International networking for the AI economy

Even the latest whizbang technology isn’t immune from the grubby realities of international power politics. Jared Cohen describes how AI is very much a part of the global economy and is very much intertwined with international cooperation and rivalries:

The AI industry also depends on a network of global commercial partners, including not only U.S. and Chinese technologies, but also Taiwan’s semiconductor fabrication plants, extreme ultraviolet lithography machines made in the Netherlands, and other critical supply chain inputs. Competition over AI has so far been dominated by debates about leading-edge semiconductors, but the next phase is also about geography and power. Specifically, where can the data centers that power AI workloads be built? And who has the capital, energy, and infrastructure needed to power the data centers where AI workloads run? [my emphasis] (1)

As has often been the case, the cutting-edge technology of the day is very much driven by military research and government funding of one sort or the other. Silicon Valley became the symbol of the tech industry and also of libertarian-billionaire ideologies like those of J.D. Vance’s patron Peter Theil. But to a large extent it was government funding both through military contracts and a world-class system of universities that made it possible. The same could be said of southern California’s aerospace industry. Prior to the Second World War, California was much more of an agricultural state, albeit an agricultural powerhouse because of its size and abundance of fertile agricultural land.

Cohen’s analysis focuses in particular on data centers:

Data centers are the factories of AI, turning energy and data into intelligence. Industry leaders estimate that a few major U.S. technology companies alone are expected to invest more than $600 billion in AI infrastructure, particularly in data centers, between 2023 and 2026. The countries that work with companies to host data centers running AI workloads gain economic, political, and technological advantages and leverage. But data centers also present national security sensitivities, given that they often house high-end, export-controlled semiconductors and governments, businesses, and everyday users send some of their most sensitive information through them. And while the United States is ahead of China in many aspects of AI, especially in software and chip design, America faces significant bottlenecks with data centers.

Bottlenecks also very much related to public infrastructure and government spending priorities:

Data centers are critical for the digital economy and AI. But the data center buildout is hitting a wall. The United States is home to the plurality of the world’s data centers, numbering in the thousands. Yet America’s aging energy grid, which powers those data centers, is under enormous strain from a complex set of factors, including rising electricity demand, delayed infrastructure upgrades, extreme weather events, and the complex transition to renewable energy. Meanwhile, surging data center demands driven by rapidly increasing AI workloads are exacerbating the grid’s vulnerabilities. [my emphasis]

Noting that the supply of data center space in the US is getting tight, he explains that this means current and future US supply chains supporting AI capabilities include China, Taiwan, Japan, Brazil, Canada, the Nordic countries (Denmark, Finland, Iceland, Norway, Sweden), India, and Vietnam.

Cohen concludes his analysis with this boilerplate US foreign policy platitude: “If the United States is successful [in accessing the needed data center capacity], it is more likely that the future world, in which machines play a greater role in daily life, will also be one with greater human prosperity and freedom.”

He mentions a number of tech industry talking points, like how AI is “contributing to the green-energy transition,” as pretty much every other company and industry claim to be doing. But a key factor in AI development is that AI functions require a tremendous amount of energy. For all the fantasies about androids, the neural networks in AI production require a tremendous amount of energy. Regular mammal brains including humans perform far more complex calculations than AI devices while using a fraction of the energy AI requires for the same process.

That could change. Just as fusion nuclear power production could become workable. Someday. But we’re still working on that one, too. In the early years of the atomic power age, Peaceful Nuclear Energy was going to soon provide abundant, cheap, and safe energy beyond our wildest dreams.

In the meantime, we’re still cooking the planet with carbon-based fuels. Climate activists like Bill McKibben (2) regularly reminds us of the available alternatives to CO2 fuels: batteries, wind turbines and solar panels.

Notes:

(1) Cohen, Jared (2024): The Next AI Debate Is About Geopolitics. Foreign Policy 10/28/2024. <https://foreignpolicy.com/2024/10/28/ai-geopolitics-data-center-buildout-infrastructure> (Accessed: 01-11-2024).

(2) McKibben, Bill (2024): Good News and Nothing But: One Day Only--Happy Earth Day. Bill McKibben Substack 04/22/2024. <https://billmckibben.substack.com/p/good-news-and-nothing-but> (Accessed: 01-11-2024).