Roel Apers

The through-line

Roel Apers argues that startups are built by exposure to real stakes, not by simulation. The recurring claim across everything he says is that the best accelerator is a paying customer or a high-profile pilot, because external pressure forces both the team and the product into readiness in a way that prototyping in the office never can [1]. The corollary runs through every other subject he touches: if the stakes are real, the engineering discipline has to change, the co-founder relationship has to hold, and the founder has to have an honest answer to the question of what they are willing to endure.

The second strand is a kind of unsentimental realism about what founders can and cannot control. Hardware startups fight battles entirely outside their influence, accelerator badges are worth only as much as the institution behind them, and the Belgian ecosystem offers an asset most founders leave on the table. What remains inside the founder's control, in his framing, is ambition, preparation and self-knowledge.

On the customer as the real accelerator

The sharpest version of his position is that a paying customer or a high-profile pilot does more for a startup than a programme does, because it removes the option of not being ready [1]. He has lived the trade-off. Doing development "on one of the most photographed stages in the world" is "very nice", he says, but "it brings pressure to the team as well" [1]. That pressure is the point, and it is also the constraint: "You can imagine that there's no room for errors on a stage like Supernova or on a Taylor Swift or on a Beyonce" [1].

On 'move fast, don't break'

The Silicon Valley formula does not survive contact with a five-ton moving ceiling. His revision is "Move fast, don't break. Iterate a lot" [1]. In practice that means front-loading the breakage deliberately and privately. Before Tomorrowland, KLSTR spent two weeks attacking its own system: "The two weeks before the festival, we really punched the out of it. Start unplugging network switches while it was running and doing emergency breaks" [1]. The payoff arrives when it matters, because "during the festival itself, it runs smooth like hell" [1]. The engineering ambition behind it is a step change in setup time rather than an incremental gain: "To model main stage for example takes around 4 days to configure and troubleshoot everything and with our solution it would take you 30 minutes" [1]. The end state he describes is infrastructural invisibility, the technology becoming a default nobody specifies any more: "It's a little bit like Bluetooth. You will not buy a phone without Bluetooth and hopefully in two years time nobody buys a fixer anymore if it doesn't have cluster natively on board" [1].

On hardware's battles outside your control

Hardware founders inherit problems they did not create and cannot solve. COVID chip shortages forced KLSTR to restart software development on a new platform, and a memory supply crisis turned a €2.6 component into a €27 one inside a €15 product [1]. He offers these as the ordinary weather of hardware rather than as exceptional misfortune, which is why the emphasis falls on the team's capacity to absorb shocks rather than on avoiding them.

On co-founders and knowing your limits

Complementary co-founders matter most in the first years, but he is clear that the arrangement is not static. Dynamics change as the company matures, and they change for reasons that have nothing to do with the business: divorces, illness, life events [1]. The practical implication he draws is that founders must know their own limits, so that the shift in the relationship is recognised rather than denied.

On ambition and the roller coaster

Ambition, for him, is not a growth target but the fuel required to survive the experience: "If you don't have the ambition to scale, you will not last this roller coaster of emotions" [1]. It is the same logic as the customer-as-accelerator argument, applied to the founder rather than the product. Sustained pressure requires a reason to keep absorbing it.

On accelerators, stamps and investor access

He distinguishes sharply between institutional credibility and generic credentials. The imec approval stamp opened investor doors that would otherwise have stayed shut, and KLSTR's first external investor started listening because of it [1]. But the lesson does not generalise to accelerators as a category: many stamps carry little weight, so founders should not join just any programme [1].

On failure, and Belgium's underrated network

Failure in Belgium remains somewhat taboo, with real consequences, since banks that funded a failed venture are unlikely to fund you again [1]. He refuses to make that the central question. "For me the thing is never is it okay in Belgium to fail it? Is it okay? Can I live with myself that I know that I didn't do everything to succeed" [1]. Against that cultural drag he sets the country's most underrated asset, which is network density. The ecosystem is small and connected enough that any advice or introduction is a single phone call away, and yet most founders do not use it: "It's so easy to just, ah, I need some help in this level, you just call somebody and five minutes later they are in a zoom call and I think that's really underestimated" [1].

Takeaways

• Treat a paying customer or a high-profile pilot as your accelerator; office prototyping lacks the stakes that force a team and product to readiness [1].
• In high-consequence deployments, adopt "Move fast, don't break. Iterate a lot" and schedule deliberate destruction beforehand, as with unplugging live network switches and triggering emergency brakes for two weeks pre-festival [1].
• Assume hardware shocks you cannot control: chip shortages that force a software restart on a new platform, or a €2.6 component becoming €27 inside a €15 product [1].
• Pick accelerators for the weight of the institution behind them; the imec stamp got a first external investor listening, while many other stamps carry little weight [1].
• Expect co-founder dynamics to change as the company matures through divorces, illness and life events, and know your own limits in advance [1].
• Without the ambition to scale, the emotional volatility of the journey will end you: "you will not last this roller coaster of emotions" [1].
• Exploit Belgian network density; help is one phone call and five minutes from a Zoom call, and most founders underuse it [1].
• Measure the product win in orders of magnitude: four days of main stage configuration and troubleshooting reduced to 30 minutes, with the goal of becoming a default like Bluetooth [1].