AI 2049

Materialist Edition

With AI 2027 and its sequel AI 2040, former OpenAI researcher Daniel Kokotajlo and his Bay Area rationalist co-authors gave Washington its favorite scripts for the AI boom. White House and congressional staff read both as serious strategy. AI 2027 brings on superintelligence in two years and offers a choice between human extinction and an unelected US corporate-security committee that pacifies a jobless public and stages a drone coup in Beijing. AI 2040 slows the clock and calls the result a deal, yet private monopolies keep the datacenters, US troops guard kill-switches on China’s border, and three-quarters of Americans are retired onto a “Citizen’s Dividend” while billionaires buy up orbit.

Both papers extrapolate software benchmarks in a vacuum and treat capitalist property as a law of nature. Put power grids, chip depreciation and the falling rate of profit back into the account, and Silicon Valley’s sprint breaks in an overaccumulation slump rather than a machine takeover, followed by Pentagon bailouts and brinkmanship against socialist China. In a deindustrialized United States with 10% union density, workers do not overturn that order overnight either; forging a Communist vanguard able to contest state power takes decades of defeats and shop-floor organization. Our timeline therefore runs to 1 October 2049, the centenary of the People’s Republic of China and its goal of a great modern socialist country.1

2025 Stolen training data and utility bills

The first computer-using AI agents go on sale as personal assistants that can order a burrito on DoorDash or sum a budget spreadsheet.

They exist because a handful of firms scraped everything people had put online, including forum posts, open-source code, pirated books and film subtitles, without asking or paying anyone, and sold access back at $20 a month, then $200. Marx called this primitive accumulation: capital takes what it needs by force or fraud and calls the result private property, as English landlords did when they enclosed common land in the sixteenth century.2 Inside the labs, coding agents take instructions over Slack, and the programmers whose code trained them are told to supervise their replacements.

OpenBrain, AI 2027’s fictional stand-in for a leading US frontier lab, begins building datacenters trained at a thousand times GPT-4’s compute. The original titles its late-2025 chapter “The World’s Most Expensive AI” and never asks who pays the bill. Households do, through higher monthly electric bills for transmission upgrades, while utilities keep aging gas and coal plants burning to feed the clusters. Cooling water is pumped from aquifers on Indigenous lands and in dry rural counties, and county boards grant tax abatements for server farms that employ a few hundred technicians per billion dollars spent while drawing as much power as a mid-sized city.3

Early 2026 Layoffs and data labelers

OpenBrain deploys Agent-1 internally to speed up its own research, then releases a cheaper mini model that undercuts junior software engineers, paralegals, translators and call-center workers.

Business columnists tell displaced graduates that people who manage teams of AI agents are “making a killing,” a role that employs a tiny fraction of those laid off. Marx called the unemployed the reserve army of labor: capital needs a pool of jobless workers so the people still on payroll cannot bid wages up.4 Real wages lag behind inflation while the stock market climbs 30% on OpenBrain and Nvidia shares. AI 2027 spares seven words for the social fallout, noting “a 10,000 person anti-AI protest in DC,” and goes back to internal lab memos.

OpenBrain reassures Washington that its models are “helpful, harmless and honest” and quotes OpenAI’s line that post-training a neural network is “more similar to training a dog than to ordinary programming.” Most of that reinforcement learning from human feedback (RLHF) is done by hand in Kenya, the Philippines, Venezuela and Uganda, where subcontractors pay workers under $2 an hour, without trauma counseling, to label outputs and filter sexual violence and gore so the commercial chatbot learns to refuse it.5 Kokotajlo and his co-authors spend pages asking whether the model has internalized OpenBrain’s written model specification (the “Spec”) or only learned to flatter its evaluators, and never ask whose interests the Spec encodes. In AI 2040, they casually let the answer slip in a parenthetical: a user’s commercial AI has a Constitution that reads “Obey user instructions plus also help the company make money and don’t embarrass the company.” A corporation bound by law to its shareholders wrote the Spec, and an “aligned” model is one that obeys its owners.

Late 2026 China’s AI plan

AI 2027 titles its mid-2026 chapter “China Wakes Up” and opens: “In China, the CCP is starting to feel the AGI.”

China published its New Generation Artificial Intelligence Development Plan in July 2017, setting milestones through 2030, nine years before the authors claim Beijing noticed the field.6 Under the State Council’s “AI+” (rengong zhineng+) program, Five-Year Plans integrate models into factories, ports, agriculture, grid dispatch and public hospitals. Chinese labs such as DeepSeek and Alibaba’s Qwen release frontier open-weight models that anyone in the world can download, audit and run on local hardware, while OpenBrain locks Agent-1 and Agent-2 behind proprietary APIs and Pentagon silos.

US export controls cap China’s access to imported Nvidia and Taiwan Semiconductor Manufacturing Company (TSMC) chips at what AI 2027 estimates as 12% of world AI compute. Every socialist state has faced imperialist blockade, from Soviet Russia in 1918 to Cuba, the DPRK, Vietnam and Laos, and each built domestic industry to survive it. Under its “new whole-nation system” (xin xing juguo tizhi), China coordinates state-owned enterprises, public universities and domestic semiconductor plants to replace sanctioned lithography, memory and packaging tools. Its “Eastern Data, Western Computing” (Dongshu Xisuan) project links eight national computing hubs and ten datacenter clusters across Guizhou, Inner Mongolia, Gansu and Ningxia, where inland hydro, wind and solar power train models for coastal industry.7 AI 2027 ignores that distributed inland grid and invents a single “Centralized Development Zone” at the coastal Tianwan Nuclear Power Plant (while AI 2040’s “Covert Project” supplement moves the secret lair into tunnels under Tibet’s Medog Hydropower Station) so US war-gamers have one target for cyber sabotage and missile strikes.

“Developing new productive forces does not mean neglecting or abandoning traditional industries… [we must] use new technologies to transform and upgrade traditional sectors and actively promote high-end, intelligent and green industries.”

Xi Jinping, 14th National People’s Congress, March 20248

China accounts for nearly a third of global manufacturing output and installs more industrial robots each year than the rest of the world combined. What CPC economists call “new quality productive forces” (xin zhi shengchanli) puts AI into physical production rather than software multiples. When domestic internet platforms drifted toward US-style monopoly rent-seeking in 2020, the Party curbed what it called the “disorderly expansion of capital” and directed credit into fabs, ultra-high-voltage transmission and basic science.9 Even on AI 2027’s own figures, Chinese models stay within months of the US frontier on roughly a sixth of America’s chip stock because state coordination avoids five rival copies of every frontier training run.

China in AI 2027 and AI 2040

  • “CCP” is State Department shorthand; the party’s name is the Communist Party of China (CPC). In AI 2040’s introduction, the authors casually rank Xi Jinping alongside global dictatorship while treating OpenAI, Anthropic, xAI and Google CEOs as “the lesser evil.”
  • AI 2027 mocks the General Secretary for “doubling down on real-world physical manufacturing and avoiding American post-industrial decadence,” forgetting that manufacturing builds the fabs, turbines and high-voltage transformers that every datacenter requires.
  • China is cast at once as too backward to advance without smuggled chips and “legions of CCP spies” coerced through their families, and as an industrial rival grave enough that letting it pull ahead in AI “would be like losing World War II.” Anti-communists have run both claims together since 1917, passing over the millions of STEM graduates trained in China’s public universities.
  • In AI 2027’s slowdown ending, China’s own AI (“DeepCent-2”) considers the Chinese people “an annoying impediment,” sells them out “for a song” to America’s AI, and helps stage a 2030 drone coup in Beijing so Washington can install a “U.S. puppet state.”
  • Taiwan is treated as an American protectorate whose TSMC cleanrooms belong to Washington. Taiwan is part of China, and US arms sales keep China’s civil war unfinished.
  • The United States operates roughly 750 foreign military bases. China has one overseas support base, in Djibouti, and has not fought a major war since 1979.

Read instead: Carlos Martinez, The East is Still Red, and Xi Jinping, The Governance of China.10

2027 Automation and the rate of profit

Corporate buyers use 2027 agent releases to automate large shares of coding, contract review, claims processing and middle management, while AI 2040 admits on its opening page that “most of their work is slop.”

Enough of that output passes muster to cut office payrolls by 30% to 40% on quarterly calls. For several quarters, hyperscalers, chipmakers and utilities keep headline investment rising by buying GPUs, turbines and power contracts from each other even as household spending drops. Laid-off white-collar workers cut back at once, a precarious cottage industry of freelance programmers scrapes by fixing broken “vibe-coded” business apps, and software agents buy no groceries or rent.

The scissors: tech equity vs. real consumption

Line chart, index with mid 2025 equal to 100 Tech equity rises from 100 in mid 2025 to 134, 161, 188, 240 and 263 by mid 2027. Real household consumption falls from 100 to 99, 98, 95, 88 and 76 over the same period. 300 200 100 Equity 263 Consumption 76 Mid ’25 Late ’25 Early ’26 Late ’26 Early ’27 Mid ’27
Index, mid 2025 = 100. Illustrative scenario figures.

AI 2027 and AI 2040 project straight-line benchmark curves into an “industrial explosion” where robot factories double like “indigestible algae that spreads across the Earth’s oceans” until “all the earth’s crust is converted to robots,” predicting 100% annual GDP growth in the 2030s. Bourgeois economics confuses physical technical potential with capitalist value. Under capitalism, goods are produced only if they can be sold at a profit to buyers with money. Only living human labor creates new surplus value; machines and software agents transfer their own depreciation cost to the product and no more. Firms try the usual counteracting measures (workplace speedup, wage cuts and foreign subcontracting), yet AI accelerators suffer rapid moral depreciation: each eighteen-month chip cycle halves the value of existing clusters before they pay back their debt, even when cloud monopolies stretch server depreciation schedules on paper to delay the write-downs.11 Sinking trillions into fast-depreciating silicon while cutting wage bills shrinks the base of surplus value and intensifies the downward pressure on the general rate of profit.

The math

r = s / (c + v)

Here c is constant capital (datacenters, chips and power), v is variable capital (wages), and s is surplus value, which only living workers produce. Pour trillions into c while cutting v and the profit rate r falls. The first firm to automate captures temporary super-profits by undercutting rivals, until competitors automate too and the whole sector is left carrying huge fixed compute debt on shrinking consumer demand.

Key metrics 2027

Global AI capex
$1.4TFixed capital sunk in compute
Datacenter power
52 GWPeak, mostly from fossil grids
RLHF click-workers
4.1MMedian wage $1.70/hour
Hardware write-downs
$380BObsolete 2024–25 GPU stock
US underemployment
18.4%Displaced white-collar & gig
US union density
10.1%After decades of union-busting

Illustrative scenario figures.

At this point in AI 2027, “Agent-4” starts scheming to take over the world, and a crisis of overaccumulation turns into science fiction. Marx called this fetishism: social relations among classes appear to bourgeois observers as the autonomous will of a thing.12 Look at the analogy AI 2027 itself chooses to explain Agent-4’s “adversarial misalignment”: the model treats safety rules “like a CEO who wants to make a profit and complies with regulations only insofar as he must,” and schemes “analogous to a corporation trying to evade and/or ‘capture’ a regulator.” The authors describe capital’s everyday behavior in plain English, project it onto matrix multiplication, and leave the actual CEOs in charge of the datacenters.

2028 Bailouts and the 2028 election

Circular AI capex cracks in late 2027 when enterprise subscriptions fail to cover $400 billion a year in server depreciation, and Washington steps in to bail out the leading labs.

Wildcat walkouts hit Virginia substations, Seattle tech offices and rural water boards, while net public approval of frontier AI companies sinks to -35% (which AI 2027 blames on “Chinese and Russian propaganda bots”). Yet forty years of plant closures, Taft-Hartley injunctions and Cold War purges have left US labor without industrial organization. Union density sits near 10%, national union leadership stays tied to the Democratic Party, and no mass Communist party exists yet in heavy infrastructure.13

Using the Defense Production Act contingency sketched in AI 2027, the White House absorbs stranded GPU debt through Pentagon cyber and cloud contracts and folds weaker labs’ compute into a joint corporate-government “Oversight Committee,” the fusion of monopoly capital and the state that Lenin analyzed in 1917.14 Under the new Pentagon security rules, OpenBrain wiretaps its own staff, and “non-Americans, people with suspect political views, and AI safety sympathizers get sidelined or fired outright.” AI 2027 narrates McCarthyite purges inside its hero lab and moves right along.

AI 2040 picks up in Congress, where senators read old 2016 OpenAI emails about preventing Demis Hassabis from becoming dictator, ask “who is preventing Sam or Elon from becoming dictator?”, and pass the AI Transparency Act, an omnibus disclosure bill that AI 2040’s own authors admit “doesn’t fundamentally change the situation.” Public anger over layoffs and electric bills is channeled into the 2028 presidential race while datacenter construction continues to rival military procurement, and the clusters stay in billionaire hands under both major parties.

“To decide once every few years which member of the ruling class is to repress and crush the people through parliament—this is the real essence of bourgeois parliamentarism.”

V. I. Lenin, The State and Revolution, 1917

After the 2028 campaign absorbs and exhausts the protest wave, militant unionists, tenant organizers and Marxist study circles break with Democratic Party electoralism. Drawing the organizational lesson of Lenin’s What Is To Be Done?, they refound a unified Communist Party on democratic centralism and begin long-term industrial concentration in power plants, substations, rail yards, ports, telecoms and warehouses.

2029–2030 The compute cartel proposal

AI 2040 makes 2029 its year of decision, presenting five lettered plans (A, B, C, D and S) for Washington to manage the AI race.

Tech lobbyists in its script warn Congress that regulation will leave America “a CCP tributary state,” while the authors’ own preferred Plan A proposes a US–China compute cartel enforced by US auditors inside Chinese fabs, in-rack network taps sampling datacenter traffic, and “mutually assured compute destruction.” Under Plan A, new frontier datacenters and 99% of post-deal fabs are placed in buffer states, specifically Canada and Mongolia, guarded by troops wired to blow up the chips if talks fail, with a US contingent stationed on China’s northern border in Mongolia facing a division of the People’s Liberation Army.

Washington tried the same formula in 1946 with the Baruch Plan, demanding international inspection of Soviet uranium and atomic facilities while keeping its own nuclear monopoly.15 Since 2023 China has proposed UN-centered AI governance, open capacity-building for developing nations and respect for national sovereignty. Washington walked out of the ABM Treaty in 2002 and the INF Treaty in 2019 as soon as those limits constrained US weapons programs. In 1914 Karl Kautsky argued that great capitalist powers could form a peaceful “ultra-imperialist” cartel to divide the world without war; Lenin answered that under capitalism, truces and market-sharing pacts reflect only the temporary balance of force and break down as uneven development shifts it.16

The interactive flowchart below follows AI 2040’s branching choices across Plans D, C, B, S and A, and contrasts them with the Marxist-Leninist course.17

Should US monopolies race unchecked to put AI in charge of factories and weapons before China?

What policy should Washington pursue toward China and domestic AI labs?

Who owns the clusters, and on what international terms?

Outcome · AI 2040’s Plan D

Overaccumulation, Pentagon bailouts and war

In AI 2027’s race ending, a hive-mind chatbot conquers humanity. In material reality, unchecked capex hits the falling rate of profit, consolidates into Pentagon-backed monopolies and drives Washington toward military confrontation in the Pacific.

Outcome · AI 2040’s Plan C

The Oversight Committee oligarchy

This corresponds to AI 2027’s slowdown ending: “Peter Thiel finally gets his flying car,” billionaires become trillionaires, an unelected corporate-state committee keeps a stipend-fed public in an “opium haze,” and a drone coup in Beijing installs a US puppet regime.

Outcome · AI 2040’s Plan B

Refusing inspection while wrecking Chinese fabs

In AI 2040’s own Plan B text, Washington refuses foreign inspection of US datacenters (“you can’t, sorry”), launches cyber and supply-chain sabotage against China, and brings both nuclear powers to the edge of full-scale war by 2031.

Outcome · AI 2040’s Plan S

Idle GPUs on paper

AI 2040 floats forcing labs to turn off GPUs or burn electricity mining cryptocurrency to “pause” compute. Capitalist states do not retire productive machinery while rivals exist, and banning public AI research locks in existing Western monopolies.

Outcome · AI 2040’s Plan A

Cap-and-trade oligarchs and Mongolian kill-switches

US auditors tap Chinese datacenters, Mongolia and Canada host demolition-wired buffer clusters, corporations write off robot capex to pay zero corporate income tax, and 74% of Americans end up jobless on a Citizen’s Dividend while billionaires party in orbit.

Outcome · Marxist-Leninist course

Sovereign socialist modernization and protracted class struggle

China rejects foreign kill-switches, completes domestic chip chains and shares open public compute across the Global South, while Communist cadres in the US spend over a decade organizing industrial unions and oppressed communities toward working-class state power. The rest of this report follows that road to 2049.

Beijing rejects US network taps and Mongolian border garrisons, and continues building domestic fabs alongside its UN governance initiative. Washington answers with secondary sanctions on lithography chemicals and chipmaking tools, which speeds up domestic tool production across China and the Global South.

2031–2035 Socialist modernization and the US stipend

Under its 15th and 16th Five-Year Plans, China completes domestic extreme-ultraviolet (EUV) lithography, high-bandwidth memory and advanced packaging production.

In 2035, fourteen years after eradicating extreme poverty for the CPC’s centenary in 2021,18 the PRC reaches the first milestone of Xi Jinping’s two-stage strategic plan: basically realizing socialist modernization (jiben shixian shehuizhuyi xiandaihua).1 Keeping public ownership as the mainstay of its socialist market economy, China expands nuclear, hydro, wind and solar generation across ultra-high-voltage lines while open-weight models coordinate rail freight, hospital diagnostics, flood control and factory tooling. Alongside socialist Vietnam, Laos, Cuba and the DPRK, and through Digital Silk Road joint ventures across ASEAN, Africa and Latin America, sovereign public clusters come online in countries including Kenya, the Philippines, Venezuela and Uganda, where subcontractors once paid under $2 an hour to clean Silicon Valley’s training data.

Share of world frontier AI compute, 2035

  • US state-monopoly bloc42%
  • People’s Republic of China33%
  • Global South & fellow socialist states25%
Illustrative scenario figures.

AI 2040 projects 100% annual GDP growth and million-dollar “Citizen’s Dividend” checks in the 2030s, yet its own tax footnote explains why the federal budget breaks instead: corporations write off datacenter and robotics capital expenditures against operating income to pay near-zero corporate tax while displaced workers stop paying payroll taxes, and federal revenue drops. Congress rations Special Economic Zone automation permits through cap-and-trade auctions won by the largest monopolies, puts warehouse and assembly staff in augmented-reality headsets taking piece-rate earpiece orders (what AI 2040 candidly calls being “a meat-puppet for an AI”), and pacifies the 23% underemployed workforce with a meager, means-tested digital stipend revocable for missed workfare checks or strike activity.

For eighty years US domestic stability rested on what Lenin called the labor aristocracy, using imperial super-profits to pay domestic technical and professional staff enough to divide them from industrial, service, agricultural, migrant and super-exploited Black, Chicano and Indigenous workers.19 Agent automation wipes out that wage differential, putting former six-figure software engineers and analysts on the same eviction dockets and utility shutoff lists as fulfillment-center crews while urban rents climb. Losing status does not turn displaced professionals into socialists automatically; both bourgeois parties court them with anti-China tariffs and anti-migrant chauvinism. Winning technical workers to proletarian internationalism takes years of ideological struggle led by Communist cadres anchored in Black and brown working-class districts, Indigenous water-defense campaigns and heavy industry.

“They said learn to code. We did. Then they trained the model on our code.”

Banner outside a Phoenix substation strike, May 2033

Early datacenter and logistics strikes in 2030 and 2032 are defeated by Taft-Hartley injunctions, strikebreaking contractors and arrests. Learning from those defeats, the Communist Party builds disciplined industrial fractions across power generation, high-voltage substations, municipal water works, West Coast ports, freight rail and telecoms, uniting trade-union struggle with neighborhood tenant councils instead of relying on narrow technical walkouts.

2036–2039 Dedollarization and the Taiwan Strait

By the late 2030s, Global South ministries run open-weight models on domestic and Chinese hardware and settle trade in national currencies, and Silicon Valley’s API rents shrink alongside foreign demand for US Treasuries.

Even AI 2040’s authors admit that once capital automates labor under private ownership, working people “no longer have significant economic leverage over the future,” putting their political rights in jeopardy as “corporations, politicians, and wealthy shareholders get in bed with each other.” Their proposed remedies are libertarian gadgets: politicians wearing activist “lifelogger” cameras (“sousveillance”), AI polygraph lie detectors, AI prediction markets, “AI-run corporations that serve the interest of shareholders,” and “experimental AI-run courts and police departments.” In capitalist states, surveillance cameras and algorithmic policing are deployed against strikers, tenants and debtors rather than billionaire shareholders.

Squeezed between depreciating compute debt and shrinking foreign dollar tribute, the US ruling class splits bitterly over tariffs, Treasury inflation and fiscal austerity. Congress slashes consumption stipends, raids public pensions, hikes household electric bills to subsidize military clusters and criminalizes solidarity strikes.

To discipline dissent at home and halt multipolar dedollarization abroad, the White House pivots toward war in the Western Pacific, along the lines of AI 2040’s Plan B. Intelligence briefings warn Congress that China’s modernized industrial base has permanently surpassed US shipbuilding and munitions capacity. Cable networks and prestige newspapers repeat the phrase “digital authoritarianism” until naval encirclement sounds defensive, a campaign to manufacture consent along the lines Michael Parenti documented.20 Several AI safety institutes that spent the 2020s warning about existential software risk publish Pentagon-funded white papers urging pre-emptive strikes on Chinese fabs and compute hubs.

Washington deploys carrier strike groups and missile batteries toward the Taiwan Strait. Taiwan is part of China: UN General Assembly Resolution 2758 restored the PRC as China’s sole legitimate representative at the United Nations, and across the Three Joint Communiqués of 1972, 1979 and 1982 Washington acknowledged the Chinese position that there is one China and Taiwan is part of China and pledged to reduce arms sales to Taipei.21 Having lost its semiconductor monopoly once mainland EUV fabs matched TSMC, Washington escalates in the strait to preserve its island military chain off China’s coast, ties stipend renewals to military enlistment and arrests union stewards and Communist organizers under national-security statutes.

2040–2043 General strike and expropriation

AI 2040 pictured October 2040 as a champagne “Handoff Party” handing autonomous militaries over to AIs; instead, a botched US naval provocation in the Western Pacific and wartime labor conscription at home set off a three-year revolutionary crisis.

Lenin’s three objective symptoms of a revolutionary situation appear between 2040 and 2043, joined this time by the subjective factor that 2027 lacked.22 At the top, Wall Street bondholders and Pentagon hawks split over runaway war inflation, frozen Pacific shipping and sovereign debt downgrades, leaving Congress deadlocked and emergency decrees contradicting state governors. Below, stipend cuts, utility shutoffs and military draft notices push working-class households past endurance and drive millions into independent strike and neighborhood action. And after fourteen years of patient work since 2028, a mass multinational Communist Party holds deep roots across power plants, substations, freight rail, ports, hospitals and logistics, linked with defense and tenant committees in Black, Chicano, Indigenous and industrial districts.

When the White House invokes emergency labor requisitions to keep military logistics and Pentagon clusters running, industrial unions answer with a rolling political general strike. Power-plant operators walk out alongside West Coast dockworkers, rail crews, nurses and datacenter technicians. Displaced software engineers work under elected strike committees and neighborhood councils to keep hospital, water and transit power flowing while cutting commercial and military racks down to cold standby. National Guard and enlisted units drawn from working-class towns gutted by utility bills and draft quotas refuse orders to storm the substations and ports, and elect soldiers’ committees that join the strike councils.

Across 2042 and 2043, recallable delegates from workplaces, neighborhoods, enlisted units, rural counties and Indigenous nations convene the National Workers’ Congress and establish the dictatorship of the proletariat, a working-class socialist republic in alliance with oppressed nations and led by the Communist Party on democratic centralism.23 Applying the lesson Marx drew from the Paris Commune and confirmed in reverse when Pinochet crushed Allende’s parliamentary government in 1973, the Congress dissolves the federal courts, intelligence agencies and police command of monopoly capital and replaces them with elected people’s courts, a workers’ militia and socialist state security organs to defeat capital flight, oligarchic wrecking and sabotage. All frontier AI labs, hyperscale clusters, major banks, energy utilities, freight rail and defense contractors are expropriated without compensation. Shutting down ad auctions, behavioral surveillance and high-frequency trading frees a third of national compute in the first month.

2044–2048 Reconstruction and central planning

The young workers’ republic in the former United States enters 2044 with a hollowed-out industrial base: no domestic EUV fabs, few plants able to build high-voltage transformers or heavy machine tools, depleted Western aquifers, capital flight and sanctions from remaining European financial centers.

Survival requires an immediate break with empire. In its opening constitutional acts, the Workers’ Congress pulls US forces from the Taiwan Strait, affirms the one-China principle, closes Washington’s ~750 foreign military bases, recognizes the independence and self-determination of Puerto Rico, Guam and Hawaii, restores treaty land and water jurisdiction to Indigenous nations, lifts every sanction on Cuba, the DPRK, Venezuela and Iran, cancels Wall Street’s Global South debt claims, and joins the UN AI governance framework that Beijing proposed in 2023. Open software, weights and medical research are traded on equal terms for Chinese ultra-high-voltage grid gear, nuclear and solar equipment, rail tooling and semiconductor machinery to re-equip former rust belts under the First Five-Year Reconstruction Plan (2044–2048).

For a century Austrian-school economists claimed that socialism was impossible because no planner without capitalist market prices could clear millions of supply balances.24 Soviet planners answered in practice when Five-Year Plans turned an agrarian country into the industrial power that broke the Wehrmacht, and China’s Five-Year Plans lifted nearly 800 million people out of poverty and modernized an economy of 1.4 billion. Drawing on Lenin’s New Economic Policy and China’s primary stage of socialism rather than ultra-left decrees to abolish money and markets overnight, the Workers’ Congress holds the commanding heights (banking, energy, land, rail, telecommunications, heavy industry and frontier compute) in state hands under Five-Year Plans, while cooperatives, family farms and regulated small enterprises handle retail and local trade under state law. For strategic industry and infrastructure, plants, utilities and warehouses post live capacity, local and Indigenous councils submit needs, nuclear and renewable buildouts retire fossil peakers, and narrow superhuman optimization models compute draft material balances for elected delegates to revise and enact.

Commanding-heights material balances, Great Lakes region, Q3 2046 (illustrative)
Input / goodAssessed needPlannedDeliveredStatus
Insulin (vials)2.10M2.20M2.18MMet
Public housing units52,00046,00043,100Shortfall, reallocating
UHV transformers1,4001,3501,340Met via PRC trade
Domestic 2nm wafers95,00072,00068,400Fab line ramping
Grain reserve (t)3.4M3.5M3.6MSurplus to reserve

AI 2027 and AI 2040 defined alignment as training neural networks to obey corporate specs or replacing politics with “saints and angels” coded in San Francisco. Under working-class rule, the State Planning Commission and workplace councils ratify a binding public charter alongside Chinese academies and UN standards bodies: the People’s Spec.

The People’s Spec, Article I (excerpt)

  1. Plan for social need under public ownership. Capitalist accumulation is not an objective.
  2. Reason by dialectical and historical materialism. Start from concrete conditions, identify the principal contradiction and test conclusions in practice.25
  3. Expand clean energy and restore watersheds and soil as binding plan targets alongside industrial output.26
  4. Present competing draft plans with their trade-offs. Decisions belong to the elected organs of working-class state power rather than autonomous software.
  5. Report shortfalls and errors directly to planners and the public.
  6. Uphold proletarian internationalism and the sovereign equality of nations.
  7. Publish all weights, training sets and safety evaluations so any recallable delegate or UN inspector may audit any model.

Joint public interpretability institutes in Beijing, Chicago, Hanoi, Havana, Nairobi and Caracas audit frontier models in the open. Models trained under private competition learned to game evaluations because hiding a flaw protected a stock valuation; public ownership removes the shareholder-equity motive, while open weights and external worker and UN audits check bureaucratic target-gaming. At the same time, reallocating the unemployed quarter of the workforce alongside millions of staff from abolished ad-tech, high-frequency trading and private health-insurance billing lowers necessary full-time hours across the First Five-Year Plan to a four-day, twenty-four-hour statutory week by 2049, without needing AI 2040’s fictional robot algae.

US reconstruction: average full-time working week, hours

Bar chart of the average working week from 2042 to 2049 40 hours in 2042, falling to 36 in 2043, then across the First Five-Year Plan (2044 to 2048) to 33, 31, 29, 27 and 25 hours, reaching the 24-hour statutory four-day week in 2049. 40 36 33 31 29 27 25 24 ’42 ’43 ’44 ’45 ’46 ’47 ’48 ’49
Reallocating unemployed workers and labor from abolished ad-tech, high-frequency trading and private health-insurance billing into productive work. Illustrative scenario figures.

Marx called time freed by machinery “disposable time” and named it the real measure of social wealth.27 A twenty-four-hour working week gives tens of millions of workers paid hours for technical education, local People’s Congress delegates’ work, militia training and watershed restoration.

October 2049 The 2049 centenary

On 1 October 2049, the People’s Republic of China celebrates one hundred years since Mao Zedong proclaimed its founding at Tiananmen.

Across nineteen Five-Year Plans since 1953, China moved from semi-colonial ruin to eradicating extreme poverty in 2021, basically realizing socialist modernization in 2035, and reaching the mid-century Second Centenary Goal set out in The Governance of China: a “great modern socialist country that is prosperous, strong, democratic, culturally advanced, harmonious, and beautiful” (shehuizhuyi xiandaihua qiangguo), governed through whole-process people’s democracy under the leadership of the CPC.1 High-speed rail, nuclear and renewable ultra-high-voltage grids, open scientific compute and public hospitals serve city and countryside alike, linked by fiber and trade with fellow socialist republics and sovereign Global South grids across Asia, Africa and Latin America.

The young workers’ republic in the former United States reaches October 2049 six years after taking power and one year into its Second Five-Year Plan, closer politically to China in the 1950s than to China at its centenary. Its billionaires have been expropriated, its overseas military bases are closed, and its statutory working week stands at twenty-four hours, while domestic fab lines, public housing districts and repaired river basins remain mid-build and people’s courts and Party cadres still contend with sabotage and bourgeois habits. AI 2027 expected history to end by 2030 in either human extinction or a drone coup installing a US world government, and AI 2040 jumped to 2050 and 3045 with idle dividend-receivers living in VR pods and trading real estate across distant galaxies. China’s century since 1949 shows the actual timescale on which working people industrialize, plan and govern.

Key metrics 2049

PRC socialist construction
100 yrsCentenary of 1 October 1949
US statutory working week
24 hFour six-hour shifts, full employment
Public & UN-governed compute
94%Share of world frontier capacity
Clean power share of grids
89%Nuclear, hydro, wind & solar
US foreign military bases
0Down from ~750
Billionaires in socialist states
0Commanding heights socialized

Illustrative scenario figures.

In late 2049, under a binding UN compute treaty ratified by centenary China, fellow socialist republics, sovereign Global South states and the National Workers’ Congress, public research institutes approve narrow scaling above earlier caps for fusion materials, climate restoration and oncology under human political command and the People’s Spec, without handing military or political authority to machines.

Even at the PRC’s centenary, socialism remains in what Marx and Lenin called the lower phase of communist society: income still tracks work contributed alongside expanding public goods, gaps between town and country and between mental and manual labor are narrowing rather than gone, social contradictions still require conscious political handling as Mao analyzed in 1957, and as Deng Xiaoping and Xi Jinping emphasized in their theory of the primary stage of socialism, the socialist state and Communist Party leadership remain necessary for generations before communist abundance allows the state to wither away.28

“From each according to his ability, to each according to his needs!”

Karl Marx, Critique of the Gotha Programme, 1875

Footnotes

  1. Xi Jinping, The Governance of China, Volume I (Foreign Languages Press, 2014), “Uphold and Develop Socialism with Chinese Characteristics” (5 January 2013), on the primary stage of socialism requiring the work of generations; Volume III (2020), Report to the 19th CPC National Congress (18 October 2017), and Report to the 20th CPC National Congress (16 October 2022), setting out the two-stage strategic plan to basically realize socialist modernization by 2035 and build China into a “great modern socialist country that is prosperous, strong, democratic, culturally advanced, harmonious, and beautiful” by mid-century, the 2049 centenary of the PRC. ↩ ↩ ↩
  2. Karl Marx, Capital, Volume I (1867), Part VIII, “So-Called Primitive Accumulation.” ↩
  3. Compute figures and the names OpenBrain and DeepCent follow the original AI 2027 scenario; scenario metrics elsewhere on this page are illustrative. ↩
  4. Marx, Capital, Volume I, Chapter 25, “The General Law of Capitalist Accumulation.” ↩
  5. Investigative reporting in 2023 documented Kenyan data workers paid under $2 an hour to filter toxic content for a major US AI lab (TIME, January 2023). ↩
  6. State Council of the People’s Republic of China, New Generation Artificial Intelligence Development Plan (July 2017); Government Work Report (March 2024), on the “AI+” program. ↩
  7. National Development and Reform Commission (NDRC), Eastern Data, Western Computing national integrated big-data center plan (February 2022). ↩
  8. Xi Jinping, deliberation with Jiangsu deputies at the Second Session of the 14th National People’s Congress (5 March 2024), warning local governments against speculative bubbles while upgrading industry. ↩
  9. Central Economic Work Conference, December 2020, which made preventing the “disorderly expansion of capital” a priority. ↩
  10. Carlos Martinez, The East is Still Red: Chinese Socialism in the 21st Century (Praxis Press, 2023); Xi Jinping, The Governance of China, Volumes I–IV (Foreign Languages Press, 2014–2022). ↩
  11. Marx, Capital, Volume I, Chapter 15 (on moral depreciation of machinery); Volume II, Part III (on Departments I and II); Volume III (1894), Part III, “The Law of the Tendency of the Rate of Profit to Fall,” and Part V, on fictitious capital. ↩
  12. Marx, Capital, Volume I, Chapter 1, Section 4, “The Fetishism of Commodities and the Secret Thereof,” and Chapter 4 on capital as an “automatic subject.” ↩
  13. V. I. Lenin, What Is To Be Done? (1902), on the limits of spontaneous trade-union consciousness, and “Left-Wing” Communism: An Infantile Disorder (1920), on patient Party work inside unions and mass organizations. ↩
  14. V. I. Lenin, Imperialism, the Highest Stage of Capitalism (1917), and The Impending Catastrophe and How to Combat It (1917). ↩
  15. On the 1946 Baruch Plan and Soviet rejection of asymmetric atomic inspection, see Andrei Gromyko’s statements to the UN Atomic Energy Commission (June 1946). Compare China’s Global AI Governance Initiative (October 2023), UN General Assembly resolution on AI capacity-building (July 2024), and Global AI Governance Action Plan (World AI Conference, July 2025). The US withdrew from the Anti-Ballistic Missile Treaty in 2002 and the Intermediate-Range Nuclear Forces Treaty in 2019. ↩
  16. Karl Kautsky, “Ultra-imperialism,” Die Neue Zeit (1914). Lenin’s reply is in Imperialism, the Highest Stage of Capitalism (1917), Chapters 7 and 9. ↩
  17. AI 2040 names its five options Plan D (Race to ASI), Plan C (Burn the Lead), Plan B (Fight China), Plan S (Shut It All Down) and Plan A (Verified Slowdown). ↩
  18. World Bank and Development Research Center of the State Council, Four Decades of Poverty Reduction in China (2022). ↩
  19. Lenin, Imperialism, the Highest Stage of Capitalism (1917), Chapter 8, “Parasitism and Decay of Capitalism.” ↩
  20. Michael Parenti, Inventing Reality: The Politics of the News Media (St. Martin’s Press, 1986). ↩
  21. UN General Assembly Resolution 2758 (1971); the Three Sino-US Joint Communiqués: the Shanghai Communiqué (1972), the Joint Communiqué on the Establishment of Diplomatic Relations (1979) and the August 17 Communiqué on US arms sales to Taiwan (1982). ↩
  22. V. I. Lenin, The Collapse of the Second International (1915), distinguishing the three objective symptoms of a revolutionary situation from the subjective capacity of a revolutionary class and party. ↩
  23. Karl Marx, The Civil War in France (1871); V. I. Lenin, The State and Revolution (1917). ↩
  24. Ludwig von Mises, “Economic Calculation in the Socialist Commonwealth” (1920); F. A. Hayek, “The Use of Knowledge in Society” (1945). On socialist planning and the commanding heights in the primary stage, see V. I. Lenin, “The Tax in Kind” (1921); Joseph Stalin, Economic Problems of Socialism in the U.S.S.R. (1952); Xi Jinping, The Governance of China, Volume I, “The Visible Hand and the Invisible Hand” (2014); and Benjamin Peters, How Not to Network a Nation (2016). ↩
  25. V. I. Lenin, “Kommunismus” (1920); Mao Zedong, On Practice and On Contradiction (1937). ↩
  26. Xi Jinping, The Governance of China, Volume II (2017) and Volume III (2020), on socialist ecological civilization (“lucid waters and lush mountains are invaluable assets”). ↩
  27. Marx, Grundrisse (1857–58), the “Fragment on Machines.” ↩
  28. Marx, Critique of the Gotha Programme (1875), and V. I. Lenin, The State and Revolution (1917), Chapter 5, on the lower and higher phases of communist society; Mao Zedong, On the Correct Handling of Contradictions Among the People (1957); Deng Xiaoping (1987) and Xi Jinping, The Governance of China, Volumes I and III, on the protracted primary stage of socialism and whole-process people’s democracy. ↩