12 meditations on deep tech startups
🔥Twelve hot takes, how many are actually hot?
I recently ran workshops on product-market fit at three accelerator programs, focusing on deep tech.
On my train back, I struck a yoga pose in the aisle and reflected on my experience. The aircrew yelled at me to stop. On one hand, I agreed with them, we were about to land, on the other, they were ruining my meditations. I told them to be quiet.
Here are 15 meditations. That’s yoga-speak for hot takes.
1. Deep tech startups are not that different
Even though many claim deep tech startups are a different breed, most startups face the same challenges. Whether it's B2C, B2B, B2G, deep tech, hardware, or SaaS, they all need to find a replicable offer they can scale for relatively similar customers. The biggest differences come down to sales cycles, navigating large organizations, and deal sizes. But, the type of challenge they need to overcome—finding a customer—is the same.
2. MVP and beachhead thinking also apply to capital-intensive startups
Deep tech—especially hardware with low tech maturity (TRL 1-3)—comes with the challenge of needing investment before anything of value can be proven. In contrast, a few developers putting in the work can validate the technical feasibility of most SaaS ideas.
Hardware often requires steep component costs, pushing the first pilot to $50K–$250K, which is a much tougher sell than a $100/month SaaS subscription. This doesn’t mean that a capital-intensive startup shouldn’t try to find the most minimum MVP possible for a niche beachhead. There’s often options to start smaller than you think.
3. Software functions are vaguer than hardware functions
For many software-based startups, understanding customer needs is trickier than for B2B hardware. Software outcomes tend to be less constrained than hardware outcomes. Think about it—how would you define and measure the function of Slack or Notion? Increase in productivity? Less emails sent? Hardware functions are usually more intuitive and can often be measured in clear metrics like torque or material strength: This material now can withstand 5 times the force of the competitor. Very clear.
4. The catch-22 of customer and proof of value
If you sell a hardware filter for liquids, you can test its filtering power in a lab and show the results to early potential customers. Many SaaS outcomes—like better scheduling software for teams—can’t be easily tested in a lab. Your first customer often plays an experimental role, essentially becoming the test case for future customers. This can create a customer-proof catch-22: No customers because no proof, no proof because no customers. Though it’s not unique to software, I see it more often there.
5. "Spaghetti strateggi" for general-purpose tech
For hardware-based startups working on general-purpose technologies—like a new broadly applicable polymer or a foundational AI model—finding the right market can be incredibly tricky. When technologies are new and broadly applicable, many markets seem like a fit. Kevlar, now famous for bulletproof vests, was originally expected to create more fuel-efficient tyres.
To overcome this, deploy a classic "spaghetti strateggi": touch many markets and see what sticks. But, if you don’t have a demoable prototype yet, it can be tough for potential customers to envision the value you bring. Running 5 to 10 problem interviews per market segment should give you a sense of whether there’s real potential.
Silicon Valley founders, this one is for you. I’m hosting a workshop on product-market fit at HackerDojo. Join others startups to get a product-market fit checkup, a critical assessment of your go-to-market strategy.
6. Emotions do count in B2B/B2G
Is deep tech only functional? Even though emotional influence on purchasing is often attributed to B2C, don’t overlook the emotions of B2B or B2G customers. The purchasing process may be rationalized, but the impact of certain problems within an organization is deeply emotional. Understanding these on a deeper level—what drives them or keeps them up at night—can be crucial.
7. Red flags with founders end-game visions
Deep tech often demands significant investment to reach a commercializable level. Some founders get stuck on their end-game vision—insisting they need $50M to do it all at once—and refuse to break it down into smaller, more actionable steps. Red flag alert: these founders are often uncoachable. Very few founders can raise $50M at an early stage. Especially first-time ones without a proven track record can’t.
8. The "nothing to learn" mindset in breakthrough technologies
Similarly, some founders with breakthrough technologies believe their tech is so unique that there’s nothing to learn from other startups—assuming that none of the usual lessons apply to them. Another red flag.
9. Challenges of corporate deals and due diligence
Corporate deals can be great for deep tech startups, offering high volume and long-term stability. The downside? Corporations have extensive due diligence processes and often require tighter IP protections than you initially had. Finalizing a deal can take 12 to 16 months—does that timeline work for your startup right now?
I once spoke with a founder who landed a deal with Microsoft. It brought in great volume but ended up consuming 90% of the business’s bandwith. Deals like these can be a major distraction from your core vision, so if you go down this path, be aware of the trade-offs.
10. LOIs as a tool for gauging customer interest
LOIs (Letters of Intent) aren’t worth much on their own, since you can’t force a deal from them. However, if your MVP is 10 months out and you want to gauge how serious your potential customers really are, proposing a few might be a good move.
There are templates online, and founders often tell me ChatGPT is helpful for making them concrete. LOIs work well for raising capital, serving as a step between problem interviews and discounted pilots. LOIs also help clarify the features, functions, or metrics you need to hit for potential customers to stay interested.
11. Founder-led sales is the way to go
Founder-led sales is the best way for almost any deep tech startup, especially when you're going after the first 1-10 clients. Don’t outsource this to an intern—you are the salesman or saleswoman now. Dong sales is customer discovery on steroids.
To do this effectively, start with a simple spreadsheet and create a lead list. Make sure you dedicate at least 2 hours per day to outreach. Don’t get caught up in setting up a CRM like HubSpot right away; a spreadsheet will work perfectly fine for those first calls.
12. If you really need funding, consider non VC-routes
Bootstrapping is sexy, but if your deep tech startup really needs capital, there are less hostile options than VC money. Some deep-tech accelerator/incubator programs in the US require capital raised to get in.
Often, friends, family, and fools rounds, along with non-dilutive grants (which don’t require giving away equity) are used. These are great for any startup, as they help you stay in control of your business. Sometimes, launching customers can even fund your MVP—without the need for diluting equity.
Enjoyed this newsletter? Consider donating to support it!
> Dong sales is customer discovery on steroids.
Awesome quote. With and without spelling mistake.
🙌🏻