Apple envy: Android makers race to catch up to Face ID

Three weeks ago they were still trying to beat Touch ID.

From KGI’s Ming-Chi Kuo, via AppleInsider:

Kuo in a report seen by AppleInsider says inquiries into potential 3D sensing solutions from Android brand vendors have at least tripled since Apple unveiled its TrueDepth camera system and Face ID in September.

Prior to the announcement of iPhone X, the Android camp was focused on delivering incremental advancements in optical fingerprint recognition technology, namely systems capable of operating through a device screen…

Compared to facial recognition, notably facial recognition of the variety introduced by Apple, traditional capacitive sensing systems are now seen as mere spec upgrades.

“3D sensing not only enables facial recognition in security applications and allows users to create fun expressions like Apple’s Animoji, on a more important level, it is a key factor in the development of AR,” Kuo writes. “We therefore believe brand vendors are willing to spend more for related components.”

My take: Kuo has a way of delivering his Apple scoops under darkening skies, in this case the threat of Asian fast followers. To Kuo’s credit, he was the first to report last February that Apple was working on facial recognition for the iPhone X. But five months later he was sill writing about it in the context of disappointments, technical challenges and shipping delays. See, for example, this gem from July 3:

“We predict the OLED model won’t support fingerprint recognition, reasons being: (1) the full-screen design doesn’t work with existing capacitive fingerprint recognition, and (2) the scan-through ability of the under-display fingerprint solution still has technical challenges, including: (i) requirement for a more complex panel pixel design; (ii) disappointing scan-through of OLED panel despite it being thinner than LCD panel; and (iii) weakened scan-through performance due to overlayered panel module. As the new OLED iPhone won’t support under-display fingerprint recognition, we now do not expect production ramp-up will be delayed again (we previously projected the ramp-up would be postponed to late October or later).”

See also: Did Apple really kill the fingerprint reader?

11 Comments

  1. David Emery said:

    iPhone X is the first device to support -3D- facial recognition, right?

    0
    October 7, 2017
      • David Emery said:

        That, of course, is a point that most of the ‘complainers’ about Apple’s facial recognition technology missed (specifically those who said, “it’s been tried and hasn’t worked well”)

        1
        October 7, 2017
  2. Ken Cheng said:

    Didn’t Ming-Chi Kuo also say Apple was a year and a half to 2 and a half years ahead due to its 3d facial recognition, or was that some other analyst?

    1
    October 7, 2017
  3. Gregg Thurman said:

    I continue to find Kuo an unreliable source. At best he is correct (sometimes) BECAUSE he puts out so many, often contradictory, statements. On some level if you throw enough spaghetti at the wall some will stick.

    I do believe, without Kuo’s recent statements, that Apple is two to three years (if not four) ahead of Android for AR and Face ID implementation.

    To make AR and Face ID possible will require a major revamp of Android by Google. Then it will require Android manufacturers to design, test those features before release. Technology developed by Apple (TrueDepth) utilizes components that the blogosphere happily reports is causing “production delay” problems for iPhone X.

    I’d wager that Apple, as has been its practice since the original iPod, has locked up supply of those KEY components (3D sensor) for the next 3 – 4 years. So even if Google upgrades Android to AR and Face ID capability, Android manufacturers will have limited access to key components as there are limited supply alternatives, with more coming online a dubious prospect without a viable customer. High end Android smart phones aren’t reason enough to invest in 3D sensor manufacturing when an elephant already exists in the marketplace. Even if such a manufacturer should appear, without major financial support from a key customer, ramp will be slow.

    Now about those “delay” rumors.
    https://9to5mac.com/2017/09/27/iphone-x-delay-report-3d-sensor/
    “The WSJ has now backed a supply-chain report that poor 3D sensor yields were delaying mass production of the iPhone X.

    Yesterday’s report said that Apple’s suppliers were hitting significant yield issues with the 3D sensors used in the TrueDepth camera module of the iPhone X. The problems were said to be so severe that production was being measured in the tens of thousands of units per day …”

    I’d like to point out that ‘tens of thousands” is an accurate descriptor up to 190 thousand units per day. Above 200 thousand units per day the descriptor wound be HUNDREDS of thousands.

    At 190K per day a 90 day production run (30 days before launch to end of quarter) yields 17 million iPhone X units during the December quarter. Its important to note that WS estimates for iPhone X unit sales has been in the 12 Million to 15 Million units during the December quarter for some time (Kuo recently reduced his estimate to 30 Million).

    Let’s put 17 Million into perspective. That volume amounts to 22% of December quarter average unit sales for the iPhone 6/Plus, iPhone 6S/Plus and 7/Plus. Does anyone thing demand for a $1000 (minimum) iPhone, in a 2 month window (vs prior years 3 month windows) is realistic? I think not.

    Minimum WS estimates at 12 Million units amounts to ~15% of December quarter average unit sales for the iPhone 6/Plus, iPhone 6S/Plus and 7/Plus. I believe that amount remains optimistic. I am modeling 10 Million iPhone Xs for the December quarter and still get 8% YoY iPhone unit sales growth with a corresponding YoY increase in ASP of nearly 4.5%.

    Lastly, has anyone else noticed that reports of iPhone X production delays have all but disappeared in the last 10 days or so?

    I firmly believe that all claims of low component yields are false, thereby making Kuo the ultimate liar.

    1
    October 7, 2017
  4. Fred Stein said:

    Image recognition requires advanced AI, like Neural Networks and Deep Learning.

    What about all those claims that Apple was way behind in AI? But we must admit, Siri is behind – at least for now.

    0
    October 7, 2017
    • Gregg Thurman said:

      I think SIRI is going to see major (leap frog?) improvements with the introduction of the Home Pod.

      0
      October 7, 2017
  5. Gregg Thurman said:

    [Does anyone thing demand for a $1000 (minimum) iPhone, in a 2 month window (vs prior years 3 month windows) is realistic? I think not.]

    Should have read, “Does anyone think that kind of demand (15% of total iPhone unit sales) for a $1000 (minimum) iPhone, in a 2 month window (vs prior years 3 month windows) is realistic? I think not.

    0
    October 7, 2017
    • Gregg Thurman said:

      Kuo’s iPhone X unit shipments for the December quarter estimate (30 Million) equals 40% of average total iPhones sold in the prior three December quarters. New model iPhones typically account for 70% to 75% of total units sold. Just how unrealistic is Kuo’s estimate when iPhone X is “production constrained”?

      You can only have it both ways if you are making stuff up AND AREN’T HELD ACCOUNTABLE by the same blogosphere that mindlessly repeat your baseless assertions.

      2
      October 7, 2017
  6. Gregg Thurman said:

    Another argument to support Apple’s lead in AI and Face ID.

    http://www.digitimes.com/news/a20170925PD201.html

    “While Google still holds the advantage in the cloud-AI sector, Apple appears to have outraced Google in the development of on-device AI products with the launch of its iPhone X and iPhone 8 devices.

    The built-in dual-core neural engine chip incorporated within the A11 processors, which power the iPhone X and iPhone 8 devices, integrates functionality such as machine learning, inference model and related algorithm to perform as an application-specific integrated circuit (ASIC) chip for image recognition and achieve hardware acceleration.”

    Android manufacturers must also rely on Qualcomm for the chip power to drive AI. Cloud AI requires an internet connection and will be hampered by throughput speed issues.

    1
    October 7, 2017

Leave a Reply