No, You Can’t Just ‘Bring the Platforms Home’
After we wrote about Israel’s progress from Start-Up Nation to “platform nation,” a reader asked a perfectly reasonable question:
If platform companies are so important, why don’t we just bring them home? Why not subsidize them, relocate them, and require them to operate in the United States?
As a reminder, a platform is an enabling layer—something other systems build on, depend on, and integrate into. Platforms become embedded inside other nations’ industries, supply chains, and critical systems that would fail without them.
The reader’s question assumes these platform companies are portable assets, like factories or warehouses. But today’s most valuable platform firms are not stand-alone assets. They are the visible nodes of dense ecosystems.
Relocating a factory requires making new logistics. Relocating an ecosystem requires remaking history.
Consider for a moment not Israel, but Taiwan. It is not only Taiwan Semiconductor (NYSE: TSM) that makes Taiwan indispensable. Yes, TSMC produces roughly 60% of global semiconductor foundry output. But in this effort, TSMC is supported by hundreds of suppliers clustered near its facilities. Tool vendors, packaging specialists, materials firms, design houses, and university-trained engineers form a tightly interwoven network.
TSMC did not merely benefit from that ecosystem; over decades, it helped create it. As it grew, suppliers moved nearby, universities adapted curricula, and adjacent firms specialized around its needs. Engineers circulate between companies. Small process improvements compound. The company and the ecosystem evolved together.
That web of relationships cannot be quickly replicated, even with generous subsidies.
Innovative Israeli firms are embedded in a dense network shaped in part by elite military intelligence units that function as real-world systems’ laboratories, exposing young engineers to adversarial problems before they enter the private sector. Contributing are universities built to foster engineering excellence and a web of more than 400 research and development centers, mostly owned by U.S. firms, that employ many of those excellent engineers and keep them engineering. These close-knit learning networks, reinforced by a culture that tolerates rapid failure and redeployment, power Israeli startups and iterative innovation.
Silicon Valley is not reducible to a few large firms, but to a mesh of talent, capital, universities, and serial founders.
Industrial policy focuses on what politicians can count: factories, headcount, and capital expenditure. Those matter. But the binding constraint in modern industry is often less visible: tacit knowledge, dense supplier relationships, and feedback loops refined over years.
Subsidies can move capital. Tariffs can shift trade flows. But ecosystems grow through accumulated experience across multiple firms. That takes time.
The true cost of “reshoring,” especially in time, should be measured in decades, not budget cycles. A new fab can be built in a few years. Building a mature semiconductor ecosystem took Taiwan a generation.
There are likely multiple reasons TSMC’s Arizona fabs still can’t match the Taiwan’s yields or that its costs run 30-50% over the home field, but the lack of ecosystem support must be one. That lack motivated the company temporarily to move hundreds of staff from Taiwan to Arizona. And yet, the performance gaps persist.
For investors, and for us at the Gilder Technology Report, this changes the analytical question. Instead of asking only what a company owns, we must ask where it sits—within which ecosystem, with which suppliers, and, above all, inside which learning networks.