Washington shores up friends in the global chip war

It’s easy to see the future of microchips as a rivalry between the U.S. and China — that’s where the economic clout is, and what drives the global policy conversation.

But the world map of the industry is far more complex than that, and new U.S. policy might make it even more so.

The semiconductor supply chain is already somewhat distributed; no single country can, or likely ever will, manage its entirety. But the Indo-Pacific region — countries like Taiwan, Japan, China and South Korea — is essential to nearly every step of the manufacturing process. So as companies look to de-risk from China, the goal of the new policy is to give them U.S.-approved places to go.

A little-known part of the CHIPS Act hands the State Department $500 million to diversify and strengthen the global semiconductor value chain among allies. Over five years, the International Technology Security and Innovation (or ITSI) Fund will use some of that money for foreign aid in recruiting partner nations that are just getting started in this highly complex web, mainly as upstream and downstream suppliers to U.S. chipmakers.

Stephen Ezell, vice president for global innovation policy at the Information Technology and Innovation Foundation, told DFD that a partner designation is intended to be “like the Good Housekeeping seal of approval,” calling it “a really strong signal from the U.S. government that we have a deep relationship with this country, and we view it as a strong place to consider for nearshoring.”

Last week, Ezell gathered with officials from the Dominican Republic and U.S. CHIPS Program Office for the release of a first-of-its-kind readiness report. It gauged whether the Caribbean country could make the leap from its current base in electronics and medical devices into packaging semiconductors and manufacturing printed circuit boards.

The consensus? The Dominican Republic should be “a leading candidate” for U.S. funding considering its low labor costs, investor-friendly regulatory environment, liberalized trade regime, established technical workforce training program and other draws.

The State Department already has partnerships with five countries Costa Rica, Panama, Vietnam, Indonesia, and the Phillipines — to explore semiconductor industry growth opportunities, as a precursor for ITSI funding. It’s reportedly expected to pick seven in total. Beyond the Dominican Republic, Ezell told DFD that other promising candidates include Mexico, Malaysia and India (which is up next for an ITIF readiness report.)

There’s a name for this diplomacy-meets-business strategy: friendshoring. And it has been picking up some momentum. Just a year and a half ago, Emily Benson, a CSIS director focused on transatlantic trade, said she asked several Western Hemisphere nations about their efforts to claim business moving away from China, to minimal response. The partnership announcements are heralding “a new chapter,” she said.

“Countries are finally clicking on and seeing that there are huge potential benefits for them,” Benson said.

There are some baseline requirements for countries seeking to make the cut for ITSI or otherwise enter the global semiconductor race, according to Ezell. The U.S. is looking for reliable infrastructure; the cost from a power outage can escalate into the billions. A skilled workforce is another important factor, as is free trade policy — ideally mirroring Singapore’s zero tariffs on nearly all imports. Countries also need regulations that offer investors certainty, especially on permitting and environmental reviews.

One potential dealbreaker might be negotiations over export controls, the main tool that Washington has used with established allies in the semiconductor supply chain, like the Netherlands and Japan, to keep a lid on China’s chip industry. The U.S. is wary of investing in countries where advanced technology could leak to a foreign adversary’s military, Benson said — and export controls are thought of as a springboard for meeting Washington’s trade demands in the future.

As it plays out, the friendshoring strategy could be what sets the U.S. and China apart. China desires self-sufficiency over trading channels that run through trusted nations, according to Ezell. Consider the mineral cobalt: Chinese companies have set up a foothold in source countries like the Democratic Republic of the Congo, resulting in a scenario where a Chinese company that’s moved to Congo sells to another Chinese company, which sells back to buyers in China.

“They want to serve as their own market,” he added. “In that way, they’re probably less interested in building up the capacity of others.”

So what could throw off the strategy? The $500 million ITSI fund is a drop in the bucket, compared to $39 billion that the U.S. is investing in American-made chips. That limits the direct impact, yet both Ezell and Benson say it’s intended as seed money, signaling an opportunity to the market. How far the money goes largely depends on countries’ ability to attract private sector investment.

In this global competition, it’s also unclear how long newcomers will be satisfied with just supplying the chip industry rather than developing their own fabrication capacity. India has already made $10 billion in subsidies available to build out its domestic semiconductor ecosystem.

And for anyone familiar with Chris Miller’s “Chip War” (required reading for the CHIPS Program Office), the idea of relocating production to other countries, even friendly ones, might give some pause. The book chronicles how the U.S. came to lose its chipmaking presence, in part by underestimating newcomers like Taiwan and allowing manufacturing to shift to that region. What’s to stop history from repeating itself?

“It’s got to be an all of the above approach to establishing that manufacturing matters here in the United States,” Ezell responded. “It only wouldn’t happen again if we treat the situation differently than we’ve been treating it for the last 20 years.”

As Washington charges forward with discussions of how to regulate the biggest and most powerful AI platforms, there’s another issue emerging: How to address what Wharton professor Ethan Mollick calls an “AI-haunted world,” where smaller and cheaper AIs are everywhere, perhaps working in small teams “managed” by the biggest and most capable models.

As with much of the AI future, we’re already partly there.

“Some of the choice about whether we want this world where AI is ubiquitous has already been made,” Mollick writes. Plenty of smaller open-source large language models are already out in the wild, customized and running locally.

This raises obvious questions about rules, policy and privacy. In a Substack post today, Neil Chilson, senior research fellow at the Center for Growth and Opportunity, argues that we’ve already been partway down this road before when Washington thought about how to regulate the Internet of Things. That vision of a future filled with networked devices raised fresh privacy and security concerns and got quite a bit of attention in Washington several years ago.

Chilson, formerly acting chief technologist at the Federal Trade Commission who helped write the agency’s IoT report, is an optimist; he sees an “Intellect of Things” developing in which smaller AIs help us run our current technology better. (Think of how many instruction manuals you’ve actually read, and how many features of your devices go unused.)

He also points out there are real limits to the Internet of Things analogy; the “AI-haunted world” may be largely autonomous, whereas the IoT was all about connection and networking.

There’s another implicit limit, too. For all its research and white-papering, the government never really did much about the IoT, and the (very legitimate) concerns about its effects on our data ended up woven into the current stalled-out federal debate about privacy and data management.

AI is a higher-profile topic, so there may still be a chance to have this debate in public, for real, before our AI interns just start having the debate for us. — Stephen Heuser

At a panel on global digital disinformation today, a British regulator flagged an emerging problem with online safety: At the corporate level, there might be a little too much going on, and not enough focus on results.

Companies like Meta, X/Twitter and Google have been responsive to government and consumer pressure, but with a flurry of internal policies and features that can be hard to track and disentangle.

“Sometimes it can be almost overwhelming,” said Jessica Zucker, director of online safety for Ofcom, the U.K.’s digital regulator.

It’s a top-of-mind topic in the U.S., where tech CEOs last week offered the Senate a laundry list of new internal policies to prove they took kids’ safety seriously, without uniting behind any kind of broad new regulation or public accountability.

Speaking Monday at a conference hosted by Silicon Flatirons, a tech-policy think tank at the University of Colorado, Zucker pointed out that the number of features and policies that firms introduce can be hard to track, and not always provably effective. One challenge Ofcom wants to tackle, she said, is whether they actually work.

One global agency trying to track it is the Global Online Safety Regulators Network, a multinational group of agencies. However its membership is still small and scattered: Australia, the U.K., Fiji, France and South Korea … but notably, not the U.S., which has no federal agency, but where the vast majority of important platforms are headquartered. — Stephen Heuser