Continue Reading With Our 7-Day Free Trial
Alibaba Group (NYSE:BABA) has delivered a fresh jolt to the global artificial intelligence race. The Alibaba Qwen3.8 Max AI Model is now at the center of that global competition. The company previewed Qwen3.8-Max-Preview, a new flagship model with 2.4 trillion parameters. Alibaba says the model ranks second only to Anthropic’s Fable 5.
That is a bold claim from a company once known mainly for online shopping. Alibaba also plans to make the model open-weight. Developers will then be able to download, modify, and build products around it.
Investors took notice. Alibaba’s Hong Kong-listed shares rose as much as 5.6% on Monday, outperforming the broader technology market.
Still, this story is about more than a benchmark ranking. Alibaba already has cloud infrastructure, business customers, consumer apps, and its own chip designs. That gives Qwen a commercial foundation that many AI startups lack.
The bigger question is simple. Is Alibaba merely catching up, or is China building a genuine rival to Silicon Valley’s AI leaders?
The Alibaba Qwen3.8 Max AI Model Is Becoming More Than A Benchmark Showpiece
A large model can generate headlines, but commercial usage provides stronger evidence. Alibaba is already turning Qwen into a growing cloud business.
Alibaba Cloud’s external revenue increased 40% year over year in the latest reported quarter. AI-related product revenue posted triple-digit growth for the eleventh straight quarter. Its annualized run rate reached roughly RMB36 billion, or about $5.3 billion.
AI products now represent around 30% of Alibaba Cloud’s external revenue. Management expects that share to cross 50% within about one year.
Most revenue on Alibaba’s Bailian Model Studio currently comes from Alibaba’s own models. That includes Qwen, along with voice, video, and image-generation tools. Usage on Bailian also increased more than tenfold between late 2025 and May 2026.
These figures suggest that Qwen is already supporting real workloads. Customers are using it for coding, reasoning, business applications, and automated tasks. The Alibaba Qwen3.8 Max AI Model could extend that commercial momentum if customers adopt it at scale.
However, the ranking claim still comes from Alibaba. Commercial adoption supports Qwen’s relevance, but it does not independently prove global model superiority. Broader testing will remain important.
China’s AI Race Is Moving Faster & Becoming More Commercial
Alibaba is not developing Qwen in isolation. China’s AI market is producing larger and more capable models at a rapid pace.
Moonshot recently introduced its Kimi K3 model with 2.8 trillion parameters. Z.AI has also released its GLM-5.2 model. These launches show how quickly leadership can change within China’s domestic market.
Alibaba believes China has already reached an important stage in AI coding. Management said coding improvements drove much of Bailian’s recent usage growth. These tools now handle tasks that stretch beyond writing basic software.
That matters because coding agents have become a major growth area in the United States. They can review files, identify errors, and complete routine digital work. Similar adoption in China could create another large market for cloud computing.
Alibaba also treats many Chinese AI startups as partners. Bailian offers access to proprietary models and third-party models. Smaller developers can specialize, while Alibaba supplies infrastructure and distribution. The Alibaba Qwen3.8 Max AI Model now gives that broader ecosystem another flagship product around which developers can build.
This creates a different kind of contest. China does not need one company to defeat every American rival. A network of model developers, cloud platforms, and application companies could collectively narrow the gap.
Still, fast releases do not guarantee lasting leadership. Models can lose their advantage within months.
Open-Weight Qwen Could Become Alibaba’s Global Distribution Weapon
Alibaba’s decision to make Qwen3.8-Max open-weight may prove more important than its parameter count. Open-weight access lets developers download and modify the model for specific needs.
That could encourage startups, researchers, and businesses to test Qwen without depending entirely on Alibaba. It may also help the model spread beyond China. The Alibaba Qwen3.8 Max AI Model could therefore reach developers who might never use Alibaba’s consumer platforms directly.
Giving away model access does not eliminate the business opportunity. Many developers still need computing power, storage, security tools, and technical support. Alibaba can provide those services through its cloud platform.
Its model and application services already had annualized recurring revenue above RMB8 billion in May. Management expected that figure to cross RMB10 billion during the June quarter. It also targeted RMB30 billion by year-end.
Alibaba has another advantage through its consumer ecosystem. The Qwen app connects with Taobao, Tmall, Alipay, Amap, and Fliggy. That could let users shop, pay, navigate, and plan travel through one assistant.
Few standalone AI startups have access to that combination of cloud infrastructure and daily consumer activity.
However, open access has limitations. Developers may run Qwen without paying Alibaba. Overseas customers may also raise questions about security, data handling, and regulation.
Open-weight availability can widen adoption, but it cannot guarantee commercial loyalty.
Catching Silicon Valley Will Require Enormous Spending
Alibaba’s AI progress comes with a large financial cost. The company is investing heavily in data centers, chips, models, and consumer applications.
Adjusted group EBITA fell 84% during the latest quarter. Free cash flow was an outflow of RMB17.3 billion. Management linked much of that pressure to AI infrastructure spending.
Alibaba has the resources to continue. It ended March 2026 with about $38 billion in net cash. However, management expects elevated investment to continue for at least two more years.
The company previously outlined an RMB380 billion investment program. It now believes spending could exceed that figure. Alibaba estimates it may need ten times more data-center infrastructure by 2033 than it had in 2022.
Hardware costs are also rising. Management said a new server costs roughly twice as much as one year earlier.
Alibaba’s T-Head chips may reduce dependence on outside suppliers. Yet deployment remains limited by domestic production capacity. Management also acknowledged that Chinese chips still trail leading foreign products in efficiency.
Then there is the model itself. More parameters do not automatically mean better answers or lower costs. Speed, reliability, safety, and accuracy matter just as much. The Alibaba Qwen3.8 Max AI Model must perform well across those areas to justify the infrastructure behind it.
Alibaba may be closing the technical gap, but maintaining that position will require sustained spending.
Final Thoughts
Alibaba’s latest model sends a clear message. China’s AI companies are no longer competing only through cheaper or smaller alternatives. They are building frontier models, commercial platforms, and large computing networks.
Qwen3.8-Max could strengthen Alibaba Cloud and attract more developers. Its open-weight strategy may also increase international usage. Still, Alibaba must prove that model demand can produce durable profits and cash flow.
The valuation presents a mixed picture. As of July 17, 2026, Alibaba traded at 1.80x LTM enterprise value to revenue and 1.77x LTM price to sales. It also traded at 4.51x enterprise value to gross profit.
Those sales-based multiples remain moderate for a company reporting rapid AI revenue growth. However, earnings-based measures reflect heavier investment. Alibaba traded at 16.71x LTM EV/EBITDA, 29.20x EV/EBIT, and 18.02x diluted earnings.
The gap between its sales and operating-profit multiples tells the story. Alibaba’s valuation reflects AI potential, but also the cost of building that future.
For now, Qwen3.8-Max is a serious warning shot. Whether it changes global leadership will depend on independent performance, developer adoption, margins, and returns on infrastructure spending.
Disclaimer: We do not hold any positions in the above stock(s). Read our full disclaimer here.
