Amazon and Alibaba: Two Giants with a $220 Billion Obsession

The race is not to the swift, nor the battle to the strong. August 3rd will be a date to remember in the global race for artificial intelligence. Amazon surpassed a market capitalization of $3 trillion, becoming the fifth company to reach this milestone, driven by AWS growth that caught everyone off guard. On the same day, thousands of miles away, Alibaba unveiled Qwen3.8-Max, its most ambitious AI model to date. Two companies, two continents, one common obsession: to never lose pace in this race that devours hundreds of billions. Key points of this article:

• On August 3rd, Amazon reached a historic market capitalization of $3 trillion, fueled by impressive growth in its AWS division.
• Alibaba revealed its most ambitious AI model, Qwen3.8-Max, with an unprecedented yet controversial opening, causing tensions with Washington.

AWS, the engine running faster than expected
The figures released by Amazon far exceed Wall Street's expectations. According to CNBC, AWS grew by 37% year-on-year, reaching $42.23 billion in revenue, with an operating margin of 39.4%, its best performance in eighteen quarters. CEO Andy Jassy clarified that AI and chip-related activities each exceeded $25 billion in annualized revenue. This sustained pace has led to an increase in the 2026 investment budget from $200 billion to $220 billion, primarily due to the surge in chip costs.

Amazon is no longer hiding the scale of its bet on generative AI. The company recently invested $50 billion in OpenAI, while signing an AWS capacity lease contract worth about $100 billion over eight years with the same company. This mechanism closely mirrors the strategy already perfected by Microsoft with Azure. The more OpenAI grows, the more guaranteed revenue AWS generates, but Amazon also ties its financial fate to the health of a company that is burning cash at an unprecedented rate in tech history.

Qwen3.8-Max, the other bet: betting on openness
While Amazon consolidates its cloud fortress, Alibaba plays a different card. Qwen3.8-Max features 2.4 trillion parameters, with only 95 billion activated per request thanks to a mixture of experts architecture, and manages a context of up to one million tokens.

The model ranks fifth in the Text Arena ranking and fourth in the Frontend Code Arena, just behind Claude Opus 5 and Kimi K3, with autonomous coding capabilities pushed to execute a complete software engineering project over sixteen days without human intervention. Notably, Alibaba promises to release the weights as open source, a first for a model of this category at Qwen.

However, Alibaba's openness deserves some nuance. According to Tom's Hardware, accounts affiliated with the group have been accused of generating nearly 29 million fraudulent exchanges with Claude's servers between April and June to extract expertise and feed Qwen's training, a practice that prompted Washington to react. Open source, appealing on paper, often says little about how a model was trained in advance.

In an ironic twist of fate, the same Amazon had briefly lost its capitalization duel to Bitcoin in May 2025, with the queen of cryptocurrencies surpassing Amazon during a "Bitcoin Pizza Day" at a $2.2 trillion market cap. A little over a year later, Amazon has regained a clear lead and is even changing categories. In the race for the largest global capitalizations, no position is ever secured for long.

This content is provided for general informational purposes only and doesn't constitute financial, investment, legal, or tax advice. Any events, rewards, online promotions, or related information mentioned herein should not be considered a recommendation, solicitation, or invitation to purchase, sell, trade, or otherwise deal in any crypto assets. Crypto assets are highly volatile and may result in loss. The availability of WEEX services, products, and related events may vary by region. You are responsible for ensuring that your participation is in accordance with applicable local laws and regulations.