Reflections on building novel hardtech: engineering, execution, and mindset

Hundreds of components and thousands of parts now make up our direct air capture system in Innisfail, Alberta. After an enormous effort from our team, our contractors and our project partners over the past 12 months of installation and commissioning, it’s edifying to be able to say: “It’s working.” It is also, more fundamentally, a relief. 

We still have some way to go with “operational ramp up.” But as the year draws to a close we’ve been reflecting on the difficulties and challenges we encountered on the way to operational success. Much of this is technical: it’s about piping, valves, electricals, blowers, concrete pad, and bubbling sorbent … lots and lots of bubbling sorbent. But most of what I will take away are the ways in which we’ve changed as a team: in experience, knowledge, and maybe most importantly, self-understanding.

And those changes are by nature deeply unpredictable. For example, it’s now late November, and once again it’s snowing on the Canadian great plains and sub-zero even in the daytime. No one tells you in advance how sub-zero temperatures will affect your lab-designed system. Like with so much else, you learn by doing — such as realising how delayed we would have been if the concrete for our pad hadn’t been poured before the deep winter set in.

I wanted to share some of the lessons we’ve learned over the past year. It’s important to cut through the fog of marketing and confidentiality that often obscures what it takes in time, sweat and treasure to design, build and prove novel hardware technologies. 

These lessons focus on project execution. Because direct air capture, and other capital-intensive climate technologies, are moving (and must move) from pilots to commercial projects that deliver on their lofty promises. The lessons are also about mindset, teamwork and relationships, because these are the underpinnings of even the most technical capital project. 

At the end of this year, my fundamental optimism about direct air capture is unchanged. But it is now more finely calibrated by realism. I am clearer-eyed about how long and with what resources it takes to deliver, where to take risk and where to play it safe, and what is and isn’t an exciting commercial opportunity. I hope this is helpful to the many other teams building hardtech and finding out just how hard it is. 

1. Good engineering starts with clarity.

 We learned early on that rushing design work almost always backfires. Spending an extra week on detailed designs, well-defined scope, and rigorous reviews can save months — and millions — down the line. One of my favourite sayings from the team is: “One dollar in design saves ten in the shop and a hundred on site.” That’s not from theory - we’ve lived it. Both when it’s worked out well but saw how it could have gone wrong, and when we did miss things in design and had to remedy later on-site. Thorough design reviews, third-party oversight, and clear documentation have become non-negotiable for the next set of projects we’re working on.

2. Be paranoid about your supply chain.

Suppliers, even the best, make mistakes — and lots of them. The lesson? Build management systems that expect (indeed, actively hunt for) problems, especially for things not in your control. You do this by enforcing accountability with clear penalties and warranties, providing incentives for over-performance, maintaining relentless oversight, and most importantly never relying too heavily on a single vendor. Competitive tension, rigorous acceptance tests at the factory and on site, and constant communication are what keep things on track. Under all of that you need strong relationships that can withstand a lot of stress. 

3. Once and done: take the time up front so you can install in one quick go.

 We discovered that incremental installation breeds complexity. It’s far more efficient to install complete, tested packages than to piece things together and test on site. When everything arrives ready and responsibilities are clear, installation is faster and more predictable. This also makes it easier to control costs.

4. Commissioning tests more than just equipment.

 Commissioning is downright painful at times. But it’s where a technology truly comes alive. We’ve learned to design with commissioning in mind. You need to have easy access to the equipment and the right tools ready to go. You also need to be realistic about timelines. This is hard because you actually just don’t know how long it’ll take, especially the first time. We also learned that at some point you have to move to night shifts and weekends when you’re into continuous operation, and having the team and contractors ready for that big shift in intensity is critical.

5. Discipline beats momentum.

 There’s a temptation in early-stage companies to charge ahead, to just get it done — especially because so much of the ethos of wider startup life is geared towards fast growth. But for capital-intensive projects with novel technologies, that bias can be dangerous. The opposite is also true: it’s dangerous for a startup to be overly cautious and take too long to move from lab to chunky engineering projects. It’s a hard balance. At the project level we’ve learned how important stage-gating execution is: every “go” decision is deliberate and tied to clear evidence and review. The stage-gate process need not be bureaucratic  — but even when it feels constraining, it’s necessary for safeguarding capital, people, and reputation.

6. Respect the weather: on “external factors.” 

You can’t pour concrete at -30°C. Commissioning is a whole lot harder when it’s -10°C in the daytime and your commissioning team has to take warm-up breaks every twenty minutes. This sounds obvious but like all “obvious” things they still need thinking about and planning for. Every project must respect its site’s unique conditions — from weather and geography to environmental constraints. We could have been delayed months if the concrete pad hadn’t gone down in time. We’ve learned to plan around the coldest months for very human reasons too: it’s hard for people to perform at their best when it’s freezing and dark all the time. 

7. Balance innovation with reliability.

 Perhaps our most profound learning: innovation and execution are different disciplines. We understood this intellectually before, but now we feel it in our bones. We separate technology development from commercial delivery, advancing R&D rapidly in-house while keeping commercial projects anchored to proven, reliable designs. That’s how we scale without losing our footing. It’s how we find the right balance between going to market for revenue generation and progressively pushing down the per tonne costs of our technology. 

8. There are limits to what you can know, especially when you’re doing it the first time.

 Executing a project that has lots of new elements requires a lot of problem-solving without a lot of visible progress. To do that means accepting what you don’t know: sometimes you simply don’t know, for example, how long it will take to bring a critical piece of equipment online. It demands resilience as you navigate delays and unexpected costs. It means that even when you experience mini-triumphs — like making that double-dump valve work, finally — you have an unpleasant sense most of the time that you’re moving too slowly. Yet those challenging stretches are where you learn the most: about your own instincts and limits, and about how your team responds to adversity. It’s not inevitable that it makes you stronger, but with the right team culture and deliberate effort, it absolutely can.

Through all of this, one truth stands out: complex engineering isn’t just about building machines. It’s about building trust, rigour, and learning. Getting our DAC system up and running at Deep Sky Alpha has made us better engineers, better project executors, better collaborators, and a stronger company as we increase our focus on the next set of commercial projects.

We’ve all, I think, arrived at the end of 2025 as what I’ve been calling ‘reality-adjusted optimists’. We have much deeper knowledge — and yeah, a bit of scar tissue — of what it takes to deliver, even if we still have much to learn. We remain optimistic about developing our technology and executing the exciting pipeline of projects we have ahead, not least because we now know which gloves to pack for the Canadian winter.