The Information Machine

Chinese AI Models and Products Gain Structural Ground on US Rivals · history

Version 13

2026-06-29 02:24 UTC · 214 items

What

Chinese AI providers have gained ground on US rivals across pricing, market share, and benchmark performance through mid-2026. A J.P. Morgan report found Chinese AI models are up to 50x cheaper per token than US equivalents, and Chinese firms held over 45% of OpenRouter traffic by April 2026, up from under 2% in late 2024 [1]. A model operating anonymously as 'Owl Alpha' on OpenRouter — reported to be Meituan's LongCat-2.0-Preview — ranks #1 on Hermes Agent and #2 on Claude Code by usage, processing 10.1T monthly tokens with 242% monthly growth [8]. DeepSeek raised $7.4B at a $50B valuation [10], reportedly after a preview of Anthropic's Mythos model prompted its CEO to pursue capital to remain competitive [11], while also publishing a new inference optimization method achieving 60-85% faster per-user token generation [9].

Why it matters

The J.P. Morgan 50x price differential and the OpenRouter traffic trajectory from under 2% to 45%+ Chinese share in roughly 18 months provide the clearest quantification yet of how fast the cost-performance gap has shifted. The Meituan LongCat episode — a Chinese model reaching #1 agent benchmark positioning while operating under an anonymous name — shows the competitive pressure extends beyond pricing to performance in the agentic coding workloads where US closed models currently earn premium revenue.

Open questions

  • If 'Owl Alpha' is confirmed as Meituan's LongCat-2.0-Preview [8], does the practice of deploying Chinese models anonymously on US platforms complicate efforts to track or restrict Chinese AI adoption?

  • Does the J.P. Morgan 50x price differential represent a durable structural advantage for Chinese providers or a temporary strategy conditional on investor tolerance for losses? [1]

  • The Information reports Anthropic's Mythos preview prompted DeepSeek's $7.4B raise [11] — does a meaningful US frontier capability lead still exist, and how durable is it given Chinese models' cost-performance gains where most enterprise work happens?

  • Can CXMT scale its buried-wordline DRAM architecture toward HBM production needed for AI accelerators, and at what cost relative to Samsung, SK Hynix, and Micron? [13][14]

Narrative

Chinese AI providers have gained ground on US rivals across model pricing, enterprise adoption, and benchmark performance through mid-2026. A J.P. Morgan report found Chinese AI models are up to 50x cheaper per token than American equivalents, with Qwen, DeepSeek, and Kimi creating direct pricing pressure on OpenAI and Anthropic [1]. Chinese firms held over 45% of OpenRouter traffic by April 2026, up from under 2% in late 2024 [1] — a trajectory confirmed by Bloomberg's report that US model token share on the same platform fell from approximately 70% to 30% in roughly one year [2]. A UBS survey found 60% of enterprises monitoring AI budgets are migrating to cheaper or open-source Chinese models, with model routing — assigning routine tasks to cheaper models and reserving premium models for complex work — now the dominant enterprise cost strategy [3].

On model performance, Z.ai's GLM-5.2, a 753B-parameter mixture-of-experts model released June 16 under an MIT license, topped the Artificial Analysis Intelligence Index v4.1 at approximately $4.40 per million output tokens [4][5]. Nathan Lambert argues it is the first open-weight model to perform credibly as a general coding agent, comparable in significance to DeepSeek R1, and that it puts economic pressure on Anthropic's Claude Code revenue [6]. Zvi Mowshowitz credits it as the strongest open-weight model available but argues it trails frontier closed models substantially, is likely heavily distilled from Claude Opus, and occupies an awkward commercial niche [7]. A separate development: a model operating anonymously as 'Owl Alpha' on OpenRouter is reported to be Meituan's LongCat-2.0-Preview — a 1.6T-parameter MoE with 33B-56B active parameters and a 1M-token context window. It ranks #1 on Hermes Agent and #2 on Claude Code by usage, processes 10.1T monthly tokens, and has grown 242% month-over-month [8]. DeepSeek also published DSpark, an inference optimization achieving 60-85% faster per-user token generation by using a Markov head and confidence scheduler to verify only high-probability drafted tokens rather than full draft blocks [9].

DeepSeek raised $7.4B at a $50B valuation [10], with The Information reporting that a preview of Anthropic's Mythos model prompted CEO Liang Wenfeng to pursue the round to remain competitive [11]. This complicates the simple cost-driven-gain narrative: it suggests US frontier development still sets a pace that Chinese providers are spending to close. Rohan Paul frames the fundraise as evidence that AI competition has become a capital-intensive systems race requiring compute reserves, talent density, secure infrastructure, and product breadth [11]. DeepSeek plans to double all departments simultaneously and shift from model research to full-stack AI product development [12].

At the semiconductor layer, CXMT is approaching China's largest semiconductor IPO on Shanghai's STAR Market, with DRAM technology tracing to approximately 2.8TB of Qimonda documentation and roughly 7,000 patents, scaled toward the 10nm class [13][14]. SMIC's N+3 process reaches TSMC N6-class logic density via DUV multi-patterning but at higher cost [15], and Alibaba's T-Head chip unit tripled its registered capital from RMB 300M to RMB 1B [16]. US Commerce Secretary Lutnick raised with ASML leadership the concern that a banned EUV lithography machine reached China [17]; ASML categorically denied it, saying it tracks all ~314-340 EUV units worldwide and has never shipped one to China [18][19].

Timeline

  • 2024-06: OpenAI cuts API access for Chinese developers; Baidu, Alibaba, and Moonshot AI actively recruit displaced users. [27][28]
  • 2026-01: Chinese models begin outpacing US models as the primary growth driver of token consumption on OpenRouter. [20]
  • 2026-06: Chinese models reach 61% of OpenRouter weekly token market share. [29]
  • 2026-06-10: DeepSeek posts a role for MW-to-GW scale data center design, interpreted as a shift to heavy-asset vertical integration. [23]
  • 2026-06-13: MiniMax M3 (428B parameters, 23B activated) released with day-0 MXFP8 weights and block sparse attention 9x faster than M2.7. [21][22]
  • 2026-06-14: SemiAnalysis SMIC N+3 teardown finds TSMC N6-class logic density via DUV multi-patterning but at higher cost; Kirin 9030 Pro trails current flagship SoCs. [15]
  • 2026-06-16: DeepSeek raises $7.4B at a $50B valuation; founder Liang Wenfeng personally contributed ~$3B. [10][30]
  • 2026-06-16: Z.ai releases GLM-5.2 (753B parameters, MIT license), topping the Artificial Analysis Intelligence Index v4.1 with a score of 51 and a 1M-token context window. [4]
  • 2026-06-19: US Commerce Secretary Lutnick raises concern with ASML that a banned EUV tool reached China; ASML publicly denies it, saying it tracks all ~314-340 EUV units worldwide. [17][18][31]
  • 2026-06-22: Nathan Lambert calls GLM-5.2 the first open-weight model to work credibly as a general coding agent, arguing it puts economic pressure on Anthropic's Claude Code revenue. [6]
  • 2026-06-22: Zvi Mowshowitz credits GLM-5.2 as the best open-weight model but argues heavy distillation from Claude Opus, a persisting frontier gap, and an awkward commercial niche. [7]
  • 2026-06-23: SemiAnalysis reports CXMT is approaching China's largest semiconductor IPO; its DRAM technology traces to ~2.8TB of Qimonda documentation and ~7,000 patents, scaled toward 10nm class. [13][14]
  • 2026-06-23: Alibaba's T-Head chip unit triples registered capital from RMB 300M to RMB 1B and moves its corporate shareholder from DAMO Academy to PingtouGe (Shanghai) Electronics. [16]
  • 2026-06-26: The Information reports Anthropic's Mythos model preview prompted DeepSeek CEO Liang Wenfeng to pursue the $7.4B fundraise to remain competitive. [11]
  • 2026-06-26: UBS survey finds 60% of enterprises monitoring AI budgets are migrating to cheaper or open-source Chinese models; model routing is the dominant cost management strategy. [3]
  • 2026-06-27: Bloomberg reports US model token share on OpenRouter fell from approximately 70% to 30% in roughly one year. [2]
  • 2026-06-27: J.P. Morgan report finds Chinese models up to 50x cheaper per token than US equivalents; Chinese firms held over 45% of OpenRouter traffic by April 2026, up from under 2% in late 2024. [1]
  • 2026-06-27: DeepSeek publishes DSpark, achieving 60-85% faster per-user token generation via Markov head and confidence scheduler for selective draft verification. [9]
  • 2026-06-28: 'Owl Alpha' on OpenRouter reported to be Meituan's LongCat-2.0-Preview: a 1.6T-parameter MoE ranking #1 on Hermes Agent and #2 on Claude Code by usage, with 242% monthly growth and 10.1T monthly tokens. [8]

Perspectives

Nathan Lambert (Interconnects)

GLM-5.2 is the first open-weight model to perform credibly as a general coding agent, comparable in significance to DeepSeek R1; it puts economic pressure on Anthropic's Claude Code revenue and raises the prospect of US government restrictions on open models that Lambert views as a dangerous concentration of AI power.

Evolution: Consistent.

[6]

Zvi Mowshowitz

GLM-5.2 is the strongest open-weight model available but is very likely heavily distilled from Claude Opus, trails frontier closed models substantially, and occupies an awkward commercial niche — not cheap enough for bulk tasks, not strong enough for the hardest ones.

Evolution: Consistent.

[7]

Rohan Paul (@rohanpaul_ai)

Tracks Chinese AI competitive gains across OpenRouter traffic (US share at ~30%, Chinese firms at 45%+), J.P. Morgan's 50x price differential, UBS enterprise budget migration, DeepSeek's capital raise and full-stack pivot, and the unverified Meituan LongCat/Owl Alpha story; frames the Mythos-triggered fundraise as evidence AI competition requires compute, talent, infrastructure, and product breadth beyond model quality alone.

Evolution: Extended this pass with J.P. Morgan data confirming Chinese firms held 45%+ OpenRouter share by April 2026 (up from under 2% in late 2024) and a 50x price advantage; added the unverified Meituan LongCat/Owl Alpha story as further evidence of Chinese model gains on agent benchmarks.

[20][10][12][3][11][2][1][9][8]

SemiAnalysis

Covers Chinese AI and semiconductor infrastructure with technical qualification: SMIC N+3 reaches TSMC N6 density but at higher cost; CXMT's DRAM technology traces to Qimonda documentation and is approaching a major IPO; Alibaba's T-Head capital raise signals expanded AI chip ambition.

Evolution: Consistent; scope has expanded across the arc to cover memory (CXMT), AI accelerator hardware (T-Head), and logic chips (SMIC).

[21][22][23][15][13][16][14]

ASML

Categorically denies ever shipping an EUV lithography machine to China, saying it actively tracks all ~314-340 EUV units worldwide and that the reports are 'inaccurate and damaging to our reputation.'

Evolution: Consistent; denial is firm and backed by claimed active unit tracking.

[18][17][19]

Grant Harvey (The Neuron)

GLM-5.2's open-weight availability at ~$4.40 per million output tokens gives developers strategic optionality against closed-model vendor risk; the emerging AI stack will likely use frontier closed models for the hardest tasks and cheaper open models for repeatable or privacy-sensitive work.

Evolution: Consistent.

[5]

Milk Road AI

Open-source models will not win the AI race, and the dominant narrative around DeepSeek's early 2025 launch misread the episode as validation of open-weight approaches.

Evolution: Consistent; directly challenges the open-weight dominance thesis.

[24]

Developer/inference-market commentators (ollobrains)

The LLM market functions as an inference spot market — developers route to whatever model delivers the best cost-performance ratio at a given moment, making Chinese models' gains real but conditional on price.

Evolution: Consistent; structurally different interpretation of OpenRouter data than the competitive-shift framing.

[25]

Tensions

  • Nathan Lambert argues GLM-5.2 is the first open-weight model to match closed frontier performance in coding agent harnesses [6]; Zvi Mowshowitz argues it is very likely heavily distilled from Claude Opus, trails frontier closed models substantially, and occupies an awkward commercial niche [7]. [6][7]
  • The US government believes a banned EUV lithography tool reached China; ASML categorically denies having ever shipped one, saying it tracks all ~314-340 EUV units worldwide [17][18][19]. [17][18][19]
  • Rohan Paul and UBS enterprise survey data treat Chinese model adoption as a structural competitive shift affecting corporate budgets [3][2][1]; developer-community commentators argue the gains reflect spot-market price routing and are conditional rather than lasting preference [25]. [3][2][1][25]
  • US hyperscalers are projected to spend approximately 8.3x more than Chinese hyperscalers on AI infrastructure by 2027 [26]; UBS enterprise migration data, J.P. Morgan's 50x price differential, and GLM-5.2's benchmark performance suggest the dollar gap is not translating proportionally into capability or market separation [3][4][1]. [26][3][4][1]
  • Lambert and open-weight advocates treat GLM-5.2's release as evidence open-weight models are closing the frontier gap [6]; Milk Road AI argues open-source models will not win the AI race and the dominant open-weight narrative was built on a misread of DeepSeek's launch [24]. [6][24]
  • The Information's report that Anthropic's Mythos preview prompted DeepSeek's $7.4B raise [11] implies the US frontier still holds a lead Chinese players are spending to close; the same period's OpenRouter share data and J.P. Morgan pricing findings [2][1][3] suggest Chinese models are already winning the cost-performance competition where most enterprise work happens. [11][2][1][3]

Sources

  1. [1] 🇨🇳🇺🇸Chinese AI models are up to 50 times cheaper than their American counterparts on a per-token basis. — Rohan Paul Twitter (2026-06-27)
  2. [2] "the share of tokens used for US models on OpenRouter has collapsed" Bloomberg — Rohan Paul Twitter (2026-06-27)
  3. [3] UBS says 60% of companies now watching AI budgets are moving to cheaper models and open-source Chinese models — Rohan Paul Twitter (2026-06-26)
  4. [4] GLM-5.2 is probably the most powerful text-only open weights LLM — Simon Willison (2026-06-17)
  5. [5] 😺 GLM 5.2 brings 1M context — The Neuron (2026-06-22)
  6. [6] GLM-5.2 is the step change for open agents — Interconnects (2026-06-22)
  7. [7] GLM-5.2 Is The New Best Open Model — Zvi's AI Roundups (2026-06-22)
  8. [8] I’m hearing that "Owl Alpha", one of OpenRouter’s fastest-growing agent models, is actually Meituan LongCat-2.0-Preview — Rohan Paul Twitter (2026-06-28)
  9. [9] Fantastic, @deepseek_ai just published their new inference optimization method. — Rohan Paul Twitter (2026-06-27)
  10. [10] DeepSeek takes the crown as China’s most valuable AI startup after a massive $7.4B raise at a $50B valuation. — Rohan Paul Twitter (2026-06-16)
  11. [11] The Information reports that Anthropic’s Mythos preview spooked DeepSeek into fundraising. — Rohan Paul Twitter (2026-06-26)
  12. [12] Reuters: DeepSeek is going on a hiring sprint, aiming to double every department. — Rohan Paul Twitter (2026-06-26)
  13. [13] China’s CXMT Is Set to Challenge DRAM Incumbents — SemiAnalysis Twitter (2026-06-23)
  14. [14] CXMT (ChangXin Memory Technologies), China’s top domestic DRAM maker, is preparing for a major IPO on Shanghai’s STAR Ma... — reactive:chinese-ai-competitive-rise (2026-06-23)
  15. [15] Is SMIC N+3’s Metal Pitch Smaller than Intel 18A’s? — SemiAnalysis Twitter (2026-06-14)
  16. [16] Alibaba's core chip entity behind their PPUs, T-Head, just filed a business registration change lifting registered capit… — SemiAnalysis Twitter (2026-06-23)
  17. [17] ASML just became the center of a US-China chip fight after Washington said it fears a banned EUV lithography tool may ha… — Rohan Paul Twitter (2026-06-19)
  18. [18] ASML denies US government report that its EUV chipmaking tool ... — reactive:chinese-ai-competitive-rise
  19. [19] If an EUV machine reached China, three years of export controls failed silently. ASML denies it - they track all 314 in ... — reactive:chinese-ai-competitive-rise (2026-06-21)
  20. [20] American AI startups are routing far more app traffic to Chinese LLMs. — Rohan Paul Twitter (2026-06-08)
  21. [21] DAY 0 ALERT: @MiniMax_AI M3 is now available on HuggingFace & has been added to InferenceX. The M3 architecture has … — SemiAnalysis Twitter (2026-06-13)
  22. [22] Congrats to @vllm_project & @lmsysorg for releasing MiniMax M3 428B on both the CUDA & ROCm stack on day 0! Mini… — SemiAnalysis Twitter (2026-06-13)
  23. [23] DeepSeek is going heavy-asset. — SemiAnalysis Twitter (2026-06-10)
  24. [24] Open source models will not win the AI race (Save this). — Milk Road AI Twitter (2026-06-20)
  25. [25] The AI model market is turning into an inference spot market. Developers are routing to whatever model gives the best co... — reactive:chinese-ai-competitive-rise (2026-06-08)
  26. [26] China is growing very quickly in AI, but the scale difference is brutal, spending gap is enormous. — Rohan Paul Twitter (2026-06-22)
  27. [27] Chinese AI startups confront challenges as OpenAI ends API services in China — reactive:chinese-ai-competitive-rise
  28. [28] Chinese AI firms woo OpenAI users as US company plans ... - Reuters — reactive:chinese-ai-competitive-rise
  29. [29] Chinese AI models hit 61% market share on OpenRouter - LinkedIn — reactive:chinese-ai-competitive-rise
  30. [30] DeepSeek has completed over 50 billion RMB in financing at a valuation exceeding $50 billion, per The Information. Found... — reactive:chinese-ai-competitive-rise (2026-06-16)
  31. [31] The US says ASML's top chip tool may be in China, but how? — reactive:chinese-ai-competitive-rise