Capex: an investor perspective

There’s money to be made, but we investors need to evolve our approach

I was speaking to a fellow deep tech investor recently who told me he was pulling back from “capex-heavy” ventures.

I was disappointed.

Modern venture capital loves capex-light. SaaS. Marketplaces. Mobile apps. Cheap to start, fast to scale. Most of the big wins have come from there.

So he’ll do well.

But there are whole new industries coming that won’t work like that. Quantum computing. Industrial biomanufacturing. Nuclear energy. Robotics. These are physical. They need hardware, infrastructure, steel in the ground. They’re not optional. They’re the platforms the rest of the next economy will stand on.

Look at Nvidia. No AI boom without their hardware. And no Nvidia without TSMC, one of the most capital-intensive companies in the world. Both are dominant. Both are essential. Both are capex-heavy to their core.

Capex Is Hard. That’s Why It’s Worth It.

I am inspired by the people that want to build physical things. But I know it needs a different playbook. We can’t use the same tactics that work for software.

My investor friend asked what we do differently.

Here’s what I told him, and a few more things we’ve learned along the way.

1. Only Build What You Absolutely Have To

Don’t default to owning the whole stack. Run every other option to ground first. Can someone else manufacture it? Can you license your plaform to others?

Every piece you add is another way to fail. Every asset is a cost. Don’t let ego push you into being a full-stack hero.

If you have to build it, fine. Let’s do it and find a way! See below.

2. Know the Value Chain Cold

Map the ecosystem. Then go deeper. Who makes real money? Who scrapes by? Who are the incumbents you’ll threaten? Who will benefit from your success?

Some will block you. Some will pull you in. You’ll need both. Know who they are.

Look for the “drop-in” — solutions that fit with how the world already works. Help incumbents make more money. The more you ask the world to change to suit you, the lower your odds.

3. Think at Industry Scale, Not Just Company Scale

You’re not just building a company. You’re nudging a whole industry forward.

• What needs to change in regulation?
• Who needs to work together?
• Are there trained people to run this?
• Where should the first examples be built?
• What’s the story that gets everyone moving?

This is market annealing. If you don’t do it, nobody will. And your company won’t go far without it.

4. Partner With the Best

Being a manufacturer is different from being a product company. Most startups that try to be both, fail at both.

Use the supply chains that already exist. Partner with people who’ve done this before. If you have to give up margin or exclusivity to get started, do it.

Speed is more important than purity. Get to market. Learn. Then improve.

5. Stack Your Financing

Capex-heavy companies get into trouble when they raise equity like they’re a software company. It doesn’t work. We need to build a stack of capital, sharing the risk and minimising dilution:

• Target 50% non-dilutive. Get serious about grants. Globally. Have someone full-time on this.
• Learn how project finance works. Know what FOAK and NOAK mean. Understand what debt investors want to see before they will engage? What kind of offtakes from customers do they want to see? How might you bring those forward? The sooner project financing can scale the business, the sooner founders and investors stop diluting.
• Get commercial pilots. Get customers to pre-pay. Try to do deals that build into larger outcomes as milestones are reached.

You can’t brute force a business like this with equity financing. It will be too dilutive for founders and investors and they will lose faith.

6. Milestones Cannot Slip

Capex sucks up cash. So you need to be crystal clear about where it’s going and deliver it on schedule so we don’t deplete reserves before you have the evidence to raise the next round.

Set tight milestones and build the team discipline to hit them.

7. Hire People Who’ve Done It

You need people who’ve lived it. Who’ve built plants. Run pilots. Closed big industrial sales. Delivered government-backed debt deals.

This isn’t a learn-on-the-job environment. Don’t hire for potential. Hire for scars in these critical roles.

8. Get Creative With Business Models

Don’t fall into the trap of commoditising early. It happens often as new companies race to the bottom to get the business that will be hard to drive back up.

• Focus on revenue now. Even for pilots.
• Show you can sell, even if the margin is thin for now.
• Build toward long-term pricing power. Consider models from other kinds of business. You don’t need to default to traditional models.
• Don’t expect a green premium. Customers and investors have lost faith in that. You must have a path to better cost structures.

Find models that give you control and flexibility. And let you scale without betting the company every 12 months.

—

Silicon Valley was founded to do hard things like this. VC rhetoric is that it is ‘Time to Build’ so let’s do it. There is no need to be afraid of capex-heavy companies. We just need to invest differently.