Martin Van den brink on R&D investment
From Oral History of Martin van den Brink · · Computer History Museum
“I have a simple rule. It needs to be between 20 and 20 and 10%. And 20 and 10% is depending on the business cycle more than anything else. So in fact if you have ex if you spent one year 15% you could spend up the next year with the same spending 20% if your business go down or the business goes up and you're ending up as spending 10%. So you I have a slow adoption rate to the to the business between 10 and 20%. And try to stay within those boundaries but we really like to get stability and on the long-term road map.”
On , Martin Van den brink, Cofounder at ASML, spoke about R&D investment during Oral History of Martin van den Brink on Computer History Museum.
About Martin Van den brink
In a September 2025 oral history interview with the Computer History Museum, Martin van den Brink, president and CEO of ASML, discussed his early life, career, and the company's technological trajectory. He described being diagnosed with dyslexia and following a specialized technical education path in the Netherlands before studying physics. Van den Brink recounted ASML's early reliance on suppliers due to its small size, stating, "We relied on suppliers as much as we can not because we felt this was the smartest way of doing it, but because when you are small with 50 people, your capability is simply limited and there's only one way to go fast: looking for friends." Van den Brink also addressed ASML's research and development spending, noting, "We have a simple rule: it needs to be between 10 and 20% of revenue on R&D. We're spending four billion a year in R&D." He expressed confidence in the future of lithography, saying, "This may sound a bit arrogant but I think the technology of lithography is unlikely to determine the end point. It's very likely that the whole chip roadmap will continue to scale at least for the foreseeable future at least 10 years from now, 15 years from now." Additionally, he offered a critical view of current AI development, stating, "The current AI is almost lazy and prefers to extend the model size and add learning cycles rather than simplify the model."