The Information Machine

China's AI Ecosystem Gaining Ground on the West · history

Version 4

2026-05-25 05:32 UTC · 180 items

What

China's AI competition with the West has moved from technical benchmarks into active US political investigation on multiple simultaneous tracks. Congressional scrutiny has expanded from probing US companies using Chinese AI (Airbnb and Anysphere, maker of the widely-adopted Cursor coding tool [1][2]) to directly investigating DeepSeek's ties to the Chinese Communist Party [3][4], while the US Senate separately asked agencies to probe DeepSeek's data security risks [5]. DeepSeek is advancing commercially regardless: its founder has declared an AGI goal ahead of a $10 billion funding round [9], and the company has permanently cut API pricing by 75% [10]. A Harvard Business School working paper frames China's underlying strategy as 'Diffusion-Forward' — deliberately prioritizing deployment scale over frontier benchmarks [16] — and a technical counter-analysis contests the widely-cited Huawei Ascend performance claims, arguing Huawei chips are a 'limited threat' to Nvidia [12].

Why it matters

The US response has shifted from restricting hardware supply chains to scrutinizing the software and open-source layer — and from investigating US adopters of Chinese AI to investigating Chinese AI companies themselves. This is a qualitative escalation: it signals that open-source model weights may not be the safe harbor that Airbnb CEO Brian Chesky argued, and that developer tools like Cursor are now inside the political perimeter alongside consumer platforms. Simultaneously, DeepSeek's AGI declaration and aggressive pricing make commercial and geopolitical competition increasingly inseparable.

Open questions

  • The Congressional probe now covers Anysphere (Cursor) [1][2], a developer tool embedded in software engineers' workflows at US companies and government contractors. Can the US restrict Chinese AI in coding tools without disrupting the Western developer ecosystem — and through what legal mechanism that would differ structurally from the TikTok ban?

  • DeepSeek's CCP ties and funding structure are under direct Congressional investigation [3][4][6], and the conflict between its ~$10B private round [7] and $45–50B government-linked valuation [8] remains publicly unresolved. Who controls DeepSeek — and does the answer change the open-source model calculus for Western companies?

  • Dr. Robert Castellano's technical analysis argues Huawei chips are a 'limited threat' to Nvidia [12], directly contesting community and industry reports that the Ascend 910C matches H100 performance [13]. Which performance regime — inference or frontier pre-training at scale — is the operative test for evaluating China's hardware independence?

  • Does DeepSeek's permanent 75% price cut [10] reflect genuine technical cost efficiency, subsidized pricing with implicit state backing, or both — and does the distinction matter for Western competitors if the competitive effect is identical?

Narrative

The US political response to China's AI rise has escalated from abstract export control debates to specific named investigations of both US adopters and Chinese AI companies themselves. House Republicans, through the Select Committee on China and the House Homeland Security Committee, announced a joint investigation into Airbnb and Anysphere — the company behind the Cursor AI coding assistant [1][2]. Anysphere's inclusion marks a significant expansion of the investigation's logic: where Airbnb's case centered on consumer data exposure through a Chinese AI API, Cursor is embedded in the development workflows of software engineers at US companies and government contractors, raising supply-chain-style concerns about code, proprietary algorithms, and intellectual property exposure. Separately, House investigators have opened a direct probe of DeepSeek's ties to the Chinese Communist Party and its funding structure [3][4], and US senators asked government agencies to investigate DeepSeek's data security risks [5] — establishing bipartisan, bicameral pressure on the executive branch to act. A detailed ownership analysis of DeepSeek's corporate structure [6] has emerged alongside these investigations, though the fundamental question of whether DeepSeek operates as a genuinely independent company or as a strategic national AI asset remains publicly unresolved. The conflict between DeepSeek's approximately $10 billion private funding round [7] and a separate Chinese government-linked investment implying a $45–50 billion valuation [8] sits at the center of that unresolved question.

DeepSeek has advanced commercially and strategically regardless of the political scrutiny. Its founder publicly declared an AGI goal ahead of the company's fundraise [9], positioning DeepSeek not merely as a commercial AI provider but as a mission-driven organization pursuing artificial general intelligence — a framing that both elevates its stated ambition and complicates Western assessments of whether it is a company or a state project. The company has also made permanent a 75% reduction in its API pricing [10], intensifying cost pressure on Western AI labs and downstream developers already grappling with the competitive availability of capable Chinese open-source models. Hugging Face's one-year retrospective on the original 'DeepSeek moment' [11] underscores how rapidly the competitive landscape shifted — the current political tensions are the downstream consequence of a single model release that restructured how the global AI industry perceived cost and capability.

The hardware competition has acquired a new critical voice. Analyst Dr. Robert Castellano's technical deep-dive argues that Huawei's latest AI chips represent a 'limited threat' to Nvidia [12], directly contesting widely-circulated reports that the Ascend 910C matches or approaches H100 performance. Community discussion, by contrast, has treated the 910C-matches-H100 claim as established fact [13]. The disagreement turns on methodology: inference workloads and frontier pre-training impose qualitatively different demands on chip architecture, and the 60% H100 inference performance figure benchmarked by DeepSeek's own research [14] does not settle whether Ascend hardware can support the most demanding training runs that determine state-of-the-art model quality. Stanford's 2026 AI Index provides related context: the US spends 23 times more than China on AI [15] — suggesting that even if Ascend chips close much of the per-chip performance gap, the aggregate US compute advantage remains substantial.

A Harvard Business School working paper introduces a strategic framing that cuts across both the bullish-on-China and US-still-leads narratives: China is pursuing a deliberate 'Diffusion-Forward AI Strategy,' prioritizing broad deployment, adoption, and application-layer integration over frontier capability benchmarks [16]. China's May 2026 guidelines — released jointly by three of its most powerful regulatory bodies — to regulate and boost innovative development of AI agents [17][18] fit this framing precisely: state attention is directed at enabling widespread deployment and commercial application rather than the frontier pre-training race that drives US lab competition. This strategic divergence helps explain why metrics like download share, patent filings, and enterprise deployment favor China even as the US maintains a measured benchmark performance lead: the two countries may be optimizing for different definitions of AI success, using different scorecards.

Timeline

  • 2025-02-05: TrendForce reports DeepSeek research showing Huawei Ascend 910C reaches 60% of NVIDIA H100 inference performance [46][14]
  • 2025-04-29: TrendForce: Huawei's DeepSeek all-in-one machine achieves 60-70% of NVIDIA H100 performance at a competitive price point [47]
  • 2025-08-05: US Senators ask government agencies to probe data security risks posed by DeepSeek [5]
  • 2026-01-01: China's raised bar for AI patent quality takes effect, directly addressing volume-vs-quality criticisms of China's patent lead [48]
  • 2026-03-09: China declared world's largest holder of AI patents [49]
  • 2026-03-24: NBER working paper on AI patent measurement in US and China circulated, framing the quality debate as an empirical question [50]
  • 2026-04-07: Harvard Business School working paper on China's 'Diffusion-Forward AI Strategy' published, arguing China deliberately prioritizes deployment scale over frontier benchmarks [16]
  • 2026-04-14: Stanford 2026 AI Index published: China leads in publications and patents; US AI performance lead quantified at 2.7%; US spends 23x more than China on AI [23][24][25][26][27][28][15][29]
  • 2026-04-17: Reuters reports DeepSeek raising approximately $300M at ~$10B valuation; DeepSeek founder declares AGI goal ahead of the fundraise [51][52][53][9][7]
  • 2026-04-30: SenseTime fully open-sources SenseNova U1 with VAE-free NEO-Unify architecture for unified multimodal understanding and generation [54][55][56][57]
  • 2026-05-08: Three of China's most powerful regulatory bodies release joint guidelines to regulate and boost innovative development of AI agents [17][18][58]
  • 2026-05-17: Analysis highlights China's lead in AI patent filings, investment approaching US private-sector levels, and high public AI enthusiasm [19][20]
  • 2026-05-20: SenseNova U1 Lite series announced; Airbnb CEO Brian Chesky tells Bloomberg the US is 'misunderstanding' Chinese open-source AI model use [21][30]
  • 2026-05-21: Bart Collet notes five Chinese AI labs have released open-source frontier models; community commentary on Silicon Valley's Chinese open-source dependency intensifies [37][59]
  • 2026-05-22: Coordinated wave of Chinese open-source models natively optimized for Huawei Ascend chips reported; Hugging Face Spring 2026 report signals Chinese models overtook US in download share [38][44][60]
  • 2026-05-23: House Republicans expand investigation to cover both Airbnb and Anysphere (Cursor) for Chinese AI model use; House separately probes DeepSeek's direct CCP ties and funding structure; China's NDRC instructs domestic LLMs to promote domestic AI tools; WSJ reports Chinese government investing in DeepSeek at $50B valuation [32][33][40][61][3][1][4][34][2][35]
  • 2026-05-25: DeepSeek makes permanent a 75% price reduction; technical analysis by Dr. Robert Castellano contests Huawei Ascend chip performance claims, arguing chips are a 'limited threat' to Nvidia [10][12]

Perspectives

Rohan Paul (@rohanpaul_ai)

Broadly bullish on China's AI trajectory, presenting interlocking structural advantages — patents, investment, public sentiment, and open-source quality — as evidence of a qualitative shift. His framing that Chinese open-source is now 'serious' rather than derivative has become the dominant community interpretation.

Evolution: Consistent across all items in this thread. His May 2026 posts on Chinese labs' open-source seriousness gained significant retweet amplification, suggesting wider resonance.

[19][20][21][22]

Stanford HAI / 2026 AI Index

The authoritative cross-country benchmark: US still leads China in AI performance at a 2.7% margin. China leads in publications and patents. The US spends 23 times more than China on AI. The Index frames both countries as genuine AI powers with different strengths rather than a clear winner.

Evolution: The 23x spending asymmetry figure adds a new dimension alongside the 2.7% performance gap: together they suggest the US maintains a large resource advantage that has not translated into a proportional performance lead, a framing that favors neither the US-is-winning nor China-is-winning camp.

[23][24][25][26][27][28][15][29]

Brian Chesky (Airbnb CEO)

Directly defends Silicon Valley's use of Chinese open-source AI, saying the US is 'misunderstanding' what open-source model adoption means for data security. Specifically denies that using Chinese open-source model weights involves sending user data to China.

Evolution: Consistent; first major US tech executive to publicly push back against Congressional pressure on Chinese AI model use. The expansion of the probe to Anysphere has not yet produced a comparable public defense from Anysphere leadership.

[30][31]

US House Republicans (Select Committee on China / House Homeland Security)

Frame Silicon Valley's use of Chinese AI models as a data security threat, conducting active investigations of named US companies (Airbnb, Anysphere) for using Chinese open-source models, while separately probing DeepSeek's direct CCP ties and funding structure.

Evolution: Expanded from Airbnb-only to also targeting Anysphere (Cursor) and pivoted from scrutinizing US adopters to investigating Chinese AI companies directly. The scope expansion signals this is a broadening inquiry, not a narrow one.

[32][33][3][1][4][34][2][35]

US Senate

Separately from House Republicans, US senators asked government agencies to investigate data security risks posed by DeepSeek — establishing bipartisan, bicameral pressure on the executive branch to act on Chinese AI security concerns.

Evolution: New voice in this thread; the Senate's involvement indicates the Chinese AI investigation is not confined to House Republicans or partisan framing, and has been building since at least mid-2025.

[5]

Harvard Business School (working paper)

Frames China's approach as a deliberate 'Diffusion-Forward AI Strategy' — prioritizing broad deployment, adoption, and application-layer scale over frontier benchmark leadership. Implies that US benchmark-centric comparisons systematically undercount China's strategic progress.

Evolution: New analytical voice; provides academic grounding for the view that China and the US are optimizing for different AI outcomes, not competing head-to-head on the same metrics — distinct from both the China-is-winning and US-still-leads camps.

[16]

DeepSeek founder (Liang Wenfeng)

Publicly declares an AGI goal ahead of the company's $10 billion funding round, positioning DeepSeek as a mission-driven organization pursuing artificial general intelligence rather than solely a commercial AI provider.

Evolution: New voice; the AGI declaration elevates DeepSeek's stated ambition and complicates Western assessments of whether it is a commercial company, a national project, or both.

[9]

Dr. Robert Castellano (analyst, Substack)

Technical counter-analysis arguing Huawei's latest AI chips represent a 'limited threat' to Nvidia. Contests widely-cited claims that the Ascend 910C approaches H100 performance, providing named expert-level skepticism of the bullish hardware narrative.

Evolution: New voice; the first named technical expert to push back against the Ascend-matches-H100 claims that have dominated coverage, creating a specific voice-vs-voice dispute where previously only aggregate benchmark claims existed.

[12]

AI Frontiers

Argues that China and the US are 'running different AI races' — the US optimizing for frontier capability benchmarks, China for scale, deployment, and hardware independence. Implies that single-metric comparisons systematically miss the structural divergence.

Evolution: Consistent; the Harvard HBS 'Diffusion-Forward' working paper provides independent academic validation for this framing.

[36]

Bart Collet (@bart)

Notes that five Chinese AI labs — including Alibaba's Qwen and Zhipu's GLM — have released open-source frontier models, contrasting with the pace of US lab open-source releases.

Evolution: Consistent; provides specific enumeration that deepens the open-source competition narrative beyond aggregate statistics.

[37]

aichina.news (@AiChinaNews)

Identifies a coordinated wave of Chinese open-source models natively optimized for Huawei's Ascend chips as the defining trend, framing it as infrastructure independence rather than simple model competition.

Evolution: Consistent; the bullish Ascend narrative now faces direct challenge from Castellano's technical counter-analysis, though the core open-source-for-Ascend coordination claim is not disputed.

[38]

MetaHacker (@metahacker_)

Skeptical of the China-winning narrative: claims Chinese open-source models are still 3-6 months behind Anthropic, and suggests adoption is partly subsidy-driven and fragile.

Evolution: Consistent, though the Harvard 'Diffusion-Forward' framing suggests the benchmark comparison MetaHacker relies on may be the wrong metric if China is optimizing for deployment breadth rather than benchmark rank.

[39]

China Tech Watch (@China_TechWatch)

Reports that China's NDRC instructed domestic large language models to actively promote domestic AI tools and services, framing Chinese AI governance as an active policy instrument rather than passive market competition.

Evolution: Consistent; China's May 2026 AI agent guidelines add another regulatory layer to the state-direction narrative.

[40]

The Next Web

Frames DeepSeek's $45 billion valuation as 'Beijing's strategic statement' — not merely a commercial financing event but a geopolitical signal of state commitment to DeepSeek's global role.

Evolution: Consistent; the DeepSeek founder's AGI declaration amplifies the 'national strategic asset' interpretation.

[41][25]

FuturMix.ai (@futurmix)

Points to 289 Chinese researchers mapped across top Western AI labs as key contributors, framing Chinese talent embedded in Western institutions as an underappreciated factor in China's AI advance.

Evolution: Consistent.

[42]

People's Daily / Chinese state media (@PDChina)

Frames Chinese large models as a foundation for global innovation, positioning China as a contributor to rather than merely a competitor with the global AI ecosystem.

Evolution: Consistent official framing.

[43]

Tensions

  • Stanford AI Index's 2.7% US performance lead [25] directly complicates the China-has-achieved-parity narrative promoted by Rohan Paul and others [19][20]. The new 23x US spending advantage figure [15] adds a second axis: a small performance gap achieved at a fraction of the spend could indicate China is far ahead on efficiency, or it could mean the benchmark gap will widen as resource asymmetry compounds. Neither interpretation has been conclusively tested. [25][19][20][24][15]
  • Brian Chesky's 'misunderstanding' defense of Chinese open-source AI use [30] directly clashes with House Republicans' data security framing [32][33][1], which has now expanded to include Anysphere (Cursor) as a named investigation target. Chesky argues open-source model weights are not a data pipeline to Beijing; Congressional investigators argue the risk is structural rather than transactional — and the Anysphere inclusion signals they believe the concern extends beyond consumer data to code and intellectual property exposure through developer tools. Neither side has engaged the other's specific technical claim. [30][32][33][1][2]
  • MetaHacker's claim that Chinese open-source models remain '3-6 months behind Anthropic' [39] clashes with Bart Collet's observation that five Chinese labs have released open-source frontier models [37] and with Hugging Face data showing Chinese models have overtaken US-origin models in download share [44][45]. The dispute turns on whether 'frontier' should be defined by top benchmark rank or by open-source availability and real-world adoption patterns. [39][37][44][45]
  • Dr. Robert Castellano's technical analysis arguing Huawei chips are a 'limited threat' to Nvidia [12] stands in direct conflict with community claims that the Ascend 910C already matches H100 performance [13] and TrendForce's benchmarked 60% H100 inference figure [14][46]. The disagreement turns on methodology — inference and frontier pre-training impose different architectural demands — but neither side has fully engaged the other's measurement claims. The US export controls strategy depends entirely on which assessment is correct for training-scale workloads. [12][13][14][46][38]
  • DeepSeek presents itself as an independent commercial lab pursuing AGI [9], but its CCP ties and funding structure are under direct Congressional investigation [3][4], and the conflict between its ~$10B private valuation [7] and $45–50B government-linked valuation [8] implies either two separate transactions with different principals, or significant ambiguity about the ultimate backer. The ownership analysis [6] has not resolved this publicly — and the resolution directly determines whether open-source DeepSeek model weights can be treated as commercially neutral technology by Western companies. [9][3][4][7][8][6]

Sources

  1. [1] Chairmen Moolenaar, Garbarino Announce Joint Investigation ... — reactive:china-ai-rising
  2. [2] House panels probe Airbnb, Anysphere over use of Chinese AI models - Nextgov/FCW — reactive:china-ai-rising
  3. [3] House Republicans Seek Info On DeepSeek Ties To CCP - Law360 — reactive:china-ai-rising
  4. [4] House investigation into DeepSeek teases out funding, security realities around Chinese AI tool | CyberScoop — reactive:china-ai-rising
  5. [5] Senators ask US to probe data security issues with DeepSeek — reactive:china-ai-rising
  6. [6] Who Owns DeepSeek? Uncovering Its Shareholders & Corporate ... — reactive:china-ai-rising
  7. [7] DeepSeek's Potential $10B+ Valuation, LinkedIn's AI Agent Win and ... — reactive:china-ai-rising
  8. [8] DeepSeek value could be up to $50 billion in first fundraising ... — reactive:china-ai-rising
  9. [9] DeepSeek Founder Avows AGI Goal Ahead of $10 Billion Funding — reactive:china-ai-rising
  10. [10] Chinese artificial intelligence startup DeepSeek will make ... — reactive:china-ai-rising
  11. [11] One Year Since the “DeepSeek Moment” - Hugging Face — reactive:china-ai-rising
  12. [12] Huawei’s Latest AI Chips: A Deep-Dive Technical Analysis Reveals Limited Threat to Nvidia — reactive:china-ai-rising
  13. [13] huawei's ascend 910c chip matches nvidia's h100. there will be 1.4 ... — reactive:china-ai-rising
  14. [14] DeepSeek research suggests Huawei's Ascend 910C delivers 60 ... — reactive:china-ai-rising
  15. [15] Stanford's 2026 AI index just dropped: the US spends 23x more than ... — reactive:china-ai-rising
  16. [16] [PDF] China's Diffusion-Forward AI Strategy - Harvard Business School — reactive:china-ai-rising
  17. [17] China unveils guidelines to regulate, boost innovative development ... — reactive:china-ai-rising
  18. [18] On 8 May 2026, three of China's most powerful regulatory bodies ... — reactive:china-ai-rising
  19. [19] 🇨🇳 China is filing and winning far more patent claims in AI. — Rohan Paul Twitter (2026-05-17)
  20. [20] 🇨🇳 China’s public is unusually positive about AI products compared to other countries, which lowers adoption friction an… — Rohan Paul Twitter (2026-05-17)
  21. [21] 3/n The release includes the SenseNova U1 Lite series: — Rohan Paul Twitter (2026-05-20)
  22. [22] RT @rohanpaul_ai: Chinese AI labs are increasingly releasing very serious open source work. — Rohan Paul Twitter (2026-05-21)
  23. [23] China leads in AI publications, patents: Stanford report - CGTN — reactive:china-ai-rising
  24. [24] @WatcherGuru 👁️ The timing is wild 👀 Stanford’s 2026 AI Index just confirmed the US lead over China in AI performance ha... — reactive:china-ai-rising (2026-05-19)
  25. [25] Stanford AI Index 2026: China narrows US lead to 2.7% while ... - TNW — reactive:china-ai-rising
  26. [26] The 2026 AI Index Report | Stanford HAI — reactive:deepmind-ai-co-clinician
  27. [27] Technical Performance | The 2026 AI Index Report | Stanford HAI — reactive:frontier-ai-cyber-capabilities
  28. [28] Inside the AI Index: 12 Takeaways from the 2026 Report — reactive:frontier-ai-cyber-capabilities
  29. [29] 20 Takeaways from Stanford’s 2026 AI Index Report — reactive:china-ai-rising
  30. [30] Airbnb's Chesky Says US 'Misunderstanding' Use of Chinese Open-Source AI Models — reactive:china-ai-rising (2026-05-23)
  31. [31] Airbnb's Chesky Says US 'Misunderstanding' Use of Chinese Open-Source AI Models — reactive:china-ai-rising (2026-05-23)
  32. [32] House Republicans probe Airbnb's use of Chinese AI models citing data security risks. CEO Brian Chesky says the company ... — reactive:china-ai-rising (2026-05-23)
  33. [33] As the US House probes Airbnb's use of Chinese AI models, CEO Brian Chesky says the company is not sharing data with Chi... — reactive:china-ai-rising (2026-05-23)
  34. [34] House GOP launches inquiry into Airbnb use of Chinese AI models — reactive:china-ai-rising
  35. [35] House investigating Airbnb over Chinese AI models - The Hill — reactive:china-ai-rising
  36. [36] China and the US Are Running Different AI Races | AI Frontiers — reactive:china-ai-rising
  37. [37] Five Chinese AI labs, including Alibaba's Qwen and Zhipu's GLM, have now released open source frontier models. Meanwhile... — reactive:china-ai-rising (2026-05-20)
  38. [38] The defining trend of this window is a massive, coordinated wave of open-source models natively optimized for Huawei's A... — reactive:china-ai-rising (2026-05-22)
  39. [39] @OrganicGPT Chinese open source models are 3-6 months behind anthropic. If AI subsidies go away, people will just switch... — reactive:china-ai-rising (2026-05-20)
  40. [40] China’s National Development and Reform Commission (NDRC) has reportedly instructed domestic large language models to ac... — reactive:china-ai-rising (2026-05-23)
  41. [41] DeepSeek's $45bn valuation is also Beijing's strategic statement — reactive:china-ai-rising
  42. [42] @bindureddy This is backed by data. We mapped 289 Chinese researchers across top Western AI labs — many are the key cont... — reactive:china-ai-rising (2026-05-16)
  43. [43] Positioning Chinese large models as a foundation for global innovation — reactive:china-ai-rising (2026-05-19)
  44. [44] One year after DeepSeek, Chinese AI models spread rapidly, overtake U.S. in download share — reactive:china-ai-rising
  45. [45] China Dominates Top 10 Open-Source Models on ... — reactive:china-ai-rising
  46. [46] [News] DeepSeek Reportedly Reveals Huawei’s Ascend 910C Reaches 60% of NVIDIA H100’s Inference Power — reactive:china-ai-rising
  47. [47] [News] Decoding Huawei’s DeepSeek All-in-One Machine: 60-70% of NVIDIA H100 Performance at an Appealing Price — reactive:china-ai-rising
  48. [48] China Raises the Bar for AI Patents: What Changes from 1 January 2026 - Mathys & Squire LLP — reactive:china-ai-rising
  49. [49] China becomes world's largest holder of AI patents - People's Daily — reactive:china-ai-rising
  50. [50] [PDF] AI Patents in the United States and China: Measurement ... - CDN — reactive:china-ai-rising
  51. [51] China's DeepSeek is raising funds at $10 billion valuation ... - Reuters — reactive:china-ai-rising
  52. [52] China's DeepSeek is Raising Money for First Time, At $10 Billion ... — reactive:china-ai-rising
  53. [53] DeepSeek seeks $300M in first outside funding at $10B valuation — reactive:china-ai-rising
  54. [54] An open-source model, 'SenseNova U1,' capable of image generation without the need for VAEs, has been released, offering significantly faster speeds and better quality than Z-Image. - GIGAZINE — reactive:china-ai-rising
  55. [55] SenseTime's SenseNova U1 ditches VAEs entirely to unify image generation and understanding - Startup Fortune — reactive:china-ai-rising
  56. [56] SenseTime Fully Open-Sources SenseNova U1: A Unified Model for Understanding and Generation-News and Blog-SenseTime — reactive:china-ai-rising
  57. [57] SenseNova-U1: Unifying Multimodal Understanding and Generation ... — reactive:china-ai-rising
  58. [58] WE ARE CHINA — reactive:china-ai-rising
  59. [59] China's open-source AI strategy is quietly reshaping global tech infrastructure. — reactive:china-ai-rising (2026-05-21)
  60. [60] State of Open Source on Hugging Face: Spring 2026 — reactive:china-ai-rising
  61. [61] China to Invest in DeepSeek at $50 Billion Valuation - WSJ — reactive:china-ai-rising