ark-disrupt-banner-139

ARK Disrupt Issue 139: Tech Cost Declines, AI, Bitcoin, Fintech, & Micro Mobility

Please enjoy ARK Disrupt Issue 139. This blog series is based on ARK Brainstorming, a weekly discussion between our CEO, Director of Research, thematic analysts, ARK’s theme developers, thought leaders, and investors. It is designed to present you with the most recent innovation takeaways and to keep you engaged in an ongoing discussion on investing in disruptive innovation. To read the previous issue, click here.

1. Sounds Nice: What’s It Gonna Cost Me?

Follow @wintonARK on Twitter

In 1980, AT&T commissioned a consultant to forecast demand for cellular telephony over the next 20 years. McKinsey concluded that this new technology was cost-prohibitive and would garner fewer than a million subscribers by 2000. As a result, AT&T T halted its cellular initiatives, giving up on a market that would grow to 109 million subscribers by 2000 and roughly 4 billion subscribers today. AT&T divested its cellular assets in 1984 and then paid dearly for its mistake with the $23 billion purchase of McCaw Cellular in the mid-nineties.

Was AT&T wise to be technology-shy based on its experience?

During the seventies, AT&T spent $3 billion on picture-phone technology, asserting that it would attract more than one million users by 1980.1 That vision did not play out, so AT&T made another attempt at a commercial picture-phone product in 1982. The picture-phone connection, which required a dedicated room and cost roughly $4,000 per hour, did not develop commercial traction, so AT&T shut it down, only to revive it and end it again in 1992.

How were the executives at AT&T to know? How were the consultants at McKinsey to know? What leads one technology into widespread adoption while another languishes?

The clue to the answer is in the costs.

In 1983, AT&T’s picture-phone meeting system cost roughly 30x more upfront and 60x more per minute than Motorola’sMSI newly launched DynaTAC cellular phone. We believe that a simple analysis of cost-declines and price-elasticity of demand would have pointed AT&T executives to a different strategic course.

Modeling the cost declines associated with new technologies lies at the core of our forecasting. Without understanding what a service will cost, today and in the future, and without considering what early adopters will be willing to pay, consultants and analysts will have difficulty forecasting the size of any market correctly.

2. IBM’s Watson for Oncology Is a Small Data Product in a Big Data World

Follow @jwangARK on Twitter

A year ago, IBM’sIBM artificial intelligence (AI) service, “Watson for Oncology”, seemed to be taking off. Hospitals in the US and Asia were signing up so that doctors and patients could get second opinions on treatment options for complex cancer cases. Since then, the same hospitals have stopped or scaled back their commitments to Watson, concluding not only that it failed to recommend new treatment options but also that it tended to generate inaccurate or unsafe recommendations. Internal IBM documents say that Watson often provide “unsafe and incorrect treatment recommendations.”

The initial vision for Watson for Oncology involved doctors at the Memorial Sloan Kettering Cancer Center (MSCC) who would train the AI system on real patient data and treatment options. With enough samples and data, Watson should have become a world class oncologist with more experience and a more rapid learning rate than any human doctor. In reality, only one or two MSCC doctors trained Watson for each kind of cancer and, instead of learning off of large data sets, it trained on just a few hundred examples.

While IBM had the right idea with Watson, poor execution, technical ineptitude, and overzealous marketing resulted in a product that failed to live up to the hype. We believe artificial intelligence and precision medicine will change the course of health care decision-making, but IBM’s Watson is unlikely to play the central role it once envisioned.

3. Do Not Confuse Bitcoin’s Base Layer Transactions Volume with Its Transactions Value

Follow @yassineARK on Twitter

Debates around Bitcoin’s scalability are rampant, with transactions-per-second one of the primary metrics for comparing blockchains to traditional payment platforms. Questioning its long-term viability, critics point to Bitcoin’s low base-layer transaction volume throughput. As shown below, Bitcoin’s annual throughput capability at the base layer is several orders of magnitude below any existing payment system, suggesting that it is wildly inefficient.

ARK Disrupt Issue 139 Graph 1

Critics fail to realize, however, that while the number of transactions might be low, thevalue per transaction on Bitcoin is quite high, as shown below. As measured by annual transactions value, Bitcoin’s base layer is supporting two times the throughput of PaypalPYPL today and is less than an order of magnitude smaller in throughput than VisaV. Since year end 2016, Bitcoin’s total transaction value has increased ten-fold, suggesting that the use case for its base layer as a settlement network for large transactions is real and growing.

ARK Disrupt Issue 139 Graph 2

Debates around throughput often neglect the potential impact of second layer scaling solutions.  Just as Internet protocols are layered to serve independent but complementary functions, ARK believes that Bitcoin’s scalability solutions will be layered on top of its base. The Lightning Network, a bidirectional payment channel network that facilitates high-volume low-value spending, is an early example of a second layer scaling solution that could enable Bitcoin to compete effectively with existing payment systems.

4. Consumer Demand for Digital Wallets Is Mounting

Follow @bhavanaARK on Twitter

Digital wallets are empowering consumers to bypass traditional bank branch networks, particularly in emerging markets like China. Alibaba’sBABA Alipay and Tencent’sTCEHY WeChat, for example, allow access to services that span payments, investments, lending, and insurance.

Digital wallets are gaining momentum in the US as well, thanks to SquareSQ Cash, AppleAAPL Pay, and Paypal’s Venmo. As of year-end 2017, Square Cash had 7M customers on its platform, and Venmo, 18M.2 According to GoogleGOOG search trends, all three services are attracting similar levels of interest today, though Square’s Cash app has entered an accelerated growth phase while Venmo and Apple Pay are growing more steadily, as shown below.

ARK Disrupt Issue 139 Graph 3

We believe that digital wallets could become the consumer’s bank branch, providing cross-selling opportunities for both banks and retailers. In addition, they will be able to connect the consumer to cryptoassets. With its bitcoin functionality, Square Cash is poised to become a crypto wallet provider.

5. Micro Mobility Might Work in Less Temperate Climates, At a Higher Cost

Follow @skorusARK on Twitter

A frequent criticism of electric scooters and bikes is that they may be effective modes of transportation in California, but not in the Midwest or on the East Coast during harsh winters. New York City’s docked bike-sharing service, CitiC Bike, offers some good perspective. As shown below, during the winter the number of Citi Bike trips drops by more than 60%. As a result, in New York City ride sharing services may have to charge more on average than those in more temperate climates to compensate for lower utilization rates.

In these early days of micro mobility, electric scooters are not distinguished in quality or design. If micro mobility is not a fad, the proliferation of ridesharing services and competition should solve those issues.

ARK Disrupt Issue 139 Graph 4

  1. This forecast is from AT&T’s 1969 annual report, which anticipated those customers spending more than $1000 per year.
  2. ARK research and calculations.


ARK's statements are not an endorsement of any company or a recommendation to buy, sell or hold any security. For a list of all purchases and sales made by ARK for client accounts during the past year that could be considered by the SEC as recommendations, click here. It should not be assumed that recommendations made in the future will be profitable or will equal the performance of the securities in this list. For full disclosures, click here.