Google Tensor G5 Archives ~ My Mobile India

Highlights Google Tensor G5 will be the first Tensor chip fully designed by Google and manufactured by TSMC using a 3nm-class process node. Key upgrades expected include a new IMG DXT GPU, a fully custom ISP for enhanced camera performance, a custom memory controller and a new Chips&Media video codec. The Tensor G5 is expected to be Google’s most advanced chip to date and will debut in the Pixel 10 series. Google Tensor G5 to succeed the previous generation Tensor G4. (Image credit - Google) The Google Tensor G5 is set to power the upcoming Pixel 10 series and is reportedly shaping up to be a major upgrade. Unlike previous Tensor chips, which were developed in collaboration with Samsung, the G5 will be entirely designed by Google and manufactured by TSMC using its 3nm-class process node. A new report from Android Authority confirms this shift and provides deeper insight into the G5’s internal components. Tensor G5 - What’s Changing? New GPU - Google is reportedly switching from the Arm Mali-G715 MP7 (used in the Tensor G4) to Imagination Technologies’ IMG DXT GPU for better graphics performance. Fully Custom ISP - Google will introduce a fully custom Image Signal Processor (ISP), moving away from the partially modified Samsung ISPs used in previous Tensor chips. This could significantly enhance Pixel 10’s camera performance. Explaining this, the Android Authority report reads, “The previous chips only had a few custom Google-designed blocks to complement the regular Samsung ISPs, but that changes now. The Tensor G5 will have a fully custom ISP from the front-end to the back-end parts of the pipeline.” Custom Memory and Power Modules - The chip will feature a custom memory controller, system-level cache, and power modules likely optimising efficiency and speed. New Video Codec - Instead of using its BigWave AV1 video codec and Samsung’s Multi Format Codec (MFC), Google will now rely on Chips&Media’s WAVE677DV, which supports AV1, VP9, HEVC, and H.264 for video encoding and decoding. The report notes, “This is where things start deviating from past generations — previous Tensor chips used a combination of two video codecs: Google’s “BigWave” AV1 codec core and Samsung’s MFC (Multi Format Codec) for the other formats. It would seem logical for Google to use something built in-house, perhaps even for all the formats, but that’s not the case. Instead, Google uses third party IP to handle all the formats. Specifically, the core used is Chips&Media’s WAVE677DV, spec’d to handle up to 4K120 encoding and decoding in AV1, VP9, HEVC and H.264.” Third-Party Components - Tensor G5 will integrate third-party USB, PCIe, and I3C components, along with interfaces for DSI (display), DisplayPort, flash storage, and LPDDR5x memory. The Tensor G5 is set to be the most advanced Tensor chip yet with these upgrades and is bringing big improvements in performance, graphics, camera processing and efficiency to the Pixel 10 series. FAQs Q1. What makes the Tensor G5 different from previous Tensor chips? Answer. The Tensor G5 is the first Tensor chip fully designed by Google and manufactured by TSMC using a 3nm-class process node, unlike earlier chips co-developed with Samsung. Q2. What are the key upgrades in the Tensor G5? Answer. Key upgrades include a new IMG DXT GPU for better graphics, a fully custom ISP for improved camera performance, and a new Chips&Media video codec supporting multiple formats. Q3. What improvements can we expect from the Tensor G5 in the Pixel 10 series? Answer. The Tensor G5 will bring significant advancements in performance, graphics, camera processing, and efficiency, making it Google’s most advanced chip yet. Also Read: Google’s Tensor G5 (Pixel 10) and G6 (Pixel 11) Major Process Upgrades Revealed Also Read: Google Pixel 10 Pro and Pro XL CAD Renders Reveal Familiar Design

Google Tensor G5

Google Tensor G5 New Leak Reveals What Key Upgrades To Expect

Google’s Tensor G5 to Use TSMC’s 3nm Process, Enhancing Performance and Efficiency in 2025

Google Shifts Tensor G5 Chip Production to TSMC for Enhanced Performance