we can. But it's not going  to be practical. In this video,  

Jon spends way too much time on space fabs.

## AI Data Centers in Space

Let's start with the reasons why people think  we should be putting AI data centers in space.

Part of this recent push seems to be  frustration with how long the data  

center buildouts have been taking here on Earth  due to energy shortages or local opposition.

With regards to the latter  blocker, there are no local  

residents in space that I personally  know of so presumably we bypass that.

With regards to the former - energy - there  is something kind of compelling. Up in space,  

a data center taking a specific earth orbit  can face the Sun for extended periods of time.

And without clouds or atmosphere, solar panels  generate anywhere from 3-10 times more power.

There are other cited reasons.  Elon has said that space is cold,  

implying that we can vent the waste heat into  it. Though many have noted that in a vacuum,  

heat dissipation can only happen  via radiation. More on this later.

And data connections with space-based fabs  are assumed to work faster because signals  

can travel faster through the vacuum  of space than within optical fibers.

I guess he might be right. Or  not. I don't know. This video  

isn't about data centers. It's about space fabs.

## The Vacuum

Before we begin, I want to  point you to Nick Pfeiffer  

and Professor Glenn Chapman  at Simon Frasier University.

Their work in the late 1990s and  early 2000s really set the tone  

for the existing space-based semiconductor  research field. It was not the only source  

I used for this video but my  favorite. Go check them out.

Perhaps the single most significant benefits  of space-based semiconductor manufacturing  

are the cleanliness and vacuum. Semiconductor  manufacturing fundamentally demands cleanliness.

Stray particles like pollutants from the  outside environment or previous steps falling  

onto the wafer or interfering with equipment  cause defects that hurt yield and profits.

Well, if you want nothing, then you will find  plenty of that in space. The vacuum of space  

in Low Earth Orbit is measured at about ten  to the minus five to the minus eight pascals.

The most abundant thing up there is atomic  oxygen produced by the sun's UV rays hitting  

the oxygen in the atmosphere. More on  this later, it is not a trivial thing.

In semiconductor manufacturing, a "high  vacuum" environment is classified as  

being at ten to the negative  one to negative five pascals.

And an "Ultra-high vacuum"  environment is classified as  

being between ten to the negative  seven to negative nine pascals.

These are non-standard measurements so  there is some variation but the point  

is that Low Earth Orbit meets the particle  density requirements for an advanced fab.

On Earth, fabs maintain big HVAC  systems to change out the air,  

filter the air of particles, and maintain  temperature. About half of a cleanroom's  

operating costs are from HVAC systems - though  I want to note this is a very rough number.

And certain tools require a vacuum. I  remember visiting the massive High-NA EUV  

machine at ASML. The second thing you notice  about it - after how big it is - is how loud  

it gets. Those are the pumps maintaining  a vacuum environment for the EUV light.

Other tools that require some form of vacuum  include those for chemical vapor deposition,  

dry etch, ion implantation, and  sputter physical vapor deposition.

Up in orbit, we can get the strong  vacuum for basically free. Just open  

a side-vent to the outside. It is  simple, saves energy and avoids  

potentially damaging pump vibrations. The  resulting environment is thus very clean.

To keep any aforementioned atomic oxygen from  wafting in and interacting with wafers, Pfeiffer  

recommends placing those side-vents in the space  fab’s wake. So away from incoming oxygen atoms.

## Space Radiation

Something that came to mind when I heard  about this was the space radiation.

Space radiation describes protons and atomic  nuclei traveling through the universe. Much  

of it is solar wind from the Sun but there are  other particles from nearby stars and galaxies.

The earth's atmosphere and magnetic field deflects  many of these particles. And most simply pass  

through things untouched, but a percentage of  these energized particles do leave an impact.

In doing so, they knock around atoms  and potentially shift the chips'  

crystal lattice. This leaves behind  vacancy defects - a missing atom in  

the lattice - or interstitial defects  - an additional atom in the lattice.

This can cause the chip to degrade in performance  and eventually fail. It is why electronic products  

sent into space have to be hardened and why  solar cells in space slowly degrade over time.

Pfeiffer addresses this, saying that a  200-millimeter wafer receives less than  

10 rads of effective radiation damage  if it does not have an electric field  

applied to it. Such damage, he says, can  be easily repaired with a simple annealing.

This seems reasonable. But  I note that this assertion  

was made 25 years ago. As Moore's Law  progresses and transistors get smaller,  

even a small amount of lattice damage  can have significant consequences.

## Temperature Control

Another thing that concerns me are  the wild temperature swings in space.

As I hinted at earlier, fab HVAC systems not  only include filters and pumps for cleaning  

the air or producing a vacuum but also chillers  regulating the temperature. More energy is spent  

on the latter. If 50% of a cleanroom's operating  cost is HVAC, then 30 of that 50% are chillers.

Heat can come from the outside, like  the sun's rays or via dissipation from  

the inside, like heat-producing  equipment. Without an atmosphere,  

temperatures can swing from negative  220 to positive 220 degrees Celsius.

These tools release a lot of heat - some are  literal ovens - while being super temperature  

sensitive. The wafers themselves are  just as sensitive too. Variations on  

their surfaces create physical distortions and  can mess with the delicate process chemistry.

As mentioned earlier, the primary method  to remove heat in a vacuum is to radiate  

it away. The International Space Station  uses liquid ammonia to actively circulate  

heat to eight big radiators - each  about as wide and long as a city bus.

Engineers have been dealing with thermal  challenges in space for as long as we have  

been sending stuff to space. And maybe the orbit’s  stability helps too. I am no satellite designer.  

But the sheer amount of heat to dissipate  away looks like a significant obstacle.

## In Microgravity

The second defining thing about  space is the microgravity.

Low earth orbit is not zero-gravity. A  satellite orbiting 400 kilometers above  

the earth's surface is just six  percent further away than a guy  

on the ground. So gravity's pull on  the satellite is 88% that of the guy.

So the weightlessness comes from  the fact that the satellite is  

constantly falling towards the earth.  It creates something equivalent to a  

zero-gravity environment - Einstein's principle  of equivalence at work here - but not the same.

Ergo people like to say microgravity  rather than zero-gravity.

Gases and liquids act differently  in microgravity. Like, the relative  

density of substances does not matter  when they mix. Which means that we do  

not have any sedimentation or buoyancy.  Particles don’t fall. Bubbles don’t rise.

Without gravity to pull them down, liquids are  most affected by their surface tension. When  

by themselves they take a spherical shape. When  in a container, they no longer fit the shape of  

those containers. And on surfaces they spread  out into thin films and tenaciously cling on.

## Liquids and Microgravity Don't Mix

Some process steps are gravity neutral  or even do better in microgravity.

The various chemical and physical vapor deposition  methods fall into this category. They deposit thin  

layers of material onto surfaces.  CVD via gaseous chemical precursors.  

PVD via smashing particles off a  source so they drift to the target.

These work. Some might even work  better in microgravity. An issue  

with CVD for instance is understanding  the complex mixing of gases of varying  

densities at different temperatures:  Convection. That is no longer an issue.

Crystal-growing is another thing. This  has been studied for some sixty years.  

The general theoretical expectation  is that we can grow larger, purer,  

and better crystals than we can on earth.

Single-crystal silicon ingots can be grown using a  method called the floating-zone method. We melt a  

portion of polysilicon feedstock and let it gently  pour down to merge with a silicon seed crystal.

It is very pure because the melt never touches  a container's sidewalls. But on earth, we are  

limited to wafers of about 200 millimeters  cuz gravity. Microgravity changes this.

The big problem are when things  involve liquids. In microgravity,  

liquids lose their predicted behavior -  causing breaking changes to the tool. We  

must either re-engineer the tool  or take on a "dry" alternative.

Also. In a vacuum, exposed liquids  start boiling off as their molecules  

start pulling apart - releasing  potentially contaminating vapors.  

An unfortunate side effect of  having a vacuum everywhere.

## No Wet Cleans

Unfortunately, we use a lot of liquids in the  semiconductor space. Wet cleans, to start.

A large portion of our existing process node  involves just cleaning contaminants like  

particles, metallic and ionic residues,  and organic residues off the wafer.

Wet cleans use ultrapure water - or in the  case of a style of cleaning called RCA clean,  

liquid acids and alkalines - for this cleaning.  They are performed in baths or with sprays.

The best dry alternative to RCA clean is probably  plasma etch. The fab can produce a weakly ionized  

plasma - perhaps using atomic oxygen collected  from space - and expose that plasma to the wafer.

This is already done to strip photoresist  layers off the wafer after lithography.  

The major challenges are that (1) It  just takes longer; And (2) It is an  

ETCH process. We can try to control it but  there will be risk of damaging the chip itself.

## No Liquid Photoresist

Another wet-based process that  will fail is spin-coating.

Lithography requires a photoresist. When  exposed to light, the resist hardens and  

retains the chip's design information  during the subsequent etch step.

Today, the photoresist is a liquid, applied  onto the wafer using a spin-coat machine.  

The machine spins the wafer at a very high rate.  We then slowly pour the resist onto the surface,  

causing it to spread out to a nice, even layer.

In microgravity, the liquid photoresist  will not stay on the wafer. The machines  

also use vacuum suction cups to hold the wafer,  

which too will fail. So we must turn  to non-liquid "dry" photoresists.

In 2000, Chapman and Pfeiffer proposed an  all-dry resist system that uses a bi-layer  

of inorganic carbon and aluminum  and can be etched using reactive  

ion etching. Overall it works, except  it is not very sensitive to UV light,  

which is not good for throughput. I have not  heard anything about it since the mid-1990s.

Fortunately, the industry is already moving  towards other dry resist solutions to adapt  

to the stricter conditions of EUV lithography.  Metal oxide resists from JSR and Lam Research  

are applied using vapor deposition. They  are also slower, but there is no choice.

## No Immersion and EUV

Today's leading edge fabs largely use EUV and  immersion 193-nanometer lithography machines.

The majority of layers are done using the latter  - which is a problem because immersion depends  

on pumping ultrapure water between the lens and  wafer to raise the tool's Numerical Aperture.

The use of liquid water is untenable in vacuum,  

it boils off immediately so we need a  pressurized mini-chamber. Plus in microgravity,  

the water acts in such unexpected ways to  necessitate an extensive redesign. Like to  

prevent defect-causing bubbles, for instance.  To me this is a no-go from the very start.

How about an EUV machine? Maybe this is better.  EUV light already needs a vacuum - though again  

I worry that oxygen atoms in LEO space  can enter and degrade the mirror surfaces.

And in microgravity, the big EUV  mirrors won't weigh anything,  

which in theory can reduce the  tool's overall size and complexity.

Just for fun, can we even send an EUV machine  into space? Per ASML, an EUV NXE:3300 machine  

(so not High-NA EUV) is 150,000 kilograms and has  a volume of 765 cubic meters. This may or may not  

include the drive laser, cooling plant, hydrogen  gas support system and other subfloor equipment.

Falcon Heavy cannot send that up, so  you would have to wait for Starship.

The problem however is that the EUV machine  famously uses an immense amount of energy - one  

machine eats about 1-2 megawatts. Producing  a megawatt in space wouldn't be impossible,  

though it would certainly  need a massive solar panel.

The real question to ask is how are you going  to dissipate 1 megawatt of heat generated by  

the machine? That is 14 times more than what  the International Space Station’s radiators  

reject. Back of the envelope, such a radiator  array might get as big as a football field.

There are other things that I am thinking  about. How do you catch and retain the tin  

droplets in space without gravity? How do  you prevent all the space fab’s jitters  

and random movements and counter-reactions  from ruining the precise lithography overlay?

I can do this all day but let's keep  going. Let us just assume that any  

space fab maker goes and rebuilds ASML's  193i or EUV machines ... but smaller,  

more energy efficient, more physically  stable, and needing less maintenance. Easy.

## No Wet Etches

After patterning the wafer,  we need to do the etch step.

There are two types of etch: Wet and Dry. The wet  

etch dunks and shakes the patterned  wafer inside a tub of acid etchant.

Since it involves a tub of liquid,  we cannot use that. Bubbles formed  

by the chemical reactions also can’t  be easily shaken off due to the lack  

of buoyancy - preventing fresh etchant from  reaching the wafer surface and creating defects.

Fortunately the industry has largely adopted dry  etch tools - Reactive Ion Etch, plasma ashing,  

etc - to take advantage of certain directional  etch benefits compared to their wet alternatives.

But wet etch is still used for a few process  steps where those benefits are not necessary.  

These would need to be replaced with dry etch. And  at this point, I am starting to wonder - between  

the cleaning and etching - how many plasma etch  tools we are going to need on this space fab.

## No CMP

Another significant missing process step  is Chemical-Mechanical Polishing or CMP.

We use CMP for "planarization", i.e.  making things flat. ICs have multiple  

layers stacked on top of each other, so a  non-flat surface can lead to serious defects.

CMP combines the mechanical action of  a grinding pad with a reactive chemical  

slurry to grind down layers of material  to produce a very flat and level surface.  

Since that slurry is liquid, it  cannot be used in microgravity.

The thing is that CMP works so well that  "dry" alternatives have not been explored.  

There are variations - adding ultrasonics  or changing pads - but the slurry has always  

been there. So we need something entirely  new. What that is, I have no suggestions.

## New Wafer Handling Systems Another issue to consider is wafer handling:

How do we transport wafers between tools and hold  

them inside the tool? And  not damage the wafers too?

I mentioned one method earlier in  passing: Vacuum suction. This one  

is untenable in a vacuum environment because we  need atmospheric pressure outside of the suction.

The brain-dead simplest way to do it on  earth is with a spatula-like effector that  

scoops up the wafer like a pancake. This  obviously does not work in microgravity.  

The robot arm can pinch the wafer  at the edges, but that risks damage.

So probably your best bet is something  along the lines of an electrostatic  

chuck. This creates an electric field to  securely attach the wafer to the clamp.  

They are already used in vacuum  environments like the EUV machine.

We cannot use this all the time however,  because the electric field's presence can  

attract space radiation - possibly leading  to more damage. In any case, every tool in  

our space fab has to be re-engineered  for this new wafer handling system.

## Logistics and Lift Costs

Unlike data centers, fabs need to be  constantly supplied with a variety of  

raw ingredients like chemicals, silicon  wafers, metals, and precursor gases.

They also need spare parts and access to  maintenance. When a tool like a lithography  

machine breaks down - and it often does  - we need to get it back online as fast  

as possible. Every minute of downtime costs a  commercial fab tens of thousands of dollars.

An engineer has to put on a clean  suit and go into the cleanroom to  

see what's going on. And if something  needs replacing, a new part needs to  

be available within hours. They cannot  wait to take the next SpaceX rocket up.

And as for lift costs, Pfeiffer's 2000  thesis did an economic analysis for the  

operating costs of an all-dry space  fab producing 5,000 wafers a month.

It assumes that we redesign all the semiconductor  tools to be "vacuum-native", removing pumps,  

structural supports, and subsystems - which  was estimated to reduce the machines' mass and  

volume by anywhere between 40-60%. The financial  analysis does not take into account R&D costs.

The base case assumed lift costs  of about $5,000 per kilogram and  

concluded that it would cost twice as  much to make a wafer in a space fab  

as compared to a ground-based fab running  wet processes. Again, only operating costs.

In the years since, the launch of  reusable spacecraft like Falcon,  

Starship and Electron has lowered lift costs  from $10,000 per kilogram to potentially as low  

as $1,000 to $2,000 per kilogram. Maybe  even $500 in the not-so-distant future.

But Pfeiffer also ran a scenario that reduced  launch costs to about $1,000 per kilogram.  

He also slimmed down the thickness of the wafers  to make them cheaper to transport to the fab. The  

model still said that space-based operating  costs would be 12% higher than that on Earth.

That was 25 years ago. Back then, the leading  edge was 130 nanometers. 193-nanometer DUV  

machines were just entering the cleanroom.  And fabs were still largely fabricating 200  

millimeter wafers. A leading edge  fab cost only a billion dollars.

Things have changed. Tolerances are  stricter. The lithography issues  

alone are untenable. And today, the  most compelling economic benefit of  

going to space - the free vacuum - is  largely a solved problem in the fab.

I asked people at several tool vendors  about how much engineering mindshare and  

mass/volume they spend on vacuum conditions  like UHV. The answer was very little. A space  

fab would be unfathomably cool, but  this juice is not worth the squeeze.

## Conclusion

Okay so where do I fall on the Space Fab? I  think it is technically possible. But so is  

me learning how to dunk. I don't think  it's anywhere economically practical.

A space node cannot use existing process nodes,  

nor can it use existing semiconductor  manufacturing equipment. So we have to  

develop both from scratch at the same time.  A complete custom job from the ground floor.

Which might excite the Silicon  Valley "disruption" boys,  

but feels like a nation-state type project.  More Huawei than two dudes in a garage.

Moreover, we did not even mention test and  packaging - much to the consternation of my  

friends in the Test industry. Neither are  anywhere near as automated as front-end.  

So we need to send the chips back to Earth and  pray the wafers don't break on the way down.

At the end of all that, you get a silicon chip  that is worse and more expensive than what Intel,  

Samsung and TSMC have been making at massive scale  

for over half a century. Why not  just send the chips up into space?

What makes more sense than space fabs  are special material-making facilities  

that leverage microgravity and relatively free  energy to produce stuff like silicon wafers.

More well-studied, smaller in scope,  and the conditions cannot be as easily  

achieved on the ground. Let's not  rush to put the most sophisticated  

and demanding mass manufacturing ever  attempted into a literally alien world.