Innovation Ecosystems
Why Great Ideas Cluster: The Hidden Physics of Silicon Valley
Silicon Valley produces a wildly disproportionate share of the world's most valuable companies, and the reason has almost nothing to do with the people who live there.
The Idea
There is a tempting story about innovation ecosystems: brilliant founders move to the right place, meet the right investors, and build great companies. But this gets the causality backwards. What makes a place like Silicon Valley, Shenzhen, or Bangalore genuinely extraordinary is not the talent it attracts — it is the density of what economists call 'knowledge spillovers.' When engineers from competing firms drink coffee at the same café, when a failed startup's employees scatter across the ecosystem, when a VC partner's casual remark at a dinner party reshapes how a founder frames her problem — information moves in ways that no contract, no Slack channel, and no conference could replicate. The key insight from economic geography is that proximity lowers the cost of this unplanned, uncompensated knowledge transfer to near zero. Ideas, unlike goods, are non-rivalrous — one person using an idea does not deplete it for anyone else — but they are still imperfectly transmissible over distance. The richest ideas, the tacit ones about how to navigate a specific technology or market, travel best face-to-face. This is why innovation clusters tend to self-reinforce rather than spread evenly. Once a critical mass of related expertise concentrates in one place, the marginal value of being there rises for everyone, pulling in more talent and capital, which generates more spillovers, which raises the value again. The cluster is not just a location — it is a compounding machine.
In the World
In the late 1990s, the Taiwanese government made a deliberate bet. It recruited Morris Chang, a veteran of Texas Instruments with deep roots in Silicon Valley, to anchor a semiconductor cluster in the Hsinchu Science Park. The ambition was not merely to build one great company — TSMC — but to construct the conditions for knowledge spillovers to occur. Engineers trained at TSMC left to start suppliers, equipment firms, and design houses that settled within a few kilometres of each other. The cluster thickened. By the 2020s, Hsinchu had become so deeply embedded in global chip supply chains that a single earthquake near the park sent shockwaves through the technology sectors of a dozen countries. What Taiwan had built was not a factory — it was an ecosystem in the precise biological sense: a web of interdependencies so tightly interlocked that removing any one node threatened the health of the whole. The lesson policymakers drew was often the wrong one. Governments around the world tried to replicate Taiwan by funding anchor firms and building science parks. Most failed. What they missed was that the spillovers in Hsinchu were not manufactured — they emerged from decades of shared professional culture, trusted personal networks, and a common language around the specific, granular problems of chip fabrication. You can build the infrastructure. You cannot mandate the conversations.
Why It Matters
Understanding innovation ecosystems changes how you evaluate opportunity — not just for investors or policymakers, but for anyone building something. If you are choosing where to locate a venture, joining an ecosystem that is already compounding will likely matter more than optimising for lower costs or tax advantages. The knowledge you absorb by proximity to people working on adjacent problems is a resource that does not appear on any balance sheet but compounds over time just as reliably as capital. It also reframes what 'failure' looks like inside a healthy ecosystem. A startup that folds but disperses its people into the network is not simply a loss — it is a redistribution of tacit knowledge that makes the whole system marginally smarter. This is why the most resilient ecosystems — Silicon Valley after the dot-com crash, London's fintech scene after the 2008 crisis — often emerge stronger from setbacks than monocultures do. Finally, it suggests a useful question to ask of any community or industry you are part of: is information actually flowing freely here, or is it pooling in silos? The answer tells you something important about where the next interesting thing is likely to emerge.
A Question to Ponder
Is there a cluster — professional, intellectual, or geographic — where you are currently on the periphery but could move closer to the centre, and what has been stopping you?
Get a new one of these every morning.
Start learning with Thinkable