Google Ditches Snapdragon for In‑House Tensor Chips in Pixel Lineup
Google has formally abandoned Qualcomm's Snapdragon CPUs in its top‑tier Pixel phones, choosing its self‑designed Tensor processors that aim to deliver tighter hardware‑software cohesion, particularly for AI and machine‑learning functions.
Announced alongside the 2021 debut of the Pixel 6, this shift mirrors a growing industry pattern of manufacturers creating custom silicon to set their devices apart. Crafting its own chip lets Google fine‑tune performance to suit Android’s requirements, including live speech recognition, computational photography, and on‑device security.
Since the Pixel line began, Snapdragon processors supplied dependable general‑purpose power but lacked the adaptability needed for the deep‑learning tasks Google now leans on. Tensor gives Google authority over the CPU, GPU, and specialized neural processing units, enabling it to fine‑adjust Android and its AI‑driven services for the silicon.
In addition to speed gains, the transition tackles supply‑chain issues. Depending on one outside vendor can leave a maker vulnerable to part shortages or limited roadmaps. With in‑house silicon design, Google can synchronize chip timelines with its product plans, possibly shortening lead times and strengthening its bargaining position with external suppliers.
Strategically, the decision reshapes the competitive field. Apple has long reaped benefits from its custom A‑series silicon, and Samsung now employs its Exynos chips across many models; Google’s Tensor seeks to level the playing field by delivering capabilities that are hard to duplicate on generic off‑the‑shelf SoCs. Early reviews have highlighted improvements in voice assistant responsiveness and photo processing, though critics note that raw benchmark scores still lag behind the latest Snapdragon offerings.
Analysts argue that cost savings over the long run may also drive Google’s choice. While chip development demands substantial capital, consolidating functions like the modem, security enclave, and AI accelerator onto one die can reduce the bill of materials as production scales.
Looking forward, Google plans to refresh the Tensor architecture each year, with every iteration promising improved energy efficiency and broader AI features. Though the firm hasn’t dismissed the possibility of external silicon for specific market tiers, the custom route seems poised to become a hallmark of its flagship Pixel series as smartphones keep emphasizing on‑device intelligence.
Comments (0)
Be the first to comment.
Join the discussion