Alibaba spent heavily across nearly every layer of AI, models, cloud, chips, while Tencent kept its bet narrower, focused on social media and content. Investors just decided which strategy they trust more.
Alibaba shares surged roughly 35% in Hong Kong this quarter, ahead of earnings due Thursday, putting the company on track for its biggest quarterly outperformance against Tencent since early 2025.
Two Different AI Strategies
The split in approach is stark. Tencent has largely folded AI into its existing social and content businesses. Alibaba built out generative models, cloud infrastructure and its own chip designs simultaneously, and cloud growth has started accelerating as a result.
That breadth is starting to pull attention away from smaller model-focused rivals too. Alibaba is reclaiming spotlight that had shifted earlier this year toward upstart labs like Z.AI.
Alibaba wasn't always the obvious pick. The company was an early winner when China's AI stock boom first took off, then fell behind as competitors captured headlines through new listings and technical breakthroughs of their own. Its resurgence has also given the company a new story to tell investors, reframing itself from an online retail giant wrestling with sluggish domestic consumption into a genuine technology platform play.
"Alibaba's AI investments have been effective in reviving both investor interest in the stock and user engagement across its broader ecosystem," said Gary Tan, a portfolio manager at Allspring Global Investments.
Why Investors Are Rewarding the Bet
Alibaba is expected to report 8.4% revenue growth for the June quarter, the fastest pace in nearly three years, even as profit is projected to decline under the weight of continued spending. Tencent and Baidu, by contrast, both saw their stocks fall after recent earnings disappointed the market.
JPMorgan analyst Alex Yao expects Alibaba's results to come in "better than feared," pointing to narrower losses in food delivery and quick commerce alongside accelerating cloud revenue and improving margins.
The market has taken notice in a way that hasn't happened in years. Alibaba shares are now trading at a valuation premium to Tencent's for the first time in more than a decade.
The Model Is Becoming a Commodity
Part of what's driving the shift is happening outside Alibaba entirely. Rapid model releases from DeepSeek and Moonshot are pushing the industry toward what analysts call a model-agnostic landscape, where businesses pick whichever system is cheapest or best suited to a task rather than committing to one vendor.
When the model itself becomes replaceable, the advantage moves to whoever controls the infrastructure underneath it. Alibaba Cloud held 37% market share in China as of the fourth quarter of 2025, according to Omdia, well ahead of Huawei's 17% and Tencent's 10%.
That lead is reinforced by how much of the stack Alibaba controls directly. Beyond cloud infrastructure, the company designs some of its own chips and offers everything from the consumer-facing Qwen app to coding tools and enterprise agents built for developers, giving it multiple points of contact with customers regardless of which specific model ends up winning any given benchmark.
"Companies with full-stack capabilities, from chips and cloud infrastructure to models and applications, like Alibaba, are better positioned to lead," said Citigroup analyst Alicia Yap.
What This Means for Miami
Alibaba's Qwen models are part of the same wave of cheap, open-weight Chinese AI already showing up in cost-conscious enterprise budgets, including here in South Florida, where businesses are increasingly routing workloads to whichever model is good enough at the lowest price.
Alibaba's stock rally is a signal for Miami's investors and enterprise AI buyers alike. The company making money isn't necessarily the one with the flashiest model release. It's the one that owns the full stack underneath it, the same full-stack logic showing up in deals like Google's recent custom chip agreement with Marvell. As AI competition increasingly plays out at the infrastructure layer rather than the model layer, that's where the durable business value, and the durable investment thesis, is starting to concentrate.
