Back to News

Silicon Valley Researchers Admit China's Edge in Multimodal AI, Says Partner

#ai#silicon-valley#china#multimodal

In an interview, Wu Yuan Capital partner Meng Xing described shifting attitudes in Silicon Valley, where OpenAI and Google researchers now acknowledge Chinese model companies' strength in areas like multimodal AI, with one Google researcher saying 'can't beat domestic [Chinese]'. He also noted that the gap between US and China varies by domain—pretraining, post-training, and multimodal—and questioned whether top Neo labs can replicate OpenAI's success, citing Thinking Machines Lab's struggles and Anthropic's ARR of $65 billion in August, below some investor expectations.

Coverage timeline

  1. 晚点 LatePost
    当硅谷开始说 “打不过国内”丨五源孟醒

    8 月的一个下午，我从旧金山机场出来，上 101 高速往南开。 湾区的夏天几乎没有云，阳光把路面晒得发白。机场附近还是仓库和低矮的办公楼，再往南，风景慢慢被广告牌接管。 这条路的广告牌，一直是硅谷最诚实的风向标。两年前挤满 SaaS，去年还能看到一点消费品牌，今年清一色全是 AI 公司。 图一：Kimi 也加入了旧金山 “AI 广告牌” 大战；图二：Codex 的广告牌在旧金山越来越多，花了血本与 Claude Code 竞争 没有任何迹象表明 AI 慢下来了，钱、卡、人还在往这里涌。真正超出我预料的是，一些过去很难听到的判断，开始出现在最核心的人群里。 几位 OpenAI 的朋友，花了很长时间复盘过去一年押错的方向，也认真拆解中国模型公司到底强在哪；Google 的 researcher 聊到多模态，很直接地说：“打不过国内。” 这当然不能概括今天中美 AI 的整体位置，但几年前，很难想象这样的判断会从硅谷一线的 researcher 嘴里说出来。 大家开始重新估计中美之间的差距到底有多大，不过这个问题也已经没有一个统一判断：在预训练、后训练和多模态上，答案可能完全不同。 另一种变化，发生在过去两年最贵的那批 Neo lab 身上。 Thinking Machines Lab 的 researcher 告诉我，他们今天经历的痛苦，其他 Neo lab 迟早还会再经历一遍。而我越来越怀疑，下一家真正跑出来的 Neo lab，不会长成今天这个样子。 这可能是我这次在硅谷最强烈的感受：过去两年形成的很多边界和共识，并没有想象中那么牢固： 最顶尖的 researcher 聚在一起，是不是就能再造一家 OpenAI？美国最好的 Lab，是不是在每一个战场上都能领先？模型、数据、应用和算力，到底谁是谁的上游？最强的模型，是不是就应该一直卖最贵的价格？…… 甚至连最漂亮的增长故事，也被重新计算：8 月，彭博报道 Anthropic 的 ARR 达到了 650 亿美元，比两个月前又涨了 180 亿。放在任何一家公司身上，这都是惊人的增长。但它仍然低于市场此前的期待，一些投资者原本预期，这个数字会达到 700 多亿美元。 五个月前我写 “全员 token-maxxing”，那时候硅谷最强烈的情绪是怕跟不上，更多 token、更多卡、更大的模型、更快的迭代，速度本身就是答案。 但五个

Related stories

  • Z.ai confirms Ox Alpha is GLM-5.3 Flash, releases open weights

    Z.ai confirmed that the anonymous model 'Ox Alpha', which went viral on OpenRouter, is a new iteration of its GLM series, specifically GLM-5.3 Flash, and released its weights on Hugging Face under a new license. The model, a 320B-parameter natively multimodal model running on domestic chips, topped OpenRouter's leaderboard and ended DeepSeek's 56-day streak on OpenCode. The license requires companies with over $10B in revenue to pass a security review.

    9 reports · 4 sources801 points · 281 comments

  • Anthropic CEO proposes 'pacing the frontier' plan; OpenAI, Musk, Hugging Face respond

    Anthropic CEO Dario Amodei published an essay on September 12, 2026, urging the AI industry to slow the pace of frontier model development to address safety risks, proposing a three-part plan: embedded third-party evaluators, coordination among democratic countries, and global coordination including with authoritarian governments. Anthropic unilaterally committed to giving third-party evaluators permanent employee-level access. OpenAI CEO Sam Altman agreed and said OpenAI would do the same, while Elon Musk endorsed the call and Hugging Face launched an Open Alignment Initiative seeking to join the embedded evaluators program.

    25 reports · 6 sources330 points · 260 comments

  • Musk proposes rival-run safety tests for top AI labs in US and China

    Elon Musk called for top US AI labs and three or four leading Chinese companies to let rivals run a 'test harness' on their models to evaluate safety. He urged the companies to work together and test each other's models, according to a CNBC report.

    1 reports · 1 sources

  • OpenAI claims internal model solved Navier-Stokes Millennium Prize problem in 88 hours

    OpenAI announced that an unreleased next-generation model, described as significantly more capable than GPT-6 Astra, solved the Navier-Stokes existence and smoothness problem, one of the Clay Mathematics Institute's Millennium Prize Problems, using about 10,000 concurrent AI agents over 88 hours. The company released a 166-page proof and Lean 4 formal verification code, and said it would forgo the $1 million prize. The announcement has sparked controversy over alleged scooping and data use, with mathematician Tristan Buckmaster disputing OpenAI's account of negotiations.

    7 reports · 5 sources178 points · 179 comments

News updates every hour. Subscribe in your RSS reader to get new stories automatically.

Subscribe