Google’s upcoming Tensor G5 chipset could mark a significant shift in the company’s approach to mobile silicon.
After relying on Samsung’s Exynos architecture and manufacturing processes since the launch of the first Tensor, Google will likely finally produce a fully in-house designed chip manufactured by TSMC.
Here’s everything we know so far.
When will the Google Tensor G5 be launched?
The Google Tensor G5 is expected to debut with the Pixel 10 series, which is likely to launch in late 2025.
Given the standard annual release schedule for Pixel devices, Google is expected to introduce the Pixel 10 around October 2025.
Dominik Tomaszewski / Foundry
This timeline lines up with reports that Google has started testing early units of the Tensor G5, which were recently discovered via shipping data.
Which phones will use the Google Tensor G5?
The Tensor G5 will make its debut in Google’s Pixel 10 series, succeeding the current Pixel 9 series, which launched in August 2024.
The Pixel 10 series is expected to include the standard Pixel 10, the Pixel 10 Pro, the Pixel 10 Pro XL, and the Pixel 10 Pro Fold.
Lucas Baker
With the Tensor G5, Google aims to address criticism of previous Tensor models, especially regarding heat, battery efficiency and raw performance.
This shift to in-house chip design and TSMC manufacturing should allow Google to better optimize its Pixel 10 series with improved performance and efficiency, something that has been hampered by previous dependence on Samsung’s foundry technologies.
What specifications and features does the Google Tensor G5 have?
In recent months, Android Authority has released two comprehensive reports (first report, second report), one of which comes directly from Google’s gChips division, that give us some details about Google’s future chipsets features.
Here’s an overview of the most anticipated specifications and features of the Tensor G5.
TSMC’s 3nm N3E manufacturing process
One of the key changes for the Tensor G5, codenamed ‘laguna’, is manufacturing on TSMC’s 3nm N3E node.
This is the same advanced manufacturing process used in Apple’s A18 Pro chip, which powers the iPhone 16 Pro and iPhone 16 Pro Max, marking a leap forward in efficiency and performance.
Dominik Tomaszewski / Foundry
Previously, Tensor chips suffered from performance issues related to Samsung’s process nodes, which lagged behind TSMC’s offerings.
The 3nm process could enable higher performance and better power efficiency for the G5, making it competitive with other leading mobile chipsets.
CPU configuration
The Tensor G5 may feature an updated CPU cluster designed to balance power and performance. This configuration would include one Arm Cortex-X4 prime core, a powerful core intended for demanding tasks.
The mid-range is said to consist of five Cortex-A725 cores, which improve multi-core performance, especially useful for multitasking. Finally, two efficiency-oriented Cortex-A520 cores can handle low-power tasks, preserving battery life when the phone isn’t under heavy load.
This new configuration, which emphasizes a larger mid-core cluster, is expected to deliver a noticeable boost in multi-core performance and day-to-day efficiency, a focus for Google.
Chris Martin / Foundry
Recently, a mysterious device believed to contain the Tensor G5 appeared on Geekbench (via GSMArena), codenamed ‘Google Frankel’, and scored 1323 points in a single-core and 4004 points in a multi-core test. As a reminder, the Pixel 9 scored 4185 points in our Geekbench 6 multi-core test.
The leaked specs closely match the expected specs of the Tensor G5, with a CPU configuration featuring one prime core (up to 3.40 GHz), five performance cores (up to 2.86 GHz), and two efficiency cores (up to 2.44GHz).
However, the listing shows a Power VR graphics unit instead of the expected Imagination Technologies DXT-48-1536 GPU. Since Power VR is also from Imagination Technologies, this discrepancy may be due to early testing, and the Geekbench scores should be viewed with caution as they may not reflect the final performance of the chip.
GPU and gaming
For the first time in the Tensor series, Google could be switching to an Imagination Technologies GPU, specifically the dual-core IMG DXT-48-1536 running at 1.1 GHz.
This GPU upgrade could introduce ray tracing, a first for Google’s Tensor chipsets, which will enable more realistic graphics and smoother gameplay experiences in mobile gaming.
Lucas Baker
Additionally, the GPU could offer virtualization support, allowing it to process accelerated graphics in virtual machines.
This addition aligns with Google’s broader focus on virtualized applications and productivity tools, where high-quality graphics can enhance functionality.
Improved AI
The Tensor G5 is expected to see moderate improvements in AI processing with an upgraded TPU (Tensor Processing Unit) that improves machine learning capabilities by around 14% over its predecessor.
This TPU can also be equipped with embedded RISC-V cores, which enable custom operations, making the TPU more versatile and capable of performing various machine learning tasks on the device.
Chris Martin / Foundry
These additions, along with the TPU’s support for on-device training, could expand the ways in which the Pixel 10 series learns over time and adapts to user preferences, allowing the Pixel line to further integrate real-time personalization .
Integrated Fan-Out Package on Package (InFO_PoP)
The Tensor G5 is said to use TSMC’s InFO_PoP technology to integrate RAM directly on top of the chipset, improving thermal management and electrical efficiency.
This arrangement could enable a thinner 1.16mm package and help reduce overall power consumption while maintaining performance.
The Pixel 10 series is expected to include 16GB of RAM, an increase over previous Pixel models, allowing for more efficient multitasking and faster response times.
Dominik Tomaszewski / Foundry
This is everything we currently know about Google’s Tensor G5. We will update this article if we learn anything new.
In the meantime, you can check out what the current flagship Pixel 9 series has to offer, as well as the history of Pixel phones, where we discuss how Google’s phones have evolved over the years.
