Google’s Path to Vertical Integration with Tensor SoC – Become Apple or Stay Google?

  • Google finally launched its long-rumored  code-named “Whitechapel” in-house smartphone System-on-Chip (SoC) yesterday.
  • Google is calling it “Tensor” SoC adding to its portfolio of Cloud & Edge TPUs to showcase its Artificial Intelligence (AI)/ Machine Learning (ML) prowess.
  • This first custom-built SoC is designed for the upcoming Pixel portfolio in the Fall’2021 – Pixel 6 / 6 Pro.

The Need for Tensor SoC

  • Sitting on tons of data coming from more than 3 billion Android users, billions of Google search, Maps, Photos, Assistant (voice), applications and services Google has transformed into  an AI powerhouse with ability to build accurate ML models to improve the user-experience across the same properties with generate so much data.
  • Advanced ML algorithms, for example, to improve the on-device photography using Night Sight-based computational photography, real-time multimodal language translation, visual search leveraging Multitask Unified Model trained across 75 languages, powerful Google Photos with features such as AI/ML powered Little Patternsintelligent app and content recommendations, real-time collaboration across Google Docs with AI/ML features such as Smart Canvas, Smart chips and other experiences.
  • Further, the AI/ML expertise while helps increase the overall UX, it also helps Google accurately profile users behaviors, patterns, predict needs, build a powerful recommendation engine and get paid top dollars for the targeted advertising from the marketers, driving Google’s core ad business model.
  • Further, looking at Apple, becoming more vertically integrated gives Google full control of the entire stack from chip to OS to middleware to cloud.
  • This gives Google a unique advantage to showcase the Google’s core software and AI/ML expertise trickling down from cloud to chip and back from chip to cloud.

Google Pixel Strategy: Vertical or Horizontal?

Google having built this vertical in-house smartphone solution with Tensor SoC (Samsung partnership), Google Android, Google Mobile Services (GMS), & Play Store (in-house), Design, IP, Product Management expertise (HTC & Moto acquisition). It has atleast two major ways in how it can execute this vertical expertise:

Google’s Pixel Strategy Option 1 :: Becoming Apple

  • According to our Smartphone Market Monitor service, Google Pixel smartphones had less than 0.3% market share by volume globally at the end of Q1 2021.
  • It remains to be seen if this full-stack controlled Pixel device performs better in the real world and can move the needle for Google
  • One gap in Google’s vertical expertise, which is an important one, is a mature 5G modem-RF subsystem expertise to offer optimal premium smartphone experience.
  • Google’s partner for Tensor, Samsung, is still generations behind Qualcomm and Samsung’s Mobile division uses Qualcomm 5G solutions on its flagships devices in key premium markets (which are Google’s key target markets as well).
  • Integrating an advanced 5G modem and optimizing the RF system-level experience is super-difficult and hence Apple is also using Qualcomm while it struggles to develop the 5G modem-RF in-house.
  • Additionally, Qualcomm is ahead by almost four generations when it comes to the 5G mmWave solutions for premium 5G smartphone experiences which is going to go mainstream in couple of years and every Google Pixel flagship will need an advanced mmWave 5G SKU.
  • Further, Google will need atleast two years to scale the Tensor down to mid-range Pixel devices as this is a premium solution
  • Additionally, Google has under-invested (on purpose) in terms of reach and market push to sell the Pixel devices across markets and channels.
  • If Pixel has to be successful and if it has to scale, Google needs to invest a lot in terms of channel partners, certifications (for the new chip), sales force, marketing and other areas.
  • So, if Google decides to go that route and scale the vertically integrated Pixel across geographies just like iPhone compared to a handful of markets right now, then it will be competing head-on with its OEM partners and close partners such as Samsung
  • So in summary, with the option 1 to follow Apple, doesn’t look entirely achievable in near-to-mid term without burning bridges and tons of investments, and ready to give up the ecosystem-led approach for Android platform.

Google’s Pixel Strategy Option 2:: Staying Google

  • Another strategy which could bode well with Google’s future is to remain a horizontal player and use the vertical expertise for Pixel just as a “showcase” of Google’s capabilities.
  • So what Google needs todo is license the AI/ML IP or silicon to other OEMs bundled with Google Android and other services, allowing partners to build “Intelligent Android Phones”
  • Depending on the level of Google expertise integration, OEMs could differentiate and opens up an entirely new and highly scalable revenue model for Google
  • Only caveat being Google will compete head-on with its key silicon partners such as mainly Qualcomm and MediaTek in long run for premium experiences.
  • With this approach, Google not only is able to scale its software but silicon and AI/ML APIs expertise to more Android phones.
  • Google should not forget its core competency – Software, Big Data, AI/ML and Cloud which in the end drives its core high-margin evergreen business modelAdvertising

Overall its good to see Google jumping in the SoC space adding new layer of chip-to-cloud expertise and could be beneficial for Google in the long-run either way. Though, Google need to be prudent on what approach it takes with this move and above all lives up to the expectations of offering a solution which is better than Apple or Qualcomm.