this week i'm thinking about – #7 – MVNOs

Hey friends,

Hope you're having a great last day of 2018! I’m writing a weekly blurb about 1) one thing I learned, and 2) one question I’m thinking about that’s broadly tech-related.

If you have thoughts, please do reply! I’ll paraphrase the best responses in the next newsletter. My goal is to start conversations with people thinking about similar topics through different lenses. I'm trying this out with a small group of VIP friends, but if you know anyone who would be interested in the discussion, please do forward this along or have them subscribe here.


What I learned about: MVNOs 📱
It’s no surprise that wireless providers are among the least-loved companies in the world. The major US wireless companies – AT&T, Verizon, Sprint – all have sub-20 NPS. But there’s been surprisingly little innovation in telecom in the past 20 years, partially because of the high capex required to get started with competitive coverage, leading to inconsistent returns of telecom startup investments. MVNOs – mobile virtual network operators – have traditionally been subsidiaries of larger telcos that want to access a new consumer segment under a new brand, or communications companies that want to access telecom infrastructure (Comcast via Xfinity, Google via Fi). But as new revenue-sharing models and customer service efficiency increases have started to work, a new set of MVNO startups have begun to take hold, particularly in Asia and Europe: Zero Mobile,, Dataxoom, etc. I expect similar players will emerge in the US, if they can convince incumbent telcos that they won't cannibalize their business.

What I’m wondering about: facial recognition and political ownership 🇨🇳🇺🇸
Chinese facial recognition startups are significantly more sophisticated than their best-in-class equivalents in the US. Sensetime has raised $1.6B so far, and Face++ has raised $1.1B to date. This seems to pose a Huawei-like risk: if the US doesn’t control its own facial recognition technology, then China will. My mental model tells me a player with massive datasets like Google is best poised to pull off a US equivalent, although the Chinese players winning this category were both founded in the last seven years. Will there be startups in the Western world that will provide an equally sophisticated, but more trustworthy alternative?

Top replies from last edition (average time to unicorn, applications of cheaper CV):
Josh mentioned beauty as a category where cheaper CV can lead to better UX by driving more interactive products, and noted Chinese companies like Meitu have already started to win this category on the other side of the Pacific.

Melisa noted that cheaper computer vision will lead to better reverse-image lookups. Google Images has had a light version of a reverse image search for exact image matches, but doesn’t have more sophisticated features like face-to-identity lookups (think Shazam for faces), which could be useful for security and government applications.

What I’m reading