Google’s Tensor chips have always been a bit of a mixed bag: ambitious in AI features, but often held back by hardware choices that feel a step behind the competition. A fresh leak suggests that trend may continue with the upcoming Google Tensor G6, and this time the spotlight is on a surprising decision around graphics.
According to the leak, Tensor G6 is expected to use the PowerVR CXT-48-1536 GPU, a graphics processor design that originally debuted in 2021. If accurate, that would mean Google’s next flagship Pixel chip could ship with a GPU architecture that’s roughly five years old by the time Pixel 11 devices arrive.
That’s a tough pill to swallow, especially because phone buyers increasingly rely on GPU performance not just for gaming, but also for smooth high-refresh-rate scrolling, image processing, advanced camera effects, and on-device AI features that can touch everything from video editing to real-time enhancements.
On the CPU side, the news sounds more encouraging. Recent reports indicate Tensor G6 will finally move to newer ARM CPU cores, including C1 Ultra and C1 Pro, with the single “big” performance core reportedly clocked at 4.11GHz. That should translate into better responsiveness for demanding tasks and improved overall performance compared to chips that leaned on older CPU designs.
However, Tensor G6 may also change its core layout. Instead of sticking with an eight-core configuration like the prior generation, the leak points to a seven-core setup in a 1+4+2 arrangement. While we can’t confirm the reasoning, a smaller configuration can sometimes be a cost and efficiency play—reducing silicon area, complexity, and potentially helping yields.
So why would Google pair newer CPU cores with an older GPU? The apparent strategy is die size reduction. By using a smaller or older graphics solution, Google could keep the chip physically smaller, which can help control manufacturing costs and protect margins—especially at a time when component pricing, including memory, remains a pressure point across the industry.
The bet, as described, is that Google’s specialized AI hardware will cover some of the workload. In other words, the NPU (neural processing unit) could shoulder more of the heavy lifting for AI-centric features, helping offset weaker traditional graphics performance in certain scenarios. That might make sense for specific machine-learning tasks, but it doesn’t fully replace what a more modern GPU can deliver in graphics-heavy apps, games, and compute workloads that still lean on GPU acceleration.
For anyone considering a Pixel 11 series phone in the future, the biggest question will be whether real-world performance matches the premium expectations that come with a flagship price. Even if AI features look impressive on paper, consumers tend to notice when gaming performance lags, when sustained performance drops under heat, or when long-term software updates outpace the hardware’s ability to keep up.
There is at least one other notable upgrade expected alongside Tensor G6: a new Titan M3 security chip. Google’s Titan security coprocessors are designed to provide hardware-level protection for sensitive user data, including things like encryption keys and biometric information. For privacy and device security, that’s a meaningful step forward—and an area where Google has historically tried to differentiate Pixel devices.
If the leaks hold true, Tensor G6 could end up being a chip of contrasts: promising CPU improvements, a stronger security story, and a potentially controversial decision to lean on an older GPU design in the name of cost control and a smaller die. The final verdict will come down to how well Google balances performance, efficiency, thermals, and the day-to-day experience that Pixel buyers actually feel.
