Background Development on a
Google-designed
system-on-chip (SoC) first began in April 2016, after the introduction of the company's
first Pixel smartphone, although Google CEO
Sundar Pichai and hardware chief
Rick Osterloh agreed it would likely take an extended period of time before the product was ready. The next year,
the company's hardware division assembled a team of 76
semiconductor researchers specializing in
artificial intelligence (AI) and
machine learning (ML), to work on the chip. Beginning in 2017, Google began to include custom-designed
co-processors in its
Pixel smartphones, namely the
Pixel Visual Core on the
Pixel 2 and
Pixel 3 series and the
Pixel Neural Core on the
Pixel 4 series. By April 2020, the company had made "significant progress" toward a custom
ARM-based processor for its Pixel and
Chromebook devices, codenamed "Whitechapel". At Google parent company
Alphabet Inc.'s quarterly earnings
investor call that October, Pichai expressed excitement at the company's "deeper investments" in hardware, which some interpreted as an allusion to Whitechapel. The Neural Core was not included on the
Pixel 5, which was released in 2020; Google explained that the phone's
Snapdragon 765G SoC already achieved the camera performance the company had been aiming for. In April 2021,
9to5Google reported that Whitechapel would power Google's next Pixel smartphones. Google was also in talks to acquire Nuvia prior to its acquisition by Qualcomm in 2021. Google officially unveiled the chip, named Tensor, in August, as part of a preview of its
Pixel 6 and
Pixel 6 Pro smartphones. Previous Pixel smartphones had used
Qualcomm Snapdragon chips, with 2021's
Pixel 5a being the final Pixel phone to do so. Pichai later obliquely noted that the development of Tensor and the Pixel 6 resulted in more off-the-shelf solutions for Pixel phones released in 2020 and early 2021. In September 2022,
The Verge reported that a Tensor-powered successor to the
Pixelbook laptop with a planned 2023 release had been canceled due to cost-cutting measures.
Design "Tensor" is a reference to Google's
TensorFlow and
Tensor Processing Unit technologies, and the chip is developed by the Google Silicon team housed within
the company's hardware division, led by vice president and
general manager Phil Carmack alongside senior director Monika Gupta, in conjunction with the Google Research division. Tensor's
microarchitecture consists of two large cores, two medium cores, and four small cores; this arrangement is unusual for
octa-core SoCs, which typically only have one large core. Carmack explained that this was so Tensor could remain efficient at intense workloads by running both large cores simultaneously at a low frequency to manage the various co-processors. Osterloh has stated that Tensor's performance is difficult to quantify using synthetic
benchmarks, but should instead be characterized by the many ML capabilities it enables, such as advanced
speech recognition, real-time language translation, the ability to unblur photographs, and
HDR-like
frame-by-frame processing for videos. == Models ==