In the past 24 hours (August 20 to August 21, 2026), the global AI industry has continued to focus on infrastructure, enterprise-level agents, and embodied intelligence. Alibaba Cloud's AI revenue continues to grow rapidly, overseas multi-model platforms and AI computing power financing have seen large transactions, and China's robotics and AI industrialization have also made new progress.
1. Alibaba Cloud's AI and computing revenue up 45% year-on-year
Alibaba's latest quarterly revenue grew 9% year-on-year, with AI cloud and computing services revenue reaching 48.44 billion yuan, up 45% year-on-year. The company has invested about half of the 380 billion yuan in its three-year AI plan and expects related capital expenditures to break even within three years.
2. Alibaba's AI investment suppresses short-term profits
Alibaba's quarterly net profit fell about 75% year-on-year, while capital expenditure increased to 67.68 billion yuan. Data shows that competition in large models is shifting from model parameters to long-term capital investment in data centers, chips, and cloud services.
3. Zhongke Fifth Generation, a Chinese embodied intelligence enterprise, has continued financing
Hangzhou's embodied intelligence company Zhongke Fifth Era announced the completion of A1 and A2 financing rounds, with participation from Bank of China and other institutions. Funds will be used for embodied brain research and product commercialization, with intelligent robot control remaining a key focus for capital.
4. The 2026 World Robot Conference focuses on "Physical AI"
The Beijing World Robot Conference showcased humanoid robots, industrial robots, and intelligent control solutions. Industry focus is shifting from demonstration capabilities to autonomous decision-making, operational capabilities, and large-scale deployment in real-world scenarios.
5. Stripe's acquisition of OpenRouter increases value in the multi-model era
Stripe has confirmed acquisition of AI model aggregation platform OpenRouter, with external reports valuing the deal at over $7 billion. OpenRouter covers hundreds of models, showing that enterprises are shifting from "picking one model" to dynamically selecting different models and controlling token costs.
6. OpenAI is catching up with Anthropic again in the enterprise user market
The latest enterprise AI usage data indicates that OpenAI is regaining some market share from enterprise customers. Enterprise loyalty to model vendors is declining, and performance, price, coding capabilities, and agent capabilities are becoming core factors in migration decisions.
7. Ramp launched the AI model routing service Router
Fintech company Ramp has launched a router that can automatically select multiple large models based on tasks. Model routing is becoming the new AI infrastructure layer, expected to help enterprises reduce inference costs and reduce reliance on single models.
8. Broadcom plans over $60 billion in AI chip financing
The report states that Broadcom is discussing over $60 billion in AI chip financing with lenders and may serve AI companies like Anthropic. As the scale of training and inference continues to expand, AI competition is simultaneously evolving into a competition for chips, financing, and energy infrastructure.
Frequently Asked Questions (Q&A)
Q: What is the most obvious AI trend in the past 24 hours?
A: Companies are moving from simply pursuing stronger models to competing simultaneously in computing power, model routing, agents, and commercial efficiency.
Q: Why has OpenRouter attracted attention from the capital market?
A: Because companies increasingly need to call multiple models simultaneously and automatically select the most suitable model based on cost, speed, and capability.
Q: What is the most noteworthy aspect of China's AI industry right now?
A: Growth in cloud computing revenue, investment in domestic computing power, and commercialization of embodied intelligence are currently the three main industry themes.
Q: What is the biggest risk of massive AI capital expenditures?
A: If model call volume and enterprise payment growth fall short of expectations, investments in data centers, chips, and energy may continue to drag down profits and cash flow.