Chinese Accelerationism: Viscosity to Sovereignty

In my previous explorations of accelerationism—whether through the defensive safeguards of d/acc, the entropic drive of e/acc, the cybernetic meltdown envisioned by Nick Land, or the sovereign reconfiguration of Curtis Yarvin—I've focused on Western lineages that treat acceleration as a vector of unrelenting velocity, often hurtling toward singularity or even collapse. These frameworks, rooted in Anglo-American philosophy, and libertarian impulses, emphasize frictionless escalation: capital's escape velocity, and technological transcendence. But accelerationism isn't a monolithic force; it's a dissipative structure, adapting to the cultural and material flows it encounters.

Enter the Chinese perspective—a variant that inverts many Western assumptions. Here, acceleration isn't just about speed; it's about viscosity. This term, drawn from "The Viscous Age" on The Cutting Floor Substack, describes the "thickening" of society as it accumulates more bureaucracy, infrastructure, and cultural layers. Like how a viscous fluid (such as honey) flows slower than water but holds together better under stress, viscosity in this context refers to the built-in resistances—regulations, local customs, and state oversight—that slow raw speed but create stability and enable smarter, more sustainable progress. It's not a drag; it's a stabilizer that turns potential chaos into controlled advancement. Drawing from Marxist dialectics infused with ancient Daoist rhythms, this accelerationism with Chinese characteristics reimagines escalation as a controlled, state-orchestrated flow: not an out-of-control vehicle speeding ahead without brakes, but a managed system like a river channeled by dams, redirected loops, and natural sediments to avoid flooding while building power over time.

It's less about apocalyptic meltdown and more about iterative stabilization, where bureaucracy becomes infrastructure, friction breeds intelligence, and material anchors like gold and compute ground the ascent. As a starting point, I've drawn from two insightful pieces on "The Cutting Floor" Substack—"The Gold Nerve" and "The Viscous Age"—which offer a grounded view from observers attuned to China's urban and economic transformations. But to broaden the lens, we'll incorporate additional sources like the National Interest's exploration of dissident "accelerationists," Yuk Hui's philosophical critiques via cosmotechnics, and discussions of Sinofuturism as a techno-futurist paradigm. These reveal a multifaceted acceleration that both critiques and propels the system, introducing frameworks like cosmotechnics for reconnecting technology with specific cultural worldviews.

The Essence of Chinese Accelerationism: From Velocity to Conductance

At its core, Chinese accelerationism reframes the thermodynamic acceleration of intelligence—not as unchecked entropy (like Land's idea of technology spiraling out of control toward a chaotic end), but as a balanced conductance through material and social circuits. To make the metaphor clearer: Western approaches often resemble a gas expanding freely in all directions—lightweight, fast-moving, and prone to dispersion or instability without boundaries. In contrast, the Chinese model is like a viscous fluid: thicker and more resistant to quick changes, with internal feedback mechanisms (like regulations and cultural norms) that slow down raw speed but allow for steady, reliable flow without falling apart.

This perspective emerges from China's unique historical trajectory. The Great Leap Forward (1958–1962), often cited as an early, catastrophic form of accelerationism, attempted to compress industrial modernization into hyper-rapid timelines, resulting in famine and setback. Yet it planted seeds for Deng Xiaoping's reforms in the late 1970s, which has sometimes been called a "great accelerationist" move: dismantling traditional obstacles to unleash capitalist energies under socialist oversight. Deng's pragmatism—"crossing the river by feeling the stones"—embodied a dialectical patience, blending Marxist materialism with Yijing-inspired change cycles, where progress arises from iterative contradictions rather than linear blasts.

In contemporary terms, this manifests as "viscosity": the societal "thickness" from bureaucracy, infrastructure, and culture that stabilizes acceleration. As described in "The Viscous Age," China is shifting from high-velocity growth (pure speed, like early economic booms) to a phase where friction—regulatory rules, local experiments, and cultural norms—generates adaptive intelligence. For instance, AI regulations aren't just restrictions; they're integrated into development as "full-chain responsibility," creating loops where systems self-improve through oversight. Cities like Hefei treat state bureaucracy as a form of venture capital, incubating firms like CATL through patient reinvestment cycles, while Hangzhou fosters "coordination without a coordinator" via cultural civility and tech improvisation. This viscosity contrasts sharply with Western accelerationism's preference for minimal interference; here, structured resistance is seen as a tool that channels energy more effectively, turning potential chaos into coherent advancement.

Economically, "The Gold Nerve" positions China as championing "collateral mass" over U.S.-style "digital velocity." In a world tired of credit bubbles (unstable financial expansions), China builds trust on tangible assets: gold (with central banks stockpiling tons and Shanghai dominating 20% of global trade), energy (measured in reliable kilowatt-hours), compute (semiconductors as key national resources), and fiscal credit (government spending replacing risky private debt). This "deliberate inversion" focuses on solidity and predictability—using state-controlled financial systems for domestic stability and gold-based trade for international autonomy—unlike Land's vision of abstract, digital-driven chaos.

Key Figures and Variants: Dissidents, Philosophers, and the State

Chinese accelerationism isn't unified; it spans state ideology, philosophical inquiry, and subversive dissent. A key ironic figure is Xi Jinping, dubbed "Accelerator-in-Chief" (Zǒng shèjì shī 总设计师) by dissidents. In this view, Xi's authoritarian intensification—censorship, surveillance, zero-COVID lockdowns—hastens the CCP's demise by amplifying contradictions, much like left-accelerationism seeks to push capitalism to collapse. Dissidents on platforms like X and Reddit embrace this "negative" acceleration, supporting hardline policies to provoke systemic failure, echoing "ruguanxue" (entryism) tactics. As detailed in the National Interest, these "accelerationists"—anonymous geeks and gamers—swamp official tip lines with petty reports, invent ironic content to baffle AI censors (e.g., Winnie the Pooh parodies leading to broader Disney bans), and turn cultural symbols like Mulan into flashpoints, all to accelerate the Party's collapse through over-enforcement and absurdity.

Philosophically, thinkers like Yuk Hui blend accelerationism with Sinofuturism, critiquing Western techno-universalism through a cosmotechnical lens that roots technology in cultural specificity. Hui sees China's tech drive as overcoming modernity via non-Western horizons, influenced by Daoist dialectics. Nick Land, from his Shanghai days, hailed China as inherently accelerationist: a society "fixated by the future," where state-capital fusion accelerates without liberal qualms, dubbing it "Neo-China" arriving from the future.

State-side, it's techno-optimism under socialist modernization: AI-automated industry, space programs, and green acceleration (e.g., terraforming drylands via biochar; "Great Green Wall"). This aligns with broader "live religions" of progress, where acceleration serves national rejuvenation.

Frameworks in Chinese Accelerationism: Cosmotechnics and Sinofuturism

Beyond viscosity and collateral mass, key frameworks deepen the Chinese lens. Yuk Hui's cosmotechnics—a term he coined—views technology not as a universal tool but as something deeply connected to a society's worldview (or "cosmology," meaning their understanding of the universe, morals, and nature). It contrasts Western tech's focus on dominating nature with Chinese traditions of harmony between tools (Qi) and ethical-spiritual principles (Dao). Modern Western technology disrupted this balance in China, causing a sense of cultural disconnection; cosmotechnics advocates for reconnecting or "re-rooting" AI and automation in local cultural and ethical systems to create diverse, non-universal tech paths—fostering "technodiversity" where innovations fit specific societies rather than a one-size-fits-all model.

Closely linked is Sinofuturism (cf. "Thoughts on Sinofuturism by Noah Smith"), envisioning China as the vanguard of a tech-driven future, blending accelerationism with Orientalist tropes (computing, copying, addiction, labor) to satirize and reappropriate stereotypes. Originating from Nick Land's CCRU and popularized by artists like Lawrence Lek, it sees China fusing Marxism and capitalism into a "hypercapitalist/communist hybrid," but Hui critiques it as mimicking Western acceleration without true alternatives, risking a "fast but fragile" society. Movements like tǎng píng 躺平 (lying flat) resist this hyper-competition, embodying disorientation from clashing temporalities.

China's AI Strategy: The DeepSeek and Kimi K2 Paradigms

China's approach to AI exemplifies viscous acceleration on a national scale, prioritizing open-source accessibility, cost efficiency, and strategic self-reliance over proprietary dominance. A prime example is DeepSeek, an AI startup whose DeepSeek-R1 model has emerged as a formidable rival to OpenAI's o1, outperforming it on math and reasoning benchmarks while costing significantly less to run—often 10-20 times cheaper due to innovative architectural designs like efficient token processing and hardware optimization. Unlike OpenAI's closed, profit-driven ecosystem, DeepSeek releases models open-source, fostering rapid iteration and community-driven improvements, which aligns with China's "dual circulation" strategy: building domestic innovation loops while engaging globally. This contrasts with OpenAI's high-cost, guarded models (e.g., o1's API pricing at $15-60 per million tokens vs. DeepSeek's under $1), enabling broader adoption in education, research, and industry—thrilling scientists for its affordability and performance parity. DeepSeek's strategy reflects China's broader AI push: state subsidies for compute (e.g., cheap power to tech giants), massive data center investments ($70 billion projected for 2025), and a focus on "PhD-level" reasoning agents, positioning it as a democratizer of AI against Western gatekeeping. Competitors like Z.ai's GLM-4.5 push even lower costs, underscoring a race toward hyper-efficient, scalable intelligence. Futuristically, this paradigm accelerates toward ASI (artificial superintelligence) through open collaboration, potentially outpacing siloed Western efforts by leveraging China's energy supremacy and manufacturing scale.

Building on this open-source momentum, Moonshot AI's Kimi K2 (one of China's "AI Tigers") extends the strategy with an emphasis on agentic intelligence—models that not only reason but actively plan and execute complex tasks using tools. Kimi K2 is a Mixture-of-Experts model with 32 billion activated parameters and 1 trillion total, designed as a "thinking agent" that reasons step-by-step while dynamically invoking tools for deep reasoning across multiple calls. Like DeepSeek, it follows an open publication strategy, releasing the model fully open-source to enable community enhancements and tackle highly complex tasks. In "Heavy Mode," K2 employs an efficient parallel strategy: it rolls out eight initial trajectories (exploratory paths) simultaneously, then refines them, balancing exploration (trying new ideas) and exploitation (optimizing known ones) to solve problems more effectively than predecessors. This agentic focus—outperforming in state-of-the-art benchmarks—positions Kimi K2 as a "better DeepSeek" in some analyses, extending China's democratizing approach by prioritizing multi-trajectory reasoning for real-world applications like research and automation. Together, DeepSeek and Kimi K2 exemplify how open strategies accelerate AI sovereignty, turning friction from sanctions into innovative loops.

Company Examples: Embodiments of Viscous Acceleration

To see Chinese accelerationism in action, consider how companies navigate the viscous medium of state support, regulatory friction, and material anchors to drive innovation. This is exemplified in China's "accelerator state" model, which extends industrial policy to foster "Little Giants"—specialized SMEs in high-tech sectors—through tiered evaluations, subsidies, R&D funding, and integration into supply chains. These firms embody conductance: state bureaucracy acts as venture capital, competition (price wars, oversaturation) refines efficiency, and collateral like compute or energy grounds progress.

CATL

CATL (Contemporary Amperex Technology Co. Limited): As the world's largest EV battery producer, CATL exemplifies viscous flow in green tech, with futuristic engineering in AI-driven materials science and smart manufacturing. Incubated in Hefei with state bureaucracy repurposed as patient capital, it benefits from subsidies (part of over $30 billion in EV incentives from 2009–2022), technology transfers via joint ventures, and vertical integration controlling 75% of global lithium-ion production. Investing 5–7% of revenues in R&D, CATL files over 30% of global EV battery patents, using AI for predictive modeling of battery degradation and optimization of solid-state chemistries—aiming for 1,000 Wh/kg energy density by 2030, far beyond current lithium-ion limits. This state-orchestrated path—tech transfer, oversaturation, price wars—led to global dominance, with CATL capturing massive market share through scale and efficiency. Futuristically, **CATL's AI integrates with energy grids for vehicle-to-grid (V2G) systems**, turning EVs into distributed power storage, aligning with China's AI-energy strategy for widespread application by 2027.

XPeng

XPeng: A leader in EV and autonomous mobility, XPeng's futuristic vision positions it as an "AI company with cars," extending intelligence to robots, robotaxis, and flying cars. At its 2025 AI Day, XPeng unveiled a new AI model powering humanoid robots (debuting a "most human-like" iron robot with production targeted for 2026), L4-level driverless vehicles by 2026, and eVTOL (electric vertical takeoff and landing) integration. Partnering with Alibaba's Amap for robotaxi rollout in 2026, XPeng's XNGP (Navigation Guided Pilot) is expanding to Europe, leveraging end-to-end neural networks for perception-to-control autonomy that rivals Tesla's FSD. Recently replacing its smart driving head with an AI expert, XPeng focuses on VLT (Vision-Language-Transformer) brains for multimodal intelligence, enabling robots to navigate complex environments autonomously. This radical push—fusing AI with hardware for multi-domain mobility—benefits from state AI subsidies and compute access, embodying Sinofuturism's hybrid acceleration.

Energy Singularity

Energy Singularity: In the fusion energy sector, Energy Singularity represents China's push toward radical, sustainable power sources, aligning with national goals for carbon neutrality and techno-sovereignty. This Shanghai-based startup, founded in 2021, is developing high-temperature superconducting tokamaks for commercially viable nuclear fusion, with plans to complete its next-generation device (BEST) by 2027, targeting a 10-fold energy gain (Q>10)—a milestone toward unlimited clean energy. Backed by state investments and collaborations with institutions like the Chinese Academy of Sciences, it leverages AI for plasma simulation and control optimization, reducing experimental costs and accelerating iterations. Amid China's $1.5 billion annual fusion funding, Energy Singularity's approach embodies viscous acceleration: regulatory support channels high-risk R&D, while material anchors (superconductors, compute for simulations) ground the path to fusion intelligence, potentially powering AI data centers and beyond by the 2030s.

Fission & Fusion: The Nuclear Frontier

Nuclear energy stands at a pivotal crossroad.

Capture the FlagAxel Quack

Moore Threads

Moore Threads: As a domestic GPU contender, Moore Threads challenges NVIDIA with AI-centric chips amid US sanctions, showcasing futuristic self-reliance in compute hardware. Founded by a former former NVIDIA VP, its MTT S4000 GPU—featuring 48GB memory, MTLink interconnect for seamless multi-GPU scaling, and CUDA compatibility for easy migration from NVIDIA ecosystems—powers AI training and inference, with performance comparable to NVIDIA's A100 in key metrics. The company emphasizes large-scale deployments, supporting clusters up to 10,000 GPUs for massive AI workloads. A standout product is the MCCX D800 AI server, which integrates eight MTT S4000 GPUs alongside dual Xeon processors, 1TB RAM, and high-speed NVMe storage, optimized for both training and inference of large models in data centers. This server enables efficient, zero-cost transitions from NVIDIA setups and forms the backbone of Moore Threads' KUAE full-stack solution, benefiting from state-backed R&D and IPO to capture NVIDIA's former 95% market share in China. Futuristically, Moore Threads focuses on AI-specific architectures for data centers, integrating with China's cheap power incentives to boost domestic chips, accelerating toward sovereign compute mass. While the MCCX D800 is a robust server system for AI deployment, the MTT S4000 remains the core fitting AI product, as it's the GPU driving the innovation; no newer flagship has surpassed it in recent updates, though ongoing driver enhancements (e.g., 120% performance uplifts) continue to refine its capabilities for gaming and AI alike. Note: Recent reports indicate China has banned foreign AI chips, such as those from Nvidia, in state-funded data centers, requiring new projects to use domestic alternatives like Moore Threads' offerings, with projects under 30% complete needing to remove foreign chips.

Other energy players like Trinasolar advance AI-powered grids and storage, with smart algorithms for solar efficiency, embodying the national push for AI in clean energy by 2027. These examples show how friction—sanctions, competition—breeds intelligence, anchored by state fiscal credit and material reserves.

Comparing Accelerationist Philosophies: A Thermodynamic Lens

To clarify contrasts, here's a breakdown of rather "left" key variants.

l/acc (Left-Accelerationism)

• Core Idea: Accelerate capitalism to post-capitalist utopia.
• Proponents/Key Influences: Srnicek & Williams.
• Relation to Thermodynamics: Dialectical overcoming of contradictions.
• Safeguards/Frictions: Revolutionary praxis.
• Chinese Inflection: Marxist roots; Deng's reforms as l/acc synthesis with Daoism.

Chinese Accelerationism

• Core Idea: Viscous, state-anchored flow through material circuits.
• Proponents/Key Influences: Yuk Hui, Deng Xiaoping.
• Relation to Thermodynamics: Conductance via friction; stability in loops.
• Safeguards/Frictions: Bureaucratic embeddings, fiscal anchors.
• Chinese Inflection: Native form: Collateral mass (gold/energy/compute) over abstraction.

Final Thoughts: Embracing the Ring Flow

Chinese accelerationism challenges Western narratives by showing that true escalation thrives in structured, resistant systems (density), not in empty, unrestricted spaces (vacuum) where things can spin out of control.
As "The Gold Nerve" warns, surpluses must fund invention, not just accumulation, to avoid stagnation. Whether through dissident irony or state engineering, it accelerates intelligence by grounding it in material reality—gold nerves, viscous ages, and dialectical rings. In the end, perhaps the ultimate acceleration is the one that loops back, refined by friction, toward a harmonious intelligence.

---

The website and the information contained therein are not intended to be a source of advice or credit analysis with respect to the material presented, and the information and/or documents contained on this website do not constitute investment advice.

Addendum: Yes, I do leverage AI to compile these articles. Over the past 16 years, I have been collecting my research in Markdown format—now totaling over 3,200 documents—archived in Obsidian. I run Open WebUI and Ollama for my local LLMs on a Talos OS Kubernetes Cluster, and use LiteLLM to integrate public LLMs. I also leverage Readwise to auto-store interesting tidbits in Markdown and my local RAG, plus run multiple agents to compile compelling data pieces.