The Soft Antichrist Of The AI Age: What Everyone Is Missing About The US-China Arms Race

Authored by Patrick Feeley via Substack,

I. The word

Peter Thiel spent the back half of last year giving private lectures on the Antichrist, and the word has been loose in the discourse ever since. His political reading is the old one. The Antichrist is not a cartoon devil. He is the figure who arrives offering peace and safety, and who uses the fear of catastrophe to install one universal order. In Thiel’s telling that figure is a regulator. He stops technology in the name of survival. The price of the peace is stagnation.

I want the frame. I do not want his conclusion.

The more probable figure is not the one who promises safety. It is the one who promises growth. Most of the world does not lie awake about existential risk from machine intelligence. It lies awake about electricity, logistics, credit, tax collection, and a median citizen who is twenty four and needs work. To that world the universal offer is not a moratorium. It is a stack. Cheap. Financed. Hosted. Present. Already attached to the handsets, the ports, and the power.

That is the soft Antichrist. It does not arrive with a speech about ending history. It arrives as the only AI that can grow your GDP.

I am not writing theology. I am writing underwriting.

II. The comforting story

The market is telling itself a story. The AI race is a contest of models. America builds the best systems. China copies. Europe regulates. Capital therefore clusters around closed labs, chipmakers, and hyperscale cloud. The scoreboard is a benchmark table.

That story is not false. It is incomplete in a way that misprices power. It treats the richest customers as the only customers that matter. It treats evaluations as destiny. It treats national power as a software demo.

I disagree.

The race will be decided by who becomes the default operating system for the economies that still have the most growth left. Those economies will not adopt AI as a lifestyle product. They will adopt it as a growth tool. They will take the stack that is cheap, present, financed, and attached to what they already run. If that stack is Chinese, Beijing does not need to conquer anyone. It only needs to become expensive to leave.

III. Where the mass actually is

When Sargasso maps AI adoption across emerging GDP rather than across model releases, the picture is not the one the market is priced for.

Start with the denominator. On purchasing power terms, IMF projections for 2026 put China at about $44.3 trillion against $32.4 trillion for the United States. India is near $18.9 trillion. Indonesia clears $5.4 trillion and Brazil $5.2 trillion. Turkey is at $4.0 trillion, Mexico $3.6 trillion, Saudi Arabia $2.9 trillion, Egypt $2.6 trillion, Nigeria $2.4 trillion. Nominal dollars still flatter America, and by a wide margin. China’s nominal print is about $20.9 trillion against the same $32.4 trillion for the United States. Both numbers are true. They answer different questions. Nominal tells you who can buy foreign assets. PPP tells you how much physical and administrative activity there is to automate. For an adoption thesis, the second number is the one that matters.

Ray Dalio has been describing the political consequence in plain terms. He calls it a tribute system. A hierarchical order in which leaders travel to Beijing to acknowledge relative power in exchange for access and stability. He ties it to a growing view abroad that American security guarantees will not be honored under stress. I take the framing seriously. I do not treat it as scripture. Ports and rail were the first set of rails. Models will be the next set.

IV. The evidence is already in the download data

This part of the argument is no longer speculative.

Over roughly four years, the American share of model downloads on Hugging Face fell from about sixty percent to the mid teens by late 2025, according to reporting in The Wire China. Hugging Face’s own one year review of the DeepSeek moment is blunt on the composition shift. DeepSeek R1 became the most liked model on the platform in its history. The top of that list is no longer majority American. Baidu went from zero public Hugging Face releases in 2024 to more than one hundred in 2025. ByteDance and Tencent raised their release counts eight to nine fold. Of newly created models under a year old, downloads for Chinese models surpassed any other country, including the United States. Western startups and researchers now routinely fine tune Chinese base models because those are the largest open weights available.

The Wire China’s Southeast Asia reporting puts the commercial logic in local language. A Jakarta lab head said developers will always pick the cheapest one. A Malaysian founder said he wants the biggest model and there is no Western open source offering at that size. Chinese cloud providers were running thirty seven availability zones across six Southeast Asian regions against thirty across four for the Western field.

None of this shows up cleanly in a frontier benchmark table. All of it shows up in switching costs three years from now.

V. The institutional layer

Beijing is not leaving the volume layer to price alone. In July 2026 it stood up the World Artificial Intelligence Cooperation Organization in Shanghai, with twenty nine founding members. Public reporting names Russia, Kazakhstan, Pakistan, Indonesia, Brazil, and a broader set spanning Africa and Latin America among the signatories. The five year commitments attached to the body, as reported by Caixin and The Diplomat, are unglamorous and therefore serious. Training placements. Joint application centers with regional blocs. A weather early warning system deployed into dozens of countries.

Read that list as an underwriter. Training placements create the administrators who will write the next procurement. Application centers create the reference deployments. Weather systems create dependency inside a ministry that cannot afford an outage. Standards get set that way. Not with a better model. With a bureaucracy that has already learned one.

VI. America is running the right play against the wrong clock

Washington understands the problem. Executive Order 14320, signed July 23, 2025, created the American AI Exports Program to push full stack packages abroad. Chips, models, applications, cybersecurity, cloud, and data centers sold together. Analysis from the Institute for Progress argues the contested emerging markets that should sit at the center of that effort include Brazil, Egypt, Indonesia, Nigeria, Thailand, the Philippines, Malaysia, Vietnam, and Bangladesh. That is the correct map. It is essentially the map in this piece.

The tension is that the same government running an export promotion program is running an export control program, and the second one moves faster than the first. The UAE’s status was upgraded only in July 2026, and even then chip access was scoped to approved entities. Meanwhile Huawei has been shopping Ascend parts into the UAE, Saudi Arabia, and Thailand, and courting Egypt directly. Beijing has published a self sufficiency ambition on a near term horizon and is moving to expand domestic AI chip output.

Set aside whether the controls are correct on the merits. Underwrite the second order effect. Capability still flows outward through commercial relationships and distillation. American open efforts start constrained at home. The result is a one way street into the volume layer of the world economy, at exactly the moment when the volume layer is where the standard gets set.

Western discourse is spending its attention elsewhere. Alignment theater. Synthetic media. White collar displacement in rich cities. Those are real problems. They are also rich country problems. The quieter failure is dependency. Once an emerging state’s logistics, credit, schools, and revenue collection run on foreign models, switching stops being a procurement decision. It becomes a sovereignty decision. Sovereignty decisions do not get made on price.

VII. The map I would force into any serious strategy memo

When we screen a country the way we screen a company, we are not asking which model it admires. We are asking what it has already installed, who financed the installation, and what it would cost to rip out.

Asia. India, Indonesia, Vietnam, Malaysia, Thailand, the Philippines, Bangladesh, Pakistan, Kazakhstan, Cambodia, Laos, Sri Lanka.

Middle East and Gulf. Saudi Arabia, the UAE, Egypt, Turkey, Iran.

Africa. Nigeria, Ethiopia, Kenya, South Africa, Angola, Ghana.

Latin America. Brazil, Mexico, Argentina, Chile, Colombia.

These are not equivalent cases. India can build its own. The Gulf can simply buy, and is buying from both sides. Vietnam and Indonesia industrialize and will take whatever shortens the industrialization. Nigeria and Ethiopia need administration and power far more than they need chat interfaces. Brazil and Mexico live between Western finance and Chinese trade and will hedge accordingly. Pakistan and Kazakhstan sit on corridors Beijing already financed once.

The common variable is growth urgency. Growth urgency selects the stack that shows up, and it selects it quickly.

VIII. The throat

Compute is the oil of this cycle, and the supply chain has a throat. Counterpoint Research put TSMC at seventy three percent of the pure play foundry market in the second quarter of 2026. Its position at the leading nodes is more concentrated than that headline suggests.

You do not need an invasion scenario to price the leverage. You need governments that come to believe only one counterpart can reliably keep the chips, the cloud, the handsets, and the financing flowing. That belief is cheaper to create than a fab and harder to reverse than a tariff.

IX. Two futures

In the first, America wins the cathedral. Benchmarks stay American. Safety papers multiply. Closed models remain impressive and expensive. Emerging economies still buy the stack attached to Chinese devices, Chinese capital, and turnkey Chinese infrastructure. Global token volume follows global GDP, which is to say it follows the parish. The United States keeps the prestige and loses the installed base.

In the second, America treats emerging GDP as the actual battlefield. Competitive open weights exist and are hostable by states that want an alternative to Beijing without becoming a tenant of a single American lab. Energy, chips, and cloud are treated as national goods rather than as line items. The public companies that can genuinely deploy AI into durable operations are valued above the ones that can only demonstrate it.

Markets are priced closer to the first future than the evidence supports. That gap is the part I care about.

X. This is the same thesis, widened

In Pilot Purgatory I argued that AI works and capital is available, and that the binding constraint is organizational. Companies cannot absorb what they have bought. Forty two S&P 500 companies captured 312 percent of the index’s price return since ChatGPT while the other 458 captured 38 percent. Fifty eight percent of small and mid cap companies claim an AI strategy and under one percent describe implementation as mature.

That was a governance problem inside public companies. Widen the aperture by one order of magnitude and it is the same problem at the level of the state. Intelligence is being manufactured at declining cost. Absorption is the bottleneck. When Sargasso underwrites a company, the question is whether the organization can metabolize the technology it has already purchased. Run that question at the level of a country with weak administrative capacity and urgent growth targets, and the answer is worse. That is precisely why the party that clears the bottleneck earns something more durable than a product cycle. It writes the rails under the next order.

XI. What I would underwrite

If you underwrite AI as a feature race between rich country labs, you will be right about the models and wrong about the century.

Underwrite instead who owns the rails that India, Indonesia, Brazil, Mexico, Saudi Arabia, the UAE, Vietnam, Nigeria, Egypt, and Turkey will actually run. Underwrite who captures adoption where governance is thin and growth is urgent. Underwrite the closed labs and the national champions as though their real competitor is not the next chat interface, but a hierarchy that intends to make itself impossible to leave.

The AI race that matters is not who builds the smartest model in the richest city. It is who becomes the operating system for the economies that still have the most growth left in them.

This is not a recommendation. It is my map.

Sargasso Capital Management is a constructivist investment firm. This post is research and commentary. It is not an offer to sell or a solicitation of an offer to buy any security or interest in any fund.

Sources

1. Fortune, “Peter Thiel is delivering 4 private sold-out lectures at a club in San Francisco, about the Antichrist,” September 2, 2025. Link; Reason, “I listened to over 7 hours of Peter Thiel’s leaked Antichrist lectures,” October 14, 2025. Link
2. Visual Capitalist, “The World’s Largest Economies in 2026, Nominal vs. PPP,” using IMF World Economic Outlook projections. Link; IMF DataMapper, GDP based on PPP. Link
3. Fortune, “Ray Dalio says China’s ascent ushers in era of ‘tribute system,'” May 16, 2026. Link; Fortune, “Ray Dalio just finished a 10-day trip to China,” June 24, 2026. Link
4. The Wire China, “Surrounding American AI from the South,” June 21, 2026. Link
5. Hugging Face, “One Year Since the ‘DeepSeek Moment,'” January 20, 2026. Link
6. Reuters, “Twenty-nine countries sign agreement to establish global AI cooperation body,” July 16, 2026. Link; Caixin Global, “China Launches Shanghai-Based AI Governance Body With 29 Founding Nations,” July 17, 2026. Link; The Diplomat, “With New AI Governance Organization, China Seeks to Formalize Its Global AI Influence,” July 2026. Link
7. Executive Order 14320, “Promoting the Export of the American AI Technology Stack,” July 23, 2025. Link; Institute for Progress, “America’s AI Exports Program.” Link
8. Morgan Lewis, “BIS Upgrades UAE Export Control Status, with AI Chip Access Limited to Approved Entities,” July 2026. Link
9. South China Morning Post, “Huawei eyes export of AI chips to Middle East, Southeast Asia to rival Nvidia,” July 12, 2025. Link
10. Bloomberg via Free Malaysia Today, “Huawei pitches AI chips to Egypt in test of US tech diplomacy,” August 26, 2026. Link
11. RCR Wireless, “China aims to triple AI chip output,” August 28, 2025. Link
12. Counterpoint Research, Global Pure Foundry Market Share, Q2 2026. Link
13. Sargasso Capital Management, Pilot Purgatory, May 6, 2026. Link