The Video Streaming Market is Getting Crowded, and That is Okay!
Follow Nick on Twitter @GrousARK
According to research firm Apptopia, Disney+ is adding roughly one million new users a day. Disney stated that its five year goal for Disney+ is 60-90 million subscribers, a figure that now seems conservative. Given its current rate of adoption, Disney+ should reach the midpoint of that target range, 75 million subscribers, two years ahead of schedule in 2022.
What does Disney’s initial success mean for the broader streaming market? A common misconception is that new services like Disney+ will detract from the potential of incumbents like Netflix. While new entrants could cause short-term turbulence we believe that in the long run, the more players in the video streaming market, the merrier.
That said, Disney+’s early success should concern the linear TV market. As shown in the chart below, the roughly 85 million Pay-TV households in the US could decline by roughly 40% by 2024 as more streaming services launch and accelerate the pace of cord-cutting.
Amazon Shows Its Silicon Ambitions at re:Invent 2019
Follow James on Twitter @jwangARK
Just as it has done with consumer goods in retail, Amazon is in-sourcing computer chips in its Amazon Web Services (AWS) division.
At its annual developer conference this week, Amazon announced three new chips, each designed to improve performance and differentiate its offerings from other cloud vendors.First, it announced Amazon Graviton 2, a second generation ARM processor designed for general server workloads. According to Amazon, this new chip offers 40% better price-performance than Intel processors. The second chip, Inferentia, accelerates artificial intelligence inference workloads including answering voice queries, classifying images, and translating text. Amazon claims that the Inferentia cloud instance is 3X the speed of a GPU instance at 40% of the cost. Finally, Amazon revealed AQUA, an FPGA accelerator that speeds up its RedShift data warehousing solution.
Because it controls the software associated with AWS, Amazon can port product offerings to run on its hardware. Over the long run, Amazon is unlikely to pay an 80% markup to Intel and Nvidia when it can build comparable processors in-house.
The rise of cloud computing should cause fundamental changes in the semiconductor industry. In our view, cloud computing companies are likely to usurp the role of the world’s largest chip companies during the next decade.
Commercial Drones Have a Wide Range of Applications Thanks to Improved Battery Range and Autonomous Technology
Follow Tasha on Twitter @TashaARK
Recently, ARK attended TransportUP, a flying car conference highlighting companies on the cutting edge of electric air taxis, batteries, and mapping. Most companies agree that the first applications will be for drones that carry packages instead of people.
As was the case last year, many companies introduced designs for different use cases as they plan to scale.Although the US has approved a number of commercial drone operations including Project Wing and UPS this year, many early stage drone firms are exploiting loopholes in Federal Aviation Administration (FAA) regulations and soliciting the support of local city authorities. Because of battery constraints, ARK’s research has focused on applications for short range flight. That said, despite the improvements in noise emission, short range drones probably will face resistance from local authorities.
Recent improvements in battery range and autonomous technology should enable electric aircraft to take flight and operate more cost-effectively than ever before. ARK estimates that the market for air taxis, parcel drones, and food delivery drones could reach $300 billion in revenue by 2030.
Ethereum Is Delaying its “Difficulty Time Bomb” Once Again
Follow Yassine on Twitter @yassineARK
Last Friday, developers at the 76th Ethereum Core Developer Meeting decided to delay Ethereum’s “Difficulty Time Bomb” for the third time. First introduced in August 2015, Ethereum’s “Difficulty Bomb” was supposed to discourage Ethereum miners from engaging on the proof-of-work chain once proof-of-stake was launched. The bomb increases the difficulty level in the proof-of-work mining algorithm, creaings a substantial lag between the production of blocks and increasing the cost to mine a block significantly.
Introducing the difficulty time bomb in 2015, Ethereum developers failed to consider the significant delays likely in transitioning from a proof-of-work to a proof-of-stake network. Fast forward 4 years and proof of stake has yet to launch on the Ethereum network. The community seems to have been left with no choice but to extend the time bomb, lowering the block time and increasing the frequency of the block rewards.
As a result, the Ethereum community is debating whether the decision to delay the time bomb is an unexpected change in monetary policy or a return to the status quo. Those in favor of delaying the time bomb argue that such a delay will not change Ethereum's monetary policy because block rewards will not change.
The Ethereum community acknowledges that extending the difficulty time bomb is necessary in the absence of a viable option to proof-of-stake. Those who have taken issue with the decision point to Ethereum’s centralized governance. The developers’ “lack of objection” delayed the decision to change the time bomb. While not an explicit change in monetary policy, readjusting the block time did change Ethereum’s daily issuance rate. While the Ethereum community on balance seems to favor defusing the time bomb, the protocol could suffer from a lack of checks and balances in future alterations.