The Information Machine

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

Version 12

2026-06-27 18:09 UTC · 183 items

What

Chinese AI providers have advanced on US rivals across model performance, enterprise adoption, and capital through mid-2026. US model token share on OpenRouter fell from approximately 70% to 30% in roughly one year [1][2], and a UBS survey found 60% of enterprises monitoring AI budgets are migrating to cheaper or Chinese open-source alternatives [3]. DeepSeek raised $7.4B at a $50B valuation [8] — with The Information reporting that a preview of Anthropic's Mythos model prompted CEO Liang Wenfeng to pursue the round to remain competitive [9] — and plans to double its workforce while building full-stack AI products beyond model research [10]. At the hardware layer, CXMT is preparing China's largest semiconductor IPO [11], and a dispute over whether a banned EUV lithography tool reached China remains unresolved [16][17].

Why it matters

The reported link between Anthropic's Mythos preview and DeepSeek's capital raise introduces a complication to the simple 'Chinese models are winning' narrative: it suggests US frontier development still sets the competitive pace even as Chinese models win on cost. The collapse in US OpenRouter share from 70% to 30% in a year — confirmed by Bloomberg — moves the adoption story from a developer-traffic statistic to a structural question about enterprise budget allocation.

Open questions

  • The Information reports Anthropic's Mythos preview prompted DeepSeek's $7.4B raise [9] — does this mean the US still holds a meaningful frontier capability lead that Chinese players are spending to close, and how wide is that gap?

  • Does the collapse in US model OpenRouter share from ~70% to ~30% in a year [1] represent a durable structural shift, or is it conditional on Chinese model pricing staying low?

  • Did a banned EUV lithography tool actually reach China? ASML categorically denies having ever shipped one and says it tracks all ~314–340 EUV units worldwide [16][17][18].

  • 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? [12][11]

Narrative

Chinese AI providers have gained ground on US rivals across model capability, enterprise budgets, and semiconductor hardware through mid-2026. On OpenRouter, US model token share fell from approximately 70% to 30% in roughly one year [1], as Chinese models became the primary growth driver from early 2026 [2]. A UBS survey of enterprises monitoring AI budgets found 60% 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][1]. Chinese models including Qwen, DeepSeek, MiniMax, GLM, and Kimi meet enterprise cost requirements through local and cloud deployment options [3].

The open-weight model drawing the most attention in June 2026 is GLM-5.2, a 753B-parameter mixture-of-experts model from Z.ai released June 16 under an MIT license, topping the Artificial Analysis Intelligence Index v4.1 with a score of 51 [4] at approximately $4.40 per million output tokens [5]. Nathan Lambert calls it the first open-weight model to work credibly as a general agent in coding harnesses, comparable in significance to DeepSeek R1, and argues it puts economic pressure on Anthropic's Claude Code revenue [6]. Zvi Mowshowitz credits it as the best open-weight model available but argues it 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 work [7]. DeepSeek raised $7.4B at a $50B valuation [8] — with The Information reporting that a preview of Anthropic's Mythos model prompted CEO Liang Wenfeng to pursue the round to remain competitive [9]. Rohan Paul frames this as evidence that AI competition has moved beyond model quality to require compute reserves, talent density, secure infrastructure, product surface area, and enough capital to survive failed training runs [9]. DeepSeek plans to double staff across every department simultaneously and transition from model research to full-stack AI product development [10].

At the hardware layer, CXMT (ChangXin Memory Technologies) is preparing what is expected to be China's largest semiconductor IPO on Shanghai's STAR Market [11]. SemiAnalysis reports CXMT built its DRAM business on approximately 2.8TB of Qimonda technical documentation and roughly 7,000 licensed patents, using a buried wordline architecture now scaled toward the 10nm class [12]. Alibaba's T-Head chip unit tripled its registered capital from RMB 300M to RMB 1B and separated into a more independent entity [13]. A SemiAnalysis teardown of SMIC's N+3 process found it reaches TSMC N6-class logic density via DUV multi-patterning but at higher cost, with the Kirin 9030 Pro trailing current flagship SoCs [14]. US hyperscalers are projected to spend approximately 8.3x more than Chinese hyperscalers on AI infrastructure by 2027 [15], though UBS enterprise migration data and GLM-5.2's benchmark performance suggest the spending gap is not translating proportionally into market separation. Separately, US Commerce Secretary Howard Lutnick raised with ASML leadership the concern that a banned EUV lithography machine reached China [16]. ASML categorically denied it, saying it has never shipped an EUV machine to China, actively tracks all ~314–340 EUV units worldwide, and called the reports 'inaccurate and damaging to our reputation' [17][18][19].

Timeline

  • 2024-06: OpenAI cuts API access for Chinese developers; Baidu, Alibaba, and Moonshot AI actively recruit displaced users. [26][27]
  • 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. [2]
  • 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. [14]
  • 2026-06-16: DeepSeek raises $7.4B at a $50B valuation, becoming China's most valuable AI startup; founder Liang Wenfeng personally contributed ~$3B. [8][28]
  • 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. [16][17][19]
  • 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]
  • 2026-06-22: Rohan Paul notes US hyperscalers are projected to spend approximately 8.3x more than Chinese hyperscalers on AI infrastructure by 2027. [15]
  • 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. [12][11]
  • 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. [13]
  • 2026-06-26: The Information reports a preview of Anthropic's Mythos model prompted DeepSeek CEO Liang Wenfeng to pursue the $7.4B fundraise to remain competitive. [9]
  • 2026-06-26: DeepSeek plans to double all departments simultaneously, spanning AI R&D, algorithms, deep learning, full-stack development, and product roles. [10]
  • 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. [1]

Perspectives

Nathan Lambert (Interconnects)

GLM-5.2 is the first open-weight model to perform credibly as a general agent in coding harnesses, 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

Tracks Chinese AI competitive gains across OpenRouter traffic (US share now confirmed by Bloomberg at 70%→30%), UBS enterprise budget migration, DeepSeek's capital raise and full-stack pivot, while noting the 8.3x US hyperscaler spending advantage; frames the Mythos-triggered fundraise as evidence that AI competition has become a capital-intensive systems race requiring compute, talent, infrastructure, and product breadth.

Evolution: Extended this pass: the Mythos/fundraise connection and Bloomberg OpenRouter confirmation reinforce his structural competitive-gain framing and add a new layer — that Chinese players are responding to US frontier moves, not just leading on cost.

[20][8][15][10][3][9][1]

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][14][12][13][11]

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.

[17][16][18]

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 [16][17][18]. [16][17][18]
  • Rohan Paul and UBS enterprise survey data treat Chinese model adoption as a structural competitive shift affecting corporate budgets [3][1]; developer-community commentators argue the gains reflect spot-market price routing and are conditional rather than lasting preference [25]. [3][1][25]
  • Rohan Paul presents an 8.3x projected US hyperscaler spending advantage by 2027 as a structural US edge [15]; UBS enterprise migration data and GLM-5.2's benchmark performance suggest the dollar gap is not translating proportionally into capability or market separation [3][4]. [15][3][4]
  • Lambert and open-weight advocates treat GLM-5.2's release as evidence that 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 [9] implies the US frontier still holds a lead Chinese players are spending to close; the same period's OpenRouter share collapse and UBS enterprise migration data [1][3] suggest Chinese models are already winning the cost-performance competition where most enterprise work happens. [9][1][3]

Sources

  1. [1] "the share of tokens used for US models on OpenRouter has collapsed" Bloomberg — Rohan Paul Twitter (2026-06-27)
  2. [2] Chinese AI models hit 61% market share on OpenRouter - LinkedIn — reactive:chinese-ai-competitive-rise
  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] 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)
  9. [9] The Information reports that Anthropic’s Mythos preview spooked DeepSeek into fundraising. — Rohan Paul Twitter (2026-06-26)
  10. [10] Reuters: DeepSeek is going on a hiring sprint, aiming to double every department. — Rohan Paul Twitter (2026-06-26)
  11. [11] 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)
  12. [12] China’s CXMT Is Set to Challenge DRAM Incumbents — SemiAnalysis Twitter (2026-06-23)
  13. [13] Alibaba's core chip entity behind their PPUs, T-Head, just filed a business registration change lifting registered capit… — SemiAnalysis Twitter (2026-06-23)
  14. [14] Is SMIC N+3’s Metal Pitch Smaller than Intel 18A’s? — SemiAnalysis Twitter (2026-06-14)
  15. [15] China is growing very quickly in AI, but the scale difference is brutal, spending gap is enormous. — Rohan Paul Twitter (2026-06-22)
  16. [16] 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)
  17. [17] ASML denies US government report that its EUV chipmaking tool ... — reactive:chinese-ai-competitive-rise
  18. [18] 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)
  19. [19] The US says ASML's top chip tool may be in China, but how? — reactive:chinese-ai-competitive-rise
  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] Chinese AI startups confront challenges as OpenAI ends API services in China — reactive:chinese-ai-competitive-rise
  27. [27] Chinese AI firms woo OpenAI users as US company plans ... - Reuters — reactive:chinese-ai-competitive-rise
  28. [28] 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)