Reshaping America’s Economy for the Superintelligence Century with Jacob Helberg

Summary

• Jacob Helberg’s core macro call is that AI, deregulation and abundant energy could shift the U.S. from a consumption-led service economy toward a high-investment industrial one. CapEx is already above 2% of GDP and could double by next year; Helberg cites analysis saying AI added a full percentage point of GDP in the last year, while the economy grew 3%.

• Supply-chain security requires insulating Western producers from China’s ability to crush prices and later restore them. Helberg highlights 90% reliance on critical minerals refined in China, reliance on semiconductors manufactured in Taiwan and the $750 million DoD–MP Materials partnership. Its anchor buyer, offtake agreement and price floor are his blueprint for countering “classic monopolistic behavior” without leaving U.S. manufacturers begging Beijing for magnet licenses.

• Helberg believes AI will erase much of developing economies’ labor-cost advantage by giving American workers “superpowers.” Companies could employ fewer people, but he expects competitors and what he argues are unlimited human wants to push them toward tenfold output instead. Jevons Paradox implies cheaper, more efficient production creates more demand. The upside case is Sarah Guo’s framing—“What if the economy was $45 trillion?”—rather than mass technological unemployment.

• The “superintelligence century” could produce a second great divergence between early AI adopters and laggards. Helberg contrasts Europe’s fall from 65% of global GDP in the early twentieth century to roughly 15% today with a tech-forward Middle East where, he says, GDP per capita in the UAE and Israel exceeds France’s and is also higher than South Korea’s. His diagnosis is that Europe repeatedly “missed the boat,” and the EU AI Act may ensure it is not a first mover.

• The next platform contest is whether the Global South imports an American stack led by NVIDIA or a bundled Chinese stack built around Huawei Ascend and DeepSeek. Elad Gil presses the open-source risk; Helberg calls Meta’s efforts important but argues DeepSeek is “not really open source,” alleging that it lied about its compute capacity, has a billion-dollar cluster and distilled ChatGPT model weights. The strategic objective is not one licensing model but making the best American models widely used.

• Nuclear is Helberg’s preferred route to a potential doubling of U.S. electricity production in the 2030s, but financing time is the decisive variable. Large plants cost dozens of billions and can take seven years; Gil notes that protests and regulatory delays can turn a five- or six-year build into twelve years, while interest and legal costs compound. Helberg favors faster permitting and a clear CFIUS path for trusted foreign capital, alongside natural gas and clean coal in an “all-of-the-above approach.”

• The broader investment regime is meant to reward builders across every layer—energy, minerals, components, chips, data centers, models, apps and logistics. Helberg describes administration policy as “shock therapy” through deregulation, lower taxes and foreign investment, with autonomous transportation offering a way to leapfrog old infrastructure. Defense is another major spending area: global spending reached a record $2.7 trillion, but how governments allocate it will determine whether their forces are real capabilities or “paper tigers.”

Deep dive

1. China’s factory-floor power begins with control of supply chains

• Helberg starts with the exposed foundation beneath America’s innovation ecosystem: 90% reliance on critical minerals refined in China, reliance on semiconductors manufactured in Taiwan and a brittle system vulnerable to geopolitical disruption. Reshoring and partnerships with other countries are therefore prerequisites for letting U.S. builders keep building.

• His causal chain extends beyond bilateral trade. China imports African raw materials, manufactures domestically and re-exports globally; Belt and Road and its influence across Africa and Latin America flow from being “the world’s factory floor.” Correct the trade imbalance, he argues, and the leverage in those third markets also weakens.

• Gil’s useful clarification is that rare earths are not especially rare; deposits exist in the U.S., Canada and India. Helberg locates the bottleneck in refining: after announcing Made in China 2025 in 2015, China aggressively added capacity, flooded markets and squeezed competing refineries—including facilities in Tennessee, Arizona and Georgia.

• The proposed defense is commercial structure, not scarcity rhetoric. The $750 million DoD partnership with MP Materials combines an anchor customer with a price floor, preventing China from depressing prices until Western competitors fail and raising them afterward: “We can fix that with offtake agreements.”

2. AI could shift the economy from consumption toward production

• Helberg sees two forces arriving together: permitting, tax and energy reforms plus rapidly improving AI. America has long been 70–80% consumption-driven at times, with more than two-thirds of activity in services and roughly 10% in manufacturing; manufacturing’s GDP share remains flat, but he calls that a lagging indicator.

• The leading indicator is CapEx above 2% of GDP, which Helberg says will “probably” double by next year. He cites analysis saying AI added a full percentage point of GDP in the last year; with the economy growing 3%, he calls that substantial. National energy demand is also rising for the first time since 2008, after total electricity supply had flatlined, while energy infrastructure, raw industrials and defense spending are emerging as major areas of activity.

• Gil challenges the idea that service-economy maturation was inevitable, pointing to Germany and other Western industrial bases. Helberg agrees that America’s premise became self-fulfilling: globalization supplied “horizontal” growth, while the past seven years have brought renewed “vertical” growth through innovation.

• His historical specimen is Britain’s First Industrial Revolution: despite its smaller population, industrial output per capita exceeded China’s by more than 50 times, while China fell from about one-third of world GDP in 1800 to 7.5% in 1913. Technology, not population or wage levels, is the variable he believes AI can reactivate.

3. Productivity—not layoffs—is Helberg’s base case for AI

• Helberg reduces agentic AI to two outcomes: if one-tenth as many workers can perform a task, companies either shed labor or produce ten times more. He expects output expansion because any company that declines the opportunity will face a competitor that does not, while he argues that human wants are unlimited.

• Jevons Paradox supplies the mechanism: when technology makes a resource dramatically more efficient, its relative cost falls and total demand can rise rather than shrink. His optimistic conclusion is that AI will not replace humans altogether; it will give workers “superpowers” and broaden what each person can produce.

• That is the upside case behind Guo’s deliberately provocative question: “What if the economy was $45 trillion?”

4. Superintelligence reshuffles countries and technology stacks

• The century’s defining event, in Helberg’s framing, is not “the rising of the East” but “the rise of superintelligence.” Early adopters could create a second great divergence, leapfrogging slow adopters while collapsing the cheap-labor advantage that supported developing economies for fifty years.

• Europe is his warning case: its share of global GDP fell from 65% in the early twentieth century to roughly one-third in the 1980s and 1990s, then 15% today. He notes that Europeans blame the 1970s oil shock but argues that Europe missed the internet, digital and consumer-app waves—and says the AI Act and digital-services taxes keep Europe “shooting themselves in the foot.”

• The Middle East is the “total plot twist.” Helberg says GDP per capita in the UAE and Israel is higher than France’s and adds that it is also higher than South Korea’s. Tech-forward leadership, capital and cheap energy could create a new kind of U.S. partnership. Compute projects there could help offset American energy constraints, provided frameworks prevent China from accessing the clusters.

• Gil presses on Chinese open-source models, sovereign AI and state support. Helberg says Meta’s ecosystem matters, but the wider contest is distribution: countries may not need “super-fancy Blackwell chips,” yet whether they receive an NVIDIA-centered American stack or Huawei’s Ascend platform bundled with DeepSeek will shape global market share and influence. Helberg also argues that DeepSeek is “not really open source” because, in his account, it distilled ChatGPT’s closed model weights; he alleges that DeepSeek lied about its compute capacity and has a billion-dollar cluster.

5. Nuclear financing is the hinge between AI demand and energy abundance

• Helberg has “no doubt” nuclear offers the best path to abundant power. He cites two centuries of correlation between cheaper energy and growth, notes U.S. electricity costs are half Europe’s and sees committed Middle Eastern capital as a possible source for productivity-enhancing domestic nuclear infrastructure.

• Gil notes that nuclear still supplies roughly 17–18% of U.S. power despite the country having added roughly no capacity since the 1970s. He recalls five- or six-year projects becoming twelve-year builds, delaying revenue while interest, legal fees and overruns compound rather than rise linearly.

• Helberg’s answer is policy certainty: shorten construction windows, reduce regulatory barriers and signal that CFIUS will permit trusted foreign investors into critical energy infrastructure. France’s roughly 75% nuclear share of total energy supply is his proof that meaningful scale is possible even under a heavy regulatory burden.

• Helberg relays Elon Musk’s point that some statistics say data centers could require America to double overall electricity production in the 2030s—and reindustrialization might push demand higher. Guo notes commitments to large data-center projects in 2028 and 2029 and the possibility of matching a single large plant with a major data center. Helberg still pairs nuclear with natural gas and clean coal, but insists the route to scale “definitely runs through nuclear.”

6. A builder economy needs every layer, including logistics and defense

• Helberg’s operating map is a layered pyramid: energy, minerals, component manufacturing, semiconductors, data centers, models and applications. He says the U.S. is in a good position at most layers, but minerals, components and chips remain the largest exposure points.

• Transportation is another strategic area. China’s Belt and Road links African extraction, Chinese refining and global exports; Helberg wants the U.S. to reconsider the kind of large transportation and logistics investments it once made through projects such as the Panama Canal, using autonomous systems to “leapfrog old infrastructure.”

• Global defense spending has reached a record $2.7 trillion, with 60% coming from the U.S., China, Russia, India and Germany. The “trillion-dollar question” is what they buy: Ukraine shows AI and autonomy changing battlefield outcomes, while poor allocation can leave militaries as “paper tigers.”

• Helberg characterizes the administration’s last six months as “shock therapy” for domestic building—faster permits, lower taxes, deregulation and foreign capital. The aspiration is a builder-friendly country whose policy environment makes America “the best destination for capital.”