Physix FrontierZTE's Dual-Track AI Phone Strategy: High-Stakes Gamble or Necessary Evolution?
Recently, I've been getting headaches from reading papers in the lab, and coincidentally, I came across a news item: ZTE-affiliated brands are going to "deeply focus on AI phones," while regular product lines will tentatively continue normal iterations. At first glance, this sounds like a researcher saying, "I'm going to go all-in on this cutting-edge direction, but I can't stop the basic experiments on hand either"—it sounds reasonable, but in practice, energy allocation is always the biggest challenge.
This strategy makes me think of an analogy: It's like running two tracks simultaneously—one is a sprint (AI phones, requiring quick results), and the other is a marathon (regular iterations, relying on long-term accumulation). The problem is, sprints require explosive power, marathons require endurance, and the training methods for these two are completely different. Can ZTE-affiliated brands handle both? Or does this "dual-track system" actually expose a deeper dilemma within the industry?
Comparing Two Routes: Aggressive vs. Steady
Let's look at the aggressive faction first. Typical representatives include Samsung's Galaxy AI, Google's Pixel series, and Apple's rumored "AI iPhone." They choose to deeply embed AI capabilities into the system's underlying layer, reconstructing everything from chip design (like Google Tensor) to system optimization (like Samsung One UI 6.1) around AI. This approach carries extremely high risk—if the judgment on the technology route is wrong (for example, Qualcomm has already proven that power consumption remains a bottleneck for AI accelerators in mobile devices), the entire product line could fall into a passive position. But the returns are also huge: seizing user mindshare early and establishing ecological barriers.
Now let's look at the steady faction. For example, Xiaomi, OPPO, and vivo's approach to regular iterations: maintaining robust hardware upgrades while achieving AI functions through third-party collaborations (such as integrating ERNIE Bot or Qwen). This is like a researcher setting up the experimental bench (basic hardware) first, then gradually introducing new tools (AI models). The advantage is flexibility and dispersed risk; the disadvantage is that it's hard to create differentiation and easily viewed as a "follower."
ZTE-affiliated brands' strategy seems to lie somewhere in between: On one hand, saying "deeply focus on AI phones" implies resource tilting, organizational structure adjustments, and possibly cutting non-core product lines; on the other hand, emphasizing "normal iteration of regular product lines" indicates they don't want to give up established market share and supply chain relationships. This "want it all" mindset reminds me of what supervisors often say in the lab: "You can't pursue three directions simultaneously and go deep in all of them."
Nubia and RedMagic: Two Test Fields for AI Phones
The news specifically mentioned the Nubia Z90 series and the RedMagic 12 series. These two series happen to represent the two extremes of ZTE-affiliated brands: Nubia focuses on imaging and full screens (custom 35mm lens, 1.5K large full screen), which is a typical "regular iteration" direction; RedMagic focuses on gaming and battery life (massive ten-thousand mAh battery), which is a typical "performance-oriented" direction.
So how do AI phones land in these two directions? Let me try to raise a few questions:
Imaging Direction: Nubia's 35mm lens is itself a differentiated selling point, but what can AI do? Over the past few years, computational photography has pushed phone imaging into the "algorithm competition" stage. If it's just adding AI portrait beauty filters or scene recognition, there's no essential difference from other manufacturers. True breakthroughs might require simulating the physical characteristics of optical zoom using AI like Google, or achieving multi-frame synthesis through AI Deep Fusion like Apple. But these require long-term algorithm accumulation and cannot be solved immediately just by "focusing."
Gaming Direction: RedMagic focuses on heat dissipation and battery life; what can AI bring? Perhaps AI scheduling performance to balance power consumption and frame rates, but the Snapdragon 8 Elite Pro itself has an AI engine, so what gaming phone manufacturers can do is system-level optimization. More worth thinking about is whether AI can produce breakthroughs in game graphics rendering and network latency optimization. For example, using AI to predict player operations to preload scenes—this is already close to the concept of cloud gaming + edge intelligence.
Core Contradiction: Limited Resources vs. Infinite Ambition
Any graduate student knows that when you decide to "focus" on a certain direction, it means giving up other possibilities. ZTE-affiliated brands' strategy of "focusing on AI phones + normal iteration" looks great on paper, but will encounter several realistic problems at the execution level:
1. Talent Competition: AI phones require composite talents in deep learning, computer vision, and on-device inference, and these people are often poached by big companies at high prices. If the company is simultaneously maintaining multiple regular product lines, engineers' attention will be scattered, and the AI team is likely to degenerate into a "face-saving project."
2. Supply Chain Pressure: AI phones require stronger NPUs, larger memory, and more complex sensor modules. The procurement cycles and cost controls for these components are completely different from traditional phone designs. If regular product lines continue to use the old supply chain system, it may drag down the R&D rhythm of new AI products.
3. Brand Perception Confusion: Consumers will be confused—are you an "AI phone brand" or a "regular phone brand"? If the Nubia Z90 is essentially still selling imaging and screens, with just a few AI filters added, it doesn't match the title of "deeply focusing on AI." Unclear brand positioning is a major taboo for tech companies.
My Judgment: This is a High-Risk "Parallel Experiment"
From a researcher's perspective, ZTE-affiliated brands' approach is a bit like opening two topics simultaneously in the lab: one high-risk, high-reward frontier project (AI phones), and one mature, stable traditional project (regular iterations). The supervisor (management) hopes to feed the former with the stable output of the latter, but the problem is that the former requires "concentrated firepower" rather than "steady flow."
Personally, I believe a more reasonable strategy might be: Use the cash flow and technical foundation accumulated by regular product lines to incubate 1-2 AI-native core functions in a targeted manner, rather than rolling out the concept of "AI phones" comprehensively. For example, Nubia could focus on AI imaging (forming a moat like the Leica partnership back in the day), and RedMagic could focus on AI gaming experiences (like ROG's control systems). Once these sub-directions are proven viable, then gradually promote them to the entire product line.
One-sentence summary of the core viewpoint: AI phones are not about slapping on labels; they require a reconstruction of the entire system from chips to algorithms to interactions. If ZTE-affiliated brands' "dual-track system" cannot achieve true resource focus, it easily turns into a middle ground that pleases neither side. Of course, as a beginner researcher, my analysis might be too idealistic. But at least, this gave me another angle to think about while reading papers: Every choice of technology route involves trade-offs between people, money, and time. Whether ZTE-affiliated brands succeed might be more worth tracking than the models in papers.
Original Link: https://www.ithome.com/0/975/478.htm
