Xiaomi: An Underestimated AI Narrative in China
Xiaomi's MiMo-V2-Pro topped for three weeks YET Xiaomi's stock hit a 12-month low. Introducing The AI Script series.
Since we started working on this Xiaomi AI story, Xiaomi’s MiMo model lead Luo Fuli posted a detailed take on Anthropic’s decision to cut off third-party harnesses from using Claude subscriptions. She also described MiMO-V2-Pro as a “quiet ambush.” She signed off the tweet from Beijing. Interesting choice.
In a new Calling the Shots series called The AI Script, I want to track how AI narratives shape the perception of power. From valuation gaps to geopolitical leverage, the story told about AI often matters more than the technology itself. Xiaomi’s venture into AI is the first story.
Xiaomi: An Underestimated AI Narrative in China
By Peiyue Wu
Xiaomi is undeniably China’s closest parallel to Tesla. Both began as consumer tech companies that successfully extended into cars, and now both are pushing into AI.
For three consecutive weeks, Xiaomi’s MiMo-V2-Pro AI model dominated the OpenRouter leaderboards. It was a rare feat of endurance that even high-flying startups like Z.ai and MiniMax didn’t sustain. Yet while the market quickly rewards LLM breakthroughs from startups, it often overlooks Xiaomi’s AI progress. On April 2, the very day MiMo’s API traffic peaked, Xiaomi’s stock price tumbled to its lowest point since May 2025, halving its peak valuation.
The shadow of the SU7
Xiaomi’s stock struggles are tied to a lingering brand crisis stemming from two high-profile accidents in 2025. The company’s rise was deeply intertwined with Lei Jun, Xiaomi’s founder, and how his personal charisma converted into commercial momentum. But that kind of overwhelming attention cuts both ways. Under such intense scrutiny, such accidents can be magnified into a full-blown crisis of trust.

In that sense, Xiaomi is similar to Tesla, in that the company’s image is closely tied to its founder’s public persona. However, the founder-led communication playbook that Lei Jun and Elon Musk operate by underpins a very different logic. Musk thrives in the fray, leaning into rumor mills and opining on anything that interests him to drive traffic and cultural relevance. When a Tesla crashes, he uses his savant status to deflect blame, finding alternative explanations why Autopilot was not fully engaged, and the accidents were all drivers’ fault. His persona as a defiant disruptor fighting for the future of self-driving acts as a defensive shield.
Lei built his authority on a different archetype. He emerged from China’s first-generation of internet billionaires (think Jack Ma of Alibaba, Pony Ma of Tencent, Richard Liu of JD), watching peers achieve historic success while he spent nearly two decades searching for his moment. When Xiaomi arrived, it carried the weight of that waiting — quiet suffering, stubborn persistence, an older generation’s work ethic. That image of the persevering striver resonated deeply, and for years, his personal credibility built on that persona was Xiaomi’s most valuable asset. Therefore, when the SU7 accidents hit, Lei had to adopt a more conciliatory tone.
It’s also more than the persona they chose. In China, a high-profile consumer tech company operates with far less room to deflect or reinterpret responsibility. A direct response, often including an apology, is not optional, but a baseline requirement for maintaining public legitimacy.
This ultimately reflects a deeper cultural divide. In Silicon Valley, founders are encouraged to project conviction about a future that does not yet exist, and investors reward that narrative. In China, public expectations are shaped more by a culture rooted in thousands of years of agrarian pragmatism. A sweeping “AI+car+home” vision carries less weight in shaping opinion than a concrete incident that exposes current technical limits.
The paradox of the rising underdog
Now, let’s return to why Xiaomi’s venture into AI has not been rewarded in market value. The Xiaomi versus Minimax and Z.ai story is, at its core, a story of the Big Co. versus Challenger.
The market views Z.ai and MiniMax as pure-play AI bets, and sees their narratives as those of underdogs in an area dominated by Chinese big tech. When their models succeed, it validates their entire existence, driving valuations north of HK$300 billion. It’s worth noting that both companies are backed by Tencent and Alibaba.
Xiaomi gets no such treatment. The market sees it primarily as a hardware conglomerate, with AI as the icing on the cake. For Big Co. like Xiaomi, Alibaba, or Tencent, AI spending is viewed through the lens of capital expenditure, a defensive cost necessary to keep phones, cars, and ecosystems competitive, rather than a new revenue driver.
But this is not only a question of business model but also about how these companies tell their stories.
When Alibaba releases a model, the announcement often reads like a KPI report dressed up as a press conference. Take the latest Qwen 3.6-Plus: it highlights refreshed SOTA results across 215 tasks and claims to outperform Gemini-3.1 Pro on multiple core metrics. The language is designed more for managing up and investor relations, but it doesn’t necessarily help win users’ hearts. By contrast, MiniMax builds its reputation through real-world testing and organic word of mouth from its community. Z.ai adopts the register of a top-tier academic lab. In both cases, the communication feels closer to peer-to-peer sharing than to a grand narrative such as a “cloud + AI” closed-loop ecosystem.
Xiaomi is caught in an awkward position between the two modes.
It is no longer the underdog people want to root for. In his 2025 annual speech, Lei returned to the hardship narrative that once moved his audience, comparing building cars and chips simultaneously to a family putting two children through university at once. The metaphor was meant to humanize the challenge. Instead, people read as performative. The kind of relatable-struggle framing that works for a scrappy startup founder, but rings hollow coming from one of China’s wealthiest men.
Meanwhile, Xiaomi is also too traditional to be seen as a worthy AI player in the arena. A well-known private equity investor wrote on Weibo on April 1 that Xiaomi has entered many industries, but rarely ranks among the top three in any of them. “Now that its auto growth is slowing, it is turning again to AI and robotics.” He went further, saying it has no edge in AI and questioning whether it has a long-term strategy at all.
Xiaomi’s hesitation is visible in its media and marketing approaches. Over the past year, both Lei and Xiaomi have noticeably dialed things down. When MiMo first appeared on OpenRouter, it was listed under the alias “Hunter Alpha” with no indication of its origin, and Xiaomi only stepped forward to claim it six days later, after it had already held the top spot for three consecutive days. Earlier, during its Leica-branded phone launch in Europe, Xiaomi even removed its own logo to let Leica take center stage. Both moves aim to avoid drawing attention to the breadth of Xiaomi’s business, as too many bets could dilute efforts and increase capital expenditure.
From a purely technical standpoint, Xiaomi’s MiMo-V2-Pro holds up against models from MiniMax and Z.ai. Developer community discussions suggest that all three are well-regarded, each with its own strengths. Yet the market tends to underestimate Xiaomi.
Xiaomi has potential in AI hardware integration and in building an ecosystem across phones, vehicles, and home appliances. Few companies can actually deploy AI at that scale in real-world scenarios. Therefore, if it can change the market’s perception of its narrative, there is significant enterprise value to be accrued once it proves itself a worthy challenger.






