It’s hardware’s turn to eat the world.

In today’s Diary:

• Steel body, silicon brain

• The 2 AM recall

• A $93 trillion blind spot

• You heard it here first

• It’s 1830 again

Hey rational optimist,

As Marc Andreessen predicted in 2011, software has eaten the world.

Well… about 15% of it.

The other 85% of the world’s economy—the physical parts like energy, manufacturing, defense, construction, transport…

Have been all but frozen for 50 years.

Software has given us a lot. It gave us tremendous stock market wealth. It gave us little luxuries like working from home, endless entertainment and burritos delivered via DoorDash.

But something big has been missing.

Planes got slower. America built just two new nuclear reactors in 25 years. We built zero major commercial airports in that time.

While software advanced, the physical world crumbled.

That era is over.

A lot of people are calling what’s about to happen the next Industrial Revolution.

It’s true we’re at the cusp of a physical transformation of our world not seen since the 1950s. Hardware is all the rage.

But after meeting with 100+ innovators and founders who are actually building that hardware, we realized something.

We’re witnessing the rise of an entirely new type of company. Not hardware, not software, but a meld of both.

We’d like to submit a name for this new type of company: Silicon Industrials.

A Silicon Industrial makes rockets, cars and robots. But its nervous system still runs on software, which runs on silicon chips.

Tesla is the original Silicon Industrial

Every Tesla on the road generates data that flows back to its centralized “brain,” which is used to constantly improve every other Tesla.

In 2023 Tesla recalled 2 million cars overnight. But this wasn’t the kind of recall that forced you to haul your car to a mechanic. It was a “software recall” pushed in the middle of the night.

Ford ships a new physical model every five years. Tesla ships an over-the-air software update every few weeks.

That, above all else, is why Tesla is now worth 26 Fords.

Is Tesla a software company or a hardware company?

Neither. Both. It’s a Silicon Industrial.

Of course, Tesla is huge already.

But most Silicon Industrials are still tiny.

Bedrock Robotics is one of our favorites.

In Proto-Town we watched a normal-looking yellow Caterpillar excavator scoop a bucketful of dirt and drop it neatly into the back of a dump truck. Except this excavator was operating itself.

Bedrock doesn’t build million-dollar robo diggers. It builds hardware that bolts onto an existing excavator. And software that then turns the construction equipment autonomous.

Bedrock collects data from every machine it operates, which is then used to improve the system. Every trench dug and every load moved makes the next deployment smarter.

The Bedrock crew told us the AI figured out how to position the bucket before the next dump truck arrived. It learned to do that by watching human operators.

Imagine how much more efficiently we can build things with autonomous construction equipment that can work through the night? Could we build the average house for 5% cheaper? 10%? 20%?

Could it solve the housing crisis?

What Bedrock is doing for the construction site, Machina Labs is doing for the factory.

Machina builds software that turns ordinary Kuka robotic arms into what founder Ed Mehr calls “robocraftsmen.” These robocraftsmen are a game-changer in producing custom, high-value parts for cheap.

For example, Legacy foundries quote four months and $300,000 to make a one-off submarine part. Machina makes it in six hours out of $300 worth of sheet metal.

When we visited, we saw a pair of giant orange robot arms pinching a sheet of steel. One arm presses from the front. The other supports from the back. They dance together, adjusting 84 times a second, rolling and nudging the metal into a curved aerospace part.

The same pair of arms that machined a nuclear submarine part this morning can switch to making an F-35 fighter component this afternoon. It’s just a software update.

Base Power wants to turn your home into a power plant.

Base does it by installing a battery the size of a small refrigerator in your garage.

The magic isn’t the battery. It’s the software that decides whether each battery in Base's fleet should charge from the grid or discharge to it.

When electricity is cheap and abundant—like at night—Base fills the batteries. When electricity is expensive—like at noon in Texas—Base sells the stored electricity back to the grid at a premium.

Customers get cheaper power and backup security when storms knock out the grid.

Base just raised $1 billion. Most American homes could have one of its battery systems in a decade.

Wire harnesses are unglamorous.

Yet you rely on them every day.

A wire harness is the thick bundle of wires snaking through your car. A modern car has 40 of them. They make up a quarter of its total cost. Every rocket, drone, fighter jet, and patriot missile needs wire harnesses.

Senra Systems is reinventing how we make them.

When we visited Senra’s LA HQ in April, we saw dozens of wire harnesses sprawled across 30-foot tables. At Senra a new employee can build a working aerospace harness in a few hours. Industry standard: 12 months!

And it’s all because Senra built software that allows engineers to easily design a new wire harness in minutes. Its software helps build hardware.

We asked all four of these Silicon Industrial founders the same question:

“Could you have been built this ten years ago?” 

The unanimous answer: “No chance.”

Zach Dell at Base Power can only build a virtual power plant because lithium batteries dropped 90%+ in cost.

Bedrock Robotics can only retrofit excavators because compute and sensors got cheap enough to run an advanced AI model on a bucket arm.

All these companies are still private. A decade from now most will be public and worth at least 10X what they are today, in our opinion.

Valuations for pure software companies have collapsed 70%+ from their 2021 peak.

Meanwhile, the companies driving the Silicon Industrial Revolution are on a tear.

SpaceX is preparing an IPO at a valuation roughly 10X Boeing’s market cap. Anduril is already one of the 10 most valuable defense companies. Tesla added more market cap in 2025 alone than the entire global auto industry produced in the prior decade combined.

The global economy is roughly $110 trillion. The 85% non-software economy represents about $93 trillion of that. If Silicon Industrials capture even 30% of that share over the next 25 years, that’s $28 trillion of annual revenue flowing into companies that did not exist five years ago.

And that’s vastly understating the opportunity. Because Silicon Industrials won’t just capture market share. They’ll grow it. A lot.

A decade from now we think at least half of the 10 most valuable companies in the world will be Silicon Industrials.

Sounds far-fetched, until you realize there are already two in the top 10: Tesla and Nvidia. SpaceX will join them when it IPOs next week.

Think of it this way: we are living in Britain just before the Industrial Revolution. In 1830 roughly 5% of Britain's economy was in mechanized factories, steam power and rail. By 1880 it had surged to 60%.

The wealth created from 1830 to 1880 was roughly 100X the wealth created in the prior 50.

How to use this knowledge to your advantage.

If you’re an investor… know the old hardware vs. software divide is obsolete. The best companies meld both.

If you’re an entrepreneur… understand all the physical industries frozen in time, like energy manufacturing, construction, defense and mining, are now being unlocked by cheap intelligence. There’s at least one Silicon Industrial in each of these sectors.

If you’re a parent… make sure your kids and grandkids know the frontier is opening again. Make sure your children know not to take on $100k in debt to go to college without learning any real skills. Have them try to intern at an emerging Silicon Industrial.

And no matter who you are, make sure you spread these stories of rational optimism.

Economic historian Joel Mokyr found between 1650 and 1850, the use of words like “innovation” and “future” in books in Britain quadrupled. People started to believe things could get better.

What followed was the greatest increase in living standards in human history:

Long live the whoosh.

What do you think of the term Silicon Industrial Revolution? Think it can catch on? You heard it here first!

—Stephen McBride