A country needs enough of them to meet its needs, must worry about securing supply, and produce as much of it as possible to benefit from high global demand. And, like oil, a small number of actors in a complex security environment control production. Microchips define geopolitics and remake world alliances.

And much as oil importers strive for “energy independence”, many large importers of microchips, such as the U.S., China and Japan, are striving for “chip sovereignty”—unplugging from the existing supply chain and capturing as much of the chip manufacturing process as they can.

The U.S., through the CHIPS Act, aims to spend over $50 billion to produce more advanced chips domestically. The EU is trying to double its share of chip production to 20% by 2030. Japan wants to triple its domestic chip revenue by 2030. China is investing tens of billions of dollars every year into chip design and production.

These countries have come to a conclusion that seems intuitive: the best solution to chip shortages is to produce chips domestically, thus getting around the problem entirely.

But they’re wrong. The solution to the chip shortage isn’t to increase onshore production, but rather to deepen supply chains, even if only between like-minded countries.

Existing semiconductor supply chains, developed over decades, span the globe. They’re incredibly complex: enterprises from around 25 countries participate directly in each segment of the semiconductor supply chains; another 23 countries provide support functions. Some advanced semiconductor products travel a distance longer than the Earth’s circumference before completion.

That’s too complicated to unwind in the short term. Building replacement ecosystems, technologies and human resources would require years, if not decades, to develop.

The existing global supply chain is also extremely efficient. High volume is needed for each segment to be economically sustainable. The current globalized supply chain allows for sustained operation even at low profits.

Bring that all in-country, and there won’t be enough local customers or consumers to sustain the whole supply chain in the long run.

Also, different countries own the different fundamental technologies required for chip design and production. Similarly, some raw materials may simply not be found in the required quantities within one’s borders (more than 60 elements of the periodic table are used in chip design) and will need continued interdependence.

Then there is the question of cost. Simply catching up to today’s leading-edge manufacturing leaders will require an annual investment of $30 billion over at least five years. $10-20 billion (or more) may be needed per year thereafter. No one government or enterprise could support this level of sustained investment. The current globalized supply chain creates the volume required to allow this investment.

But there is a way to keep the scale and efficiency of a globalized supply chain the security of a consolidated one: create a chip alliance with like-minded countries. Carefully selected countries can pool together financial and technological resources to build a sustainable and secure semiconductor supply chain—that may be more impervious to external shocks, whether economic or geopolitical.

One feasible chip alliance would be between the U.S. and other liberal democracies with a strong presence in the chip supply chain, like Japan, South Korea, the Netherlands, the U.K., Israel, Germany and France.

These countries collectively represent most of today’s semiconductor spending, control most segments of the semiconductor supply chain—the U.S., Japan, and the Netherlands together produce 90% of global semiconductor manufacturing equipment—and align on governance, rule of law, and human rights.

They also happen to complement each other. The U.S. lacks the photolithography and advanced node manufacturing that the Netherlands and South Korea specialize in.

Instead of following an isolationist approach that splits apart the existing supply chain, it makes sense for these countries to collaborate on a supply chain that minimizes disruption and that leverages existing strengths and interdependencies. In the short term, these countries could co-fund chip manufacturing and technology R&D using joint consortia. As a medium-term project, they could collaborate on investment screening, and intelligence and information sharing to prevent unwanted technology transfer. As the key long-term project, they could gradually and incrementally reconfigure the existing supply chain to reduce exposure to factors external to the alliance.

Stated like this, the benefits of an alliance for shared chipmaking may seem obvious. So why aren’t countries doing this?

In short, they hope that they can capture as much of the semiconductor industry as possible, perhaps worth as much as $600 billion in 2022. A nation might want to own as much of it as possible—sharing might dilute the profits. And any alliance would still be vulnerable to the risk of disruption—a push for cooperation will always be counteracted by the pull of autonomy.

In addition, transitioning to an alliance-based approach won’t be easy. Businesses will resist any change to the existing supply chain, as efficiency will be compromised, at least in the short term. Nations who lose out from the reconfiguration may retaliate by withholding materials, tools, and technologies that they control. Friction between allies may also threaten the alliance—in 2019, Japan blocked exports of chemicals required for chip manufacturing to South Korea due to a long-running dispute over wartime forced labor.

The idea of chip sovereignty has an instinctive appeal. But it’s logistically unfeasible and prohibitively expensive. Selective interdependence, not isolation, will solve our chip crisis.

Rakesh Kumar is a Professor in the Electrical and Computer Engineering department at the University of Illinois and author of Reluctant Technophiles: India’s Complicated Relationship with Technology.

The opinions expressed in Fortune.com Commentary pieces are solely the views of their authors, and do not reflect the opinions and beliefs of Fortune.