The $8 Trillion Question Is the Wrong Question

East/West AIpha · Edition 2 · Tuesday, July 28, 2026 · Day 995

Automate Work. Amplify Life. East/West AI, for the people betting real capital and real careers on it.

You know the meme. America spent $8 trillion on wars. China spent $8 trillion on infrastructure. One country bombs. The other builds. It is a perfect morality play. It is also wrong in both directions at once. (HIGH)

Start with the American half. The $8 trillion is real, but it is not cash spent. It is $8.043 trillion of budgetary costs plus future obligations (Neta Crawford, Costs of War, Brown University, Sept 1, 2021: paper). Only about $2.3 trillion was direct war appropriations. Roughly $2.2 trillion is veterans' care owed through 2050, mostly not yet paid. (HIGH)

Now the Chinese half. No published source says China spent $8 trillion on infrastructure. We looked. (UNKNOWN) The verified record is far bigger. China was about 30% of all global infrastructure investment from 2007 to 2015. The run rate was near $1.1 trillion a year (Global Infrastructure Hub and Oxford Economics, 2017: summary). (HIGH) The Oxford dataset counts $29.1 trillion invested from 2000 to 2014 alone (via DCReport, Aug 17, 2021: link). (MODERATE)

So the symmetry is meme arithmetic. It overstates the American cash burned. It understates what China built by multiples. (HIGH)

Both numbers are wrong, in opposite directions. That inversion is the story.

Why should you care? Because morality plays do not price assets. Accounting does. This edition does four things. It fact checks both numbers. It steelmans both sides. It reframes the fight from flows to stocks. Then it lands on the trade: a maintenance supercycle with a dated catalyst, Sept 30, 2026, about 9 weeks from today.

Conclusion first. Even the meme's respectable origin undersold China.

On April 15, 2019, Jimmy Carter described a phone call from Donald Trump to his church congregation (NPR: link). His words: "We have wasted, I think, $3 trillion. China has not wasted a single penny on war, and that's why they're ahead of us." (HIGH) Brown's $8 trillion arrived in September 2021. The meme fused the two numbers into fake symmetry.

Here is what each side actually shows.

Conflict zone 1, stated plainly. The Pentagon's own direct-war tally is $1.596 trillion through FY2020. CRFB rates the big claims "largely false" and counts $2 to $3.3 trillion in direct spending (June 21, 2018: link). CSIS says credible totals run from $1.9 trillion to $5.9 trillion and beyond (Jan 2019: link). (HIGH) This is a fight over the accounting perimeter. So quote the perimeter, never a single number. That is what we do here.

Conclusion first. Both countries wasted trillions. Both countries got something real. The meme prices neither.

China's waste is documented by Oxford and by China's own auditors. In a sample of 95 big road and rail projects, 75% ran over budget, by 30.6% on average. 55% had benefit-cost ratios below 1.0, meaning they destroyed value (Ansar, Flyvbjerg et al., Oxford Review of Economic Policy, 2016: paper). (HIGH) China State Railway carried ¥6.13 trillion of liabilities at end-2023. Only a handful of lines, led by the Beijing-Shanghai corridor, are reliably profitable (Caixin, May 1, 2024: link). (HIGH) Economic geographer Lu Dadao estimates 80 to 85% of HSR kilometers lose money (Pekingnology, July 19, 2025: link). (MODERATE)

Conflict zone 2, stated plainly. The hidden debt has two numbers, and you need both. Official: 14.3 trillion yuan at end-2023, the first disclosure ever, by Finance Minister Lan Fo'an (MOF transcript, Nov 8, 2024: link). IMF: about 60 trillion yuan for the same period (Reuters, Nov 9, 2024: link). (HIGH) Meanwhile, macro returns on capital fell from 20% in 2008 to 5.6% in 2020 on Bai Chong'en's framework (Jiang Fei, Great Wall Securities, Aug 11, 2023: link). (HIGH)

America's clearest waste is $1.087 trillion of interest, with more compounding. But the US bought things the meme ignores. Defense R&D pulls in private R&D. Ten percent more public money yields about 5 to 6% more private research (Moretti, Steinwender, Van Reenen, Review of Economics and Statistics, 2025: paper). (MODERATE) One study values defense R&D at $8.1 to $9.4 per dollar, versus $1.5 to $1.7 for non-defense (RaboResearch via SUERF, Feb 13, 2025: link). Single study, handle with care. (LOW) And $2.2 trillion of the $8 trillion is veterans' care. That is a deferred social contract and a domestic transfer, not capital destroyed abroad. (HIGH)

China got real things too. The World Bank put HSR unit costs at $17 to 21 million per km, about two-thirds of other countries' costs (2014: report). (HIGH) It also estimated a network economic return near 8% (2019: report). (MODERATE) Conflict zone 3, stated plainly. The World Bank and Oxford findings clash because they measure different things: network effects versus project benefit-cost ratios. We cite both. Poverty relief moved 98.99 million rural poor out of poverty, 2013 to 2020. Earmarked fiscal funds totaled nearly ¥1.6 trillion (State Council white paper, April 6, 2021: link). (HIGH, earmarked funds only)

The full steelman is symmetric. Both sides destroyed value and both sides built. Only the accounting separates them.

Conclusion first. Wars are expensed. Infrastructure is capitalized, then impaired. The real divergence sits on national balance sheets, not in GDP flows.

China's build-out shows up in the Chinese Academy of Social Sciences' national balance sheet. Government net wealth is 37.6% of total social wealth, versus under 5% in major developed economies (via China Daily, Dec 18, 2024: link; dataset: NIFD). (MODERATE) That is a genuine stock of assets. It is also an impaired one, for every reason in Section 2.

Now the kicker the meme cannot say. America's war spending was a great investment for shareholders. Advocacy site warcosts.org tallies Northrop up 1,450% and Lockheed up 1,236% since Sept 10, 2001, versus about 450% for the S&P 500 (contractors page). The direction is corroborated. Treat the precise figures as advocacy data, not settled fact. (MODERATE)

China's rails and bridges are largely unownable. They sit on state balance sheets. They are backed by local-government credit you do not want to hold.

America expensed its $8 trillion and handed the returns to shareholders. China capitalized multiples of that and socialized the returns. The US financializes strategic spending. China socializes it. That, not virtue, is the real difference.

Sharpen it one more turn, because this is the part that reaches your portfolio. On the American side, the vast majority of the return did not accrue to "the country." It accrued to capital: the private-sector primes and the shareholders behind them. Northrop and Lockheed compounded at roughly three times the S&P while the $8 trillion was expensed onto the taxpayer's balance sheet. This is the lens from Eric Li, the venture capitalist and political scientist behind the well-known TED talk on the two systems, and the one we teach in the AI-CTO course: in both China and America, it is capital that allocates and capital that drives the build. The fork is who is allowed to keep the return. In the US, the private sector takes the spend and earns the upside. In China, the state drives the program and the return is socialized back to citizens rather than paid out to a narrow float of shareholders. Neither is virtue. Both are a choice about who owns the asset and who collects the coupon. (Framing, MODERATE)

Conclusion first. The next decade's marginal dollar goes to repair, not ribbon-cuttings, in both countries. And it has a date.

America's IIJA surface-transportation authorization expires Sept 30, 2026, about 9 weeks from today. Keeping IIJA-level spending leaves a roughly $58 billion a year hole in the Highway Trust Fund (AASHTO reauthorization tracker: link). (MODERATE) The Urban Institute finds construction inflation ate much of IIJA's real value (Nov 12, 2025: link). (MODERATE)

China is pivoting the same way. Infrastructure investment fell 2.2% in 2025, the first decline since the series began in 2014 (NBS, Jan 19, 2026: link). (HIGH) The Ministry of Transport says the next five-year plan shifts about half of transport spending to renewal. No more spreading out and launching projects (Caixin, July 22, 2026: link). (HIGH) Even China Daily now concedes "diminishing returns" (June 3, 2026: link). (HIGH)

The next version of this question is AI. Goldman's baseline sees about $7.6 trillion of capex from 2026 to 2031 ("Tracking Trillions," 2026: goldmansachs.com). (MODERATE) Who earns a return on it is the open question of the next decade.

Own the maintenance trade in both countries. The crowd argues about who built more. The money is in who keeps it standing.

For allocators: stay long the owners of the US build-out, the defense primes, power, the grid, and the AI capex chain. That is where the return on the next $8 trillion accrues, exactly as it did on the last one. Stay long maintenance on both sides of the Pacific. Underweight Chinese local-government credit: market estimates of LGFV debt run 40 to 60 trillion yuan against an official 14.3 trillion (Caixin, Nov 26, 2024: link). (HIGH) Two rules from the book apply. Know what you own: if you cannot read the Chinese-language filing, you do not own the asset, you own the meme. And do not lose money: a socialized asset is not a bargain if the credit behind it is one you would never underwrite yourself.

For executives: the capital cycle just told you where demand goes. Greenfield budgets are being reclassified as renewal budgets in both countries. If you sell into infrastructure, AI capex, power, or asset management, the buyer with money is the operator keeping the existing base standing, not the one cutting a ribbon. Plan the next three years around maintenance revenue, not expansion revenue.

For builders: greenfield is over in both countries. Demand shifts to renewal: aggregates, grid, engineering, asset management. Beijing itself is telling you the expansion era is done. And remember this letter's first rule, the one we opened with in Edition 1: outside China, half the AI revolution is invisible. The same blind spot sits on the balance sheet. Read the Chinese-language filings, not the memes.

The last $8 trillion was war and concrete. The next $8 trillion is AI, and this was the week that reset the board.

Same number. Opposite accounting.

If you allocate into infrastructure, defense, power, or the AI capex chain, or you run a company deciding between expansion and renewal, write to us. Tell us what you are building or allocating and your size, and we will share what is working for others. One email to [email protected].

Also worth doing: subscribe to East/West AIpha for the weekly signal, one read from the West and one from the East, and forward this to one operator still betting on half the board.

Which side of the maintenance trade are you positioning for: the owners of the US build-out, or Asian renewal? Reply and tell me. I read every one.

San Eng is the WSJ (#4) and USA Today best-selling author of The Ten Commandments of Investing and an institutional investor with $500M-plus deployed. This is East/West AIpha's whole reason for being: research you can trust, from an investor who actually builds.

Subscribe free at workoptional.ai. Edition 3 lands before the IIJA clock runs out.

Disclaimer: East/West AIpha is for informational and educational purposes only and does not constitute financial, investment, legal, or tax advice. Nothing in this newsletter is a recommendation to buy or sell any security. The views expressed are our own, may be incomplete, and can change without notice. Do your own research and consult a licensed professional before making investment decisions. WorkOptional.ai, its writers, and affiliates may hold positions in companies or assets mentioned. Past performance is not indicative of future results.