October 30, 2025
·
5
min read

What founders get wrong about speed and scale

Insights
What founders get wrong about speed and scale
Every founder faces the same question: how fast should you go? In conversation with our Venture Partner, Duleepa “Dups” Wijayawardhana, we explored why so many startups struggle to balance speed, quality and scale, and how to get it right (at least most of the time). As Dups puts it, for every rule there is an exception, and for every exception there is something to learn.

Myth 1: Fast always wins

The old Facebook mantra of “move fast and break things” makes sense when applied carefully. But as Dups says, “speed only matters when you’re learning.”

In the early days, getting something into users’ hands was the only way to discover what works. Once customers start paying, breaking things only breaks trust. “You can’t scale a business on apologies,” he says. Real velocity comes from stability that allows you to move confidently, not recklessly.

Myth 2: Perfect code builds great companies

Early-stage engineers often want to craft elegant systems. Dups calls that a trap. “If you’re spending months perfecting code without customers, you’re building for yourself.”

Perfection too early drains energy and time from learning. The goal is to understand customers fast, not to showcase engineering skill.

With new AI tools, there are now infinite ways to reach an outcome that might once have taken years. According to Dups, if you spend months getting to a prototype for testing with customers, depending on the industry and the problem, someone else is likely already there.

Myth 3: Scaling is a technical problem

Most scaling failures are human, not technical. “Clarity, empathy and understanding what people actually do, that’s what scales teams,” Dups says.

He stresses that leaders must stay close to their teams’ work. When management loses touch, processes multiply and dysfunction follows. “If you don’t understand what your people do, you can’t lead them.” A good leader, he adds, stays close enough to maintain understanding. A great leader hires people who can scale that understanding both across the team and towards management.

Myth 4: AI will fix everything

AI promises efficiency, but Dups sees another side. “While already incredible, AI tools will improve even further in the coming years.”

He believes AI is reshaping how engineers think, shifting focus from the tools themselves to product outcomes. Yet skipping the fundamentals can be costly. “If you don’t understand the tools deeply, they will hurt you.” Founders must bridge those who build with those who think in outcomes.

While the future is unwritten, Dups believes engineers who combine strong technical skill with a deep understanding of generative AI, machine learning and creative thinking will become some of the most sought-after professionals in the market.

Myth 5: Scaling is linear

Growth never moves in a straight line. “It’s easy to go from one to ten million,” Dups says. “It can be painful to go from fifty to a hundred million.”

Processes that once helped start to hide problems. Leaders need observability, understanding and empathy in both their systems and their teams to see what is really happening and fix it before it slows them down. While Dups believes scaling is always a team effort, having the right people with the right experience, skills, and mindset at the right time can make all the difference. In the end, it is always about people.

The outcome

Speed without direction is noise. Quality without agility is paralysis. The best founders know when to trade one for the other. As Dups puts it, “You can’t know the exact formula, you just need to know your people and your purpose.”

Quantum computing in Europe: lessons from IQM and OpenOcean

What we’ve learned from the front lines of AI-enabled services

The market map for AI-enabled services: where to watch in 2025

The state of agentic AI in 2025: what’s working, what isn’t, and what’s next

News from our network

All news
September 29, 2025
·
9
min read

Quantum computing in Europe: lessons from IQM and OpenOcean

When we met Jan Goetz, Professor Mikko Möttönen and the rest of the IQM founding team in late 2018, we knew we were seeing something rare. The science was hard to grasp, even for my colleague Patrik Backman, who had studied physics and some quantum theory, but the ambition was unmistakable.
August 22, 2025
·
10
min read

The state of agentic AI in 2025: what’s working, what isn’t, and what’s next

At the start of the year, “2025: The Year of Agents” sounded like headline bait. Eight months in, it feels like an understatement. Sam Altman said agents would “join the workforce” in 2025. Satya Nadella expects them to replace segments of knowledge work. Marc Benioff wants Salesforce to be the “#1 provider of digital labour.” That’s not future tense anymore - today, agents are moving tickets, shipping code, and digesting documents.
June 23, 2025
·
9
min read

What we’ve learned from the front lines of AI-enabled services

Last November, we laid out a vision for AI-native services: businesses rebuilt from the ground up to combine the margins of software with the resilience and depth of services markets. Since then, we have taken dozens of first meetings, explored more than 15 verticals, backed our first deals, and sharpened our thesis along the way.