The Chip Control Fantasy

The current push for semiconductor export controls ignores a fundamental reality: you can't undo forty years of willful technology transfer with regulations. For decades, Western companies systematically dismantled their vertical integration capabilities in pursuit of quarterly profits and shareholder returns. This wasn't an accident or theft—it was a conscious choice to trade long-term sovereignty for short-term gains. Today, as geopolitical tensions rise and supply chain vulnerabilities emerge, we face the consequences of these decisions.

The depth of this self-inflicted wound becomes clear when we examine the history. In the 1970s, U.S. semiconductor firms operated as vertically integrated powerhouses, controlling everything in-house, from R&D and chip design to fabrication and packaging. Companies like Texas Instruments, Fairchild, Motorola, and IBM handled their entire production chain. This model—the Integrated Device Manufacturer (IDM)—gave firms end-to-end control over production, helping protect intellectual property and ensuring quality.

But by the 1990s, incentives for executives, capital providers, and short-term shareholders drove a corporate demand for higher margins, which caused a massive shift. Manufacturing was seen as a liability that hurt quarterly numbers. Companies rushed to go “fab-light” or “fabless,” outsourcing production to save on costs. The rise of Taiwan Semiconductor Manufacturing Company (TSMC) in 1987 accelerated this trend, proving that design and production could be separated. This sparked a gold rush of companies eager to shed their manufacturing operations.

This change represented an intentional transfer of institutional knowledge and capabilities. When IBM formed partnerships in China sharing PowerPC chip designs, or AMD licensed its x86 CPU technology to Chinese firms for quick cash, they weren't just outsourcing production—they were training their future competitors. U.S. tax and trade policies encouraged this behavior, offering deferrals on foreign profits and eliminating tariffs on semiconductor imports. 

When companies wanted to sell in China, they were often pressured to form joint ventures with Chinese firms and “transfer technology.” China leveraged gaps in WTO rules and its huge market to coerce access to intellectual property from foreign companies, and the U.S. government and international trade system went along. Western policymakers prioritized market access and consumer benefits instead of maintaining power over: 

• Means of production (including access to raw material, direct relationships with vendors, etc.)

• Manufacturing capabilities (understanding how to turn an input into an output, maintaining control over the system) 

• Intellectual property

The numbers tell a devastating story: The U.S. share of global semiconductor manufacturing capacity plunged from 37% in 1990 to just 12% by 2020. But more critically, despite the U.S. inventing this technology, we currently can not make a single one of the most advanced chips. In our pursuit of short-term profits, we exported not just our factories but the crucial “learning by doing” knowledge that comes from actually making things. (If you’re interested in understanding more about how we got where we are, I strongly recommend reading Chip War and The Big Score.)

The recent administration’s export controls rushed through in Biden’s final weeks create an intricate web of licenses, exceptions, and validated end-user categories. But this complexity only provides more opportunities for creative trade structuring. (For example, the export controls announce the creation of categories like “Universal” and “National” validated end users. One implication is it essentially creates more paths for commodity flow, not fewer, which makes the hardware more difficult to control).

The attempt to control compute at the Total Processing Performance (TPP) level also reveals a fundamental misunderstanding of how physical commodities move in global markets. You can't effectively control computational capacity when you can't even control the physical chips themselves. The market will simply re-route around these restrictions, using the very complexity of the regulatory framework as cover for legitimate-appearing transactions.

For the G5000 companies that want to dictate their future instead of being subjected to the whims of politicians who don’t understand technology, the choice couldn’t be more clear. The path forward isn't through hoping for protection but through rebuilding sovereign capabilities.

This requires a fundamental shift in strategy, culture, and operations. It starts with companies industrializing their InfraOps, taking ownership of their infrastructure operations, and building the internal expertise needed to operate at scale efficiently. It means investing in people, processes, and technologies that allow companies to end outsourcing of critical systems and return operational authority to where it belongs.

To achieve this, organizations must embrace several critical tenets of InfraOps that directly address their current predicament:

• First, prioritize possible progress over static incrementalism. While regulators attempt incremental controls on chip flows, companies need to fundamentally reimagine their approach to infrastructure sovereignty. Breaking free of the mindset of marginal improvements means refusing to be enticed by the endless cycle of new regulations and instead envisioning what's truly possible through rebuilding their capabilities.

• Second, treat hardware as the commodity it is. The very attempt to control chip flows through export restrictions ignores this fundamental reality. Technical hardware is a commodity that will flow where market demands dictate. Instead of pretending governments and policies can constrain this flow, companies should focus on building the best products in the world with these inputs.

• Third, let their people build. The loss of manufacturing expertise over the past 40 years happened partly because companies pulled their engineers away from actual building and focused instead on vendor management and global procurement. A proper InfraOps framework ensures your technical talent focuses on developing internal capabilities rather than managing external dependencies—exactly what was lost when companies abandoned vertical integration in the nascent silicon business.

• Fourth, and perhaps most critically, companies must increase their sovereignty. After decades of willingly surrendering control over critical operations, they can't pin their hopes on regulatory band-aids restoring their capabilities. True sovereignty—the kind that comes from actually building and controlling critical infrastructure—is fundamental for future success.

The good news is that the market is already signaling the value of this approach. The world’s leading technology companies—including Microsoft, Apple, and Meta—design their own chips. Similarly, Tesla’s vertical integration and command of its critical infrastructure enables it to consistently outperform other car manufacturers. 

The sovereignty premium is real, and it's growing as global supply chains become more uncertain. And the market will increasingly reward companies that control their technological destiny over those dependent on external providers.

The era of depending on regulations to protect competitive advantage is over. We can't turn back the clock and change the past. Decades of technology transfer and knowledge sharing can't be undone with regulation or wishful thinking. It’s time to be realistic about where we are and what we can do now.

The companies that will thrive in the coming decade won't be those praying for protection through export controls, but those building the internal capabilities to compete effectively in a globalized world. The time has come for G5000 companies to embrace this reality and act accordingly.