ARK Disrupt Issue 68 Banner

ARK Disrupt Issue 68: Deep Learning, Seamless Payments, and Space Exploration

Please enjoy ARK Disrupt Issue 68. 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. Competition in the Deep Learning Chip Space Should Intensify

2016 was a breakthrough year for deep learning. NVIDIANVDA more than doubled its data center revenues thanks to its monopoly on deep learning chips. Its success has jolted other companies into action: 2017 will be the first year that NVIDIA’s Graphics Processing Units (GPUs) will face serious competition.

For starters IntelINTC, the world’s largest semiconductor company, will introduce the Stratix 10 chip, a Field Programmable Gate Array (FPGA) with specifications comparable to NVIDIA’s flagship Pascal GPU. In a new paper, Intel states that future deep learning networks will use reduced precision operations in which the Stratix 10 chip should outperform GPUs by 60-500%. Later this year, Intel also will release a dedicated deep learning processor from Nervana, which it acquired in 2016. Nervana’s chip aims to improve deep learning performance by tenfold.

ARK Disrupt Issue 68 GPU
Source: The Next Web, Intel.

Startups also are battling for a slice of the deep learning pie. Among the more than seven startup companies working on deep learning processors, Mythic just came out of stealth mode. Reportedly, its chips will enable deep learning to take place directly in memory, without the need for an additional processor, not only cutting power consumption dramatically but also allowing Artificial Intelligence (AI) to be embedded in wearable applications.

NVIDIA’s processors are the de facto standard among developers, giving it a meaningful competitive edge in the deep learning space. As the sole survivor of the GPU wars of the 1990s and no stranger to competition, NVIDIA is likely to innovate its processor architecture aggressively, ultimately dedicating a chip exclusively to deep learning. Its annual GTC developer conference in May should be quite illuminating.

2. With Apple Pay, Start Your Transaction on Your Mac…Complete it on Your Watch!

Apple Pay is redefining “seamless payments” as it eliminates friction in paying for goods and services. As discussed in its recent security whitepaper, Apple Pay is enabling a feature in which a transaction starts on one device, say a MacBook, and completes on another, say an Apple Watch or an iPhone.

To secure this platform, AppleAAPL is using an end-to-end encrypted Intrusion Detection System (IDS) including software that monitors networks for malicious activity. IDS uses keys for encryption, each one unique to a device and identified by a credit card and other metadata. Importantly, Apple secures the data on its device. Once users authorize payments on an Apple device, Apple transfers the encrypted data from users’ iPhones or Apple Watches to their Macs,1 ultimately delivering it to merchants’ websites.

This feature should reduce “cart abandonment,” increasing customer conversion with a simpler and more secure payment process. Its integrated security platform, including hardware and software, should protect customers from security concerns as they complete transactions across different devices.

3. Tencent and Softbank Invest in Autonomous Mobility-as-a-Service

This week, TencentTCEHY took a 5% stake in TeslaTSLA. Immediately following the announcement, Tencent’s President Martin Lau created a twitter account to tweet several messages to Elon Musk:

ARK Disrupt Issue 68 Tweet

ARK Disrupt Issue 68 Tweet 1

A Tencent spokeswoman told the Wall Street Journal, “Elon Musk is the archetype for entrepreneurship, combining vision, ambition, and execution.”

Tesla is not Tencent’s first investment in autonomous electric mobility. It also invested in Didi Chuxing, China’s largest ridesharing company, as well as in HERE, the autonomous map provider, and Nio and Future Mobility, two Chinese electric autonomous vehicle manufacturers.

In other news this week, SoftbankSFTBY is considering a $6 billion investment in Didi. If the deal materializes, Softbank’s could be the largest funding vehicle for a Chinese tech company to date, according to Bloomberg. These funds could support Didi’s autonomous technology development efforts.

According to ARK’s research, the autonomous mobility-as-a-service market will reach $10 trillion in gross revenues by the early 2030’s, with China dominating the market. Didi is well positioned as a local player in China, and now Tesla should benefit from Tencent’s 5% position.

To read more about driverless cars in China, please stay tuned for ARK’s research articles on China’s autonomous vehicle landscape.

4. SpaceX Does It Again!

On Thursday evening, for the first time in history SpaceX successfully launched and landed a refurbished orbital class rocket! In addition to reusing and then recovering a Falcon 9 rocket, SpaceX was able to recover the $6 million nose cone for the first time. For a launch priced at $62 million, the nose cone cost save is no small matter.

Following the successful mission, Musk tweeted that his next goal for spaceX is to achieve a relanding within 24 hours, as opposed to the one-year rule of thumb previously established. SpaceX might attempt to recover the upper stage of the rocket and potentially could move the Falcon Heavy to reuse two Falcon 9 first stages.

ARK’s research suggests that SpaceX’s strategy for reusable rockets will focus on economies of scale as it attempts to reach and inhabit Mars. As SpaceX’s success rate in recovering rockets improves, we will scrutinize future rocket launches and evaluate their long-term viability.

5. Allogeneic Transplantation Should Democratize Stem Cell Therapy

The world’s first stem cell transplant to treat macular degeneration using donor stem cells could herald the rise of induced pluripotent stem cell (iPSC) banks. These banks could make heretofore expensive and lengthy stem cell therapy accessible to the masses in a fraction of the time it currently takes. In a procedure this week, Japan’s Yasuo Kurimoto received retinal cells from an anonymous donor’s reprogrammed stem cells. While such an allogeneic transplantation usually would run the risk of tissue rejection, iPSC pioneer, Shinya Yaminaka, found a way to bypass that problem and potentially democratize stem cell therapy.

Three genes that code for Human Leukocyte Antigens (HLAs) – proteins that trigger an immune response – also can match stem cell donors with appropriate recipients. As a result, Yamanaka is establishing an iPSC bank in which stem cell lines will be preserved for future therapeutic use cases, shaving both costs and months off of preparation and treatment time. According to Yamanaka, given the relative homogeneity of Japan’s population, 100 donor stem cell lines would be adequate to cover Japan’s entire population. By contrast, 200 donor stem cell lines would be necessary to provide coverage for the US population.

iPSC banks offer a promising future for stem cell therapy by making the inputs for research both accessible and cost-effective. That said, these early times for allogeneic stem cell treatments are beckoning long-term follow-up studies, including the safety and efficacy of potentially important treatment breakthroughs.

  1. Using iCloud Keychain


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.

Join the Conversation

Be the First to Comment!

avatar
wpDiscuz