双十一作为中国电商的标志性发明之一，已经走过十个年头。
每年都不例外的是，在双十一（和其它购物节），电商平台都尽力用数据表现出成交量之大，增长之高，购物速度之疯狂，等等。
除了常见的秀业绩的必要性外，疯狂的数据还可以 1. 加强运用消费者从众心理，让双十一买买买成为再正常不过的事情，“大家都在这时候买嘛” 2. 强调抢购的速度给消费者更少的思考时间，“要第一时间抢，手慢无”
虽说我相信数据上不会假，但数据背后的内容值得细想。
比如 消费积聚在双十一所以九十十一三个月整体消费总额的影响？为了双十一数字好看牺牲的前后消费？消费总额实际没有增加多少？这个集聚效应应该近些年更明显。
比如 依靠平台整合，把更多渠道平台流量计算到阿里/京东等头部平台里，但实际没有额外增量（对于阿里/京东可能没有额外营收或极低；对于去年整体消费可能增长很低甚至为负，但更大一部分消费被计入了头部平台双十一数字）
两个字难讲。
可以肯定的是，天花板越来越近，高增长信手拈来时代已过。
这周五的WTI原油价格历史性连跌创 34 年以来最长记录（1984年7月）。借这个机会，来谈谈未来石油的角色。
内燃机是一项伟大的发明，由此诞生的汽车经济体量难以估量。很难想象没有内燃机的当代社会能有什么工业生产，交通运输会是怎样。在人类文明史上，内燃机应该有与互联网一样的地位。
从全球原油用途的角度看，当下约有 60% 的需求来自交通运输（陆路，航空、航运等）。其中陆路运输占其中近 80%。大概是每20桶油有9桶用于陆路运输这样的比例。
另一个全球共识，各国家地区都已宣布并付诸实践的，是大力推动新能源汽车的发展，最具代表性的就是Tesla为首的电动汽车。按照各自的时间表，内燃机车的销售还能有10-20年。实际我认为电动汽车时代会来得更快，就像iPhone短短两三年就能让诺基亚和黑莓成为历史。虽然汽车相比于手机周期会更长，但变革来临时，用脚投票的消费者往往超出现有预期。最好的证明 – Tesla Model 3 在今年第三季度已经成为美国最受欢迎销量最好的车型。
眺望一下全球45%的石油需求可能消失的2050年，油价是什么？普通消费者完全不需要有油价的概念，这是一个淡出视野，出现在历史课本里的词汇。
所以说，依靠原油作为主要经济来源的国家都得找出路。沙特首当其冲，也最具代表性。沙特土豪形象深入人心，靠着地下天然资源硬生生造出了一股强大的经济力量。
最能体现沙特的焦虑的，大概就是参与软银的基金。作为 Softbank Vision Fund 最大 LP，沙特在软银帮助下，频频出手在全球找未来行业进行大手笔投资。科技虽然是一个行业，但科技也是一个改变各行业的行业。所以，沙特／软银的投资不能单纯说科技导向，而是diversify + growth + market leader导向，比如 uber（交通），wework（商业地产），compass（住宅地产），guardant health（癌症检测）等等。
贵没有关系，要的是有分量的股权。
Tesla的创始人Elon Musk曾说有可能私有化，依靠沙特的资金支持。这是一笔很合理的买卖。Tesla是沙特地底黄金贬值的始作俑者，买Tesla算是一笔国运hedge。不过想靠私有化Tesla延缓电动汽车时代的到来并不现实，Musk也不会让出他的愿景。有科学的投资，沙特不一定要买Tesla；而且软银也投了其它电动汽车，比如Lucid Motor。
Read More On…
It is not necessarily something new/surprising – traditional media are PAYING tech companies to be promoted.
I found this really ironic in the sense that originally most media should BE PAID to promote something. The reach to readers is built in the gene of media and defines one of the two pillars of media business model.
Tech companies like Twitter/Toutiao, are not just selling promotions/ads. To me, they are more like real estate developers. They are building towns/offices for people to live/work and renting out the places such as tops of buildings… What’s more, they are doing this everyday and creating places everyday with little physical limits and marginal efforts.
When the town is getting crowded, it becomes Manhattan, where real estates are pricey. Those who were invited as “Official Accounts” now needs to pay more money to be seen. The tech companies can run those auctions which no one questions their eligibility (and why Manhattan is in New York).
高通财报 FY Q4 收入低于预期，净利高于预期，苹果手机不采用高通芯片的影响在数字上反映得很明显，包括对于下个季度的预期（9-12月）。
回顾一下智能手机和苹果手机的近几年发展，到与高通的关系，有这么一个逻辑：
2014年开始全球智能手机增量市场感受到相当大阻力，以出货量增长放缓乃至不增长为标志。苹果依靠iphone建立起的帝国开始面临一系列问题。这在苹果公司内部或高层应该被经常讨论，也早已市场感受或预料到（比如2011/2012年）。苹果整体利润水平开始有下滑，尤其用2016&2017年比较之前。收入无非是单价*数量。数量增上不去，只能在单价上做文章，iphone的提价路线开始清晰。另一方面，高通芯片和技术授权的成本开始变得更重要。在高增长时期，大家合作共赢，有问题拖到后面再说；增长放缓，则要把分蛋糕的事情再拿出来讨论。如果高通定价为固定数额，那iphone提价是妥妥地提高苹果利润空间；但事实并非如此。高通以手机单机为基础的计价方式让苹果很头大。2015甚至更早的时候，高通苹果应该就有过相关讨论。但作为行业风向标，苹果的谈判结果势必影响高通与其它手机厂商的合同；从这点上说，高通也绝对不想让。大家都盯着这一代智能手机最后的赚头不放手。矛盾爆发在2016，苹果让生产商拒付高通部分费用。也是为2017年iphone X提价到$999做了铺垫。经过两年多准备和研究，2018年起的iphone已经不再采用高通，转用英特尔。
整个事件对与高通来说自然是不好过。苹果手机卖得最贵，贡献的收入自然多。高通不希望失去苹果这个大客户和行业巨头。但同时，高通很明确自己作为供应商的角色。想成为有话语权的供应商，一定要保证下游客户数量和平衡。苹果固然重要，但这一代智能手机的后半场苹果的声音越来越小。虽然苹果依旧可以赚很多钱，依然是最受欢迎，整体体验最好的手机；但在后半场里，印度中国等市场中购买力不出众的消费者撑起了半片天。苹果做不到的，抓不到的消费者，高通可以通过其它OEM捕捉到。另一个角度来说，苹果高通的纠纷让高通提前跳下了有点下沉的巨轮。即使再往前开，高通从苹果身上也赚不到多少。
苹果7-9月财报和未来预期已经说明了很多；不再报告iphone销量数字也成为让投资人不安的点之一。
所以，即便平行空间里高通苹果还是朋友，有一个修改过的合作，高通这季度开始之后几个财报也不会因此亮眼。
而我觉得高通 sexy 的地方还是在于 5G 和 IoT 的时代。5G 概念下包括可下一代手机，AR/VR，车联网，自动驾驶等；IoT 概念下可以算上智能家居等，越来越多设备将连上网络/云端，但不一定需要 5G。
2019将是高通的转折年。可能看不到大涨的基础；好的话应该是为下一个5年的领导地位打下基础；差的话，很多今天挂在嘴边的应用还要过数年才能实现/实用，以及新技术新标准新共识的出现。
Crystal City near Washington was reported a few days ago by Bezos’ Washington Post as the choice for Amazon’s HQ2; only after 2 days, medias reported that it’s going to be split into two cities.
While the result will be announced soon, I think it is worth to review tech firms’ real estate moves by looking at Facebook, Google, etc. This post will focus on Facebook.
Facebook’s original office (since 2004) was in downtown Palo Alto on Emerson Street. It had 5 or 6 offices, one conference room, and a large common area.
The following expansion in Palo Alto including leases on S. California Avenue and 1050 Page Mill Road.
Facebook moved to Menlo Park in 2011, the old campus of Sun (acquired by Oracle) – an 11 building, 57 acre, 1 million square foot property on the Bayfront Expressway. The purchase was a sale-leaseback with a 15 year long-term lease, with an option to purchase the campus after five years. There were no tax breaks included in the deal with Menlo Park. Facebook has also bought a 22 acre building connected to the campus by a tunnel. It was rumored that Facebook had been considering a leaseback of the property, with the purchase being assigned to a state pension fund. A 15 year leaseback was confirmed today, with the option to buy after 5 years. [TechCrunch]
Besides the move in Silicon Valley, in December 2010, it leased two floors with up to 150,000 square feet at 335 Madison Avenue in New York City, accommodating up to 600 people. It began construction on a $450 million data center in Forest City, North Carolina in 2010 after the first one announced earlier that year in Prineville, Oregon. Facebook also announced a 500-person office in Hyderabad, India, as the first office in mainland Asia, investing up to $150 million to the 50,000 square foot facility.
The number of employees start to explode since 2010.
In a 2013 report, FB has the project, called Anton Menlo, cost an estimated $120 million. It will sprawl over 10 acres of land off Highway 101 in Menlo Park, California. The 630,000 square-foot Facebook town will be walking and biking distance to Facebook’s headquarters. There will be 35 studios, 208 one-bedroom apartments, 139 two-bedroom apartments, and for top Facebook employees, there will be 12 three-bedroom apartments. [Business Insider]
In 2015, FB acquired a 56-acre industrial park immediately south of its current Menlo Park headquarters. The purchase of Prologis Inc.’s 21-building Menlo Science & Technology Park — which industry sources pegged at roughly $400 million. [Silicon Valley Business Journal]
In 2016, Facebook goes from rent to own in Menlo Park in $202 million deal.
During 2017-2018, FB has opened offices and hired at a blistering pace, with enormous new leases in Sunnyvale, Mountain View and Fremont. This month, Facebook leased all the office space in San Francisco’s new 43-story Park Tower, vaulting it into the ranks of the largest tech tenants in San Francisco. Instagram, a major Facebook subsidiary, recently moved 200 employees into another San Francisco office tower. [San Francisco Chronicle]
In June 2018 reports, Facebook has leased an additional 754,000 square feet for offices in Fremont.
In 2017, FB said it will two more Prineville data centers, followed up by a report of $750 million in Sep. 2018, bringing total investments to $2 billion and total footprint to 3.2 million sqft. The expansion will take nearly two years and employ 1,500 construction workers at peak.
Reference / Read More on…
CIIE 2018 is actually the first of its kind so itself doesn’t have a rich “history” to tell. But there are two other exhibits in China worth noting – 1. Canton Fair (广交会) in Guangdong & 2. Expo 2010 (2010 世博会) in Shanghai
Canton Fair was first held in 1957 Spring and has two sessions (spring session & autumn session) a year (62 years and 124 sessions so far) . Its official name was Chinese Export Commodities Fair, which highlights its origin and main function – exports. Its name was changed after the 100th session and has been China Import and Export Fair since then.
World Expo (or World’s fair, International Exposition, etc.), held every 5 years, is more of a showcase than an purposeful event. China won the bid to host Expo 2010 in 2002, one year after it won the 2008 Summer Olympic Games in 2001. CIIE is held in Shanghai.
I watched live keynote and took the following notes –
Part I – welcome & intro
Part II – importance of a global economy
Part III – what China did and will do
Part IV – be confident in China’s growth potential
Part V – Shanghai is great and has its unique positions
Part VI – closing
Shanghai’s role reinforced.
Intense sub-planck undulations are seen as glimmers of distortions when zooming out, and are smoothed out on a higher level. Vice versa, seemingly flat world could be surprisingly dynamic if viewed in a sufficient close manner.
Fluctuations in prices within a short time frame, however dramatic, could be (at least partially) smoothed out in a longer period of time. Vice versa, more details or clues of changes could possibly be found behind the scene even if prices stay the same in a week/month/year.
While sub-planck variations are cancelling out with each other, some changes are constant and going in certain directions, although might be immaterial/unsensible on a daily basis – e.g. our universe is ever expanding.
Certain underlying trends (company, industry, economy, society, species, planet, etc.) are happening definitely in the long run. Trying to uncover those trends and making investments according to those should be an effective long term approach.
[personal notes]
In ITU standards,
Extremely high frequency (EHF) = 30 to 300 GHz in frequency = 10 to 1 mm in wavelength, so called millimeter-wave (mmW)
Ultra high frequency (UHF) = 300 MHz to 3 GHz in frequency = 1 to 0.1 m in wavelength
UHF is where our current mobile networks live on, with 4G mostly on 700 MHz, 1700-2100 MHz, 1900MHz and 2500-2700 MHz across the globe.
Under the Friis Transmission Law, higher frequency has much higher loss (attenuation) in free-space. For mmW, additional transmission losses occur when traveling through the atmosphere are absorbed by molecules of oxygen, water vapor and other gaseous atmospheric constituents.
Important absorption peaks occur at 24 GHz (for water vapor) and 60 GHz (for oxygen).
As range of 5G signals are limited, small cells deployment, collaboration and integration will be essential. Current experiments found a 200-meter range doable.
Appendix – Frequency Allocation
Read more on…
On Oct 8, Intergovernmental Panel on Climate Change (IPCC) issued a special report on the impacts of global warming of 1.5 °C above pre-industrial levels and related global greenhouse gas emission pathways. And a summary version for policymakers.
1.5°C was the target to strive for, agreed in 2015 Paris Agreement. (countries agreed to limit warming to below 2°C above preindustrial levels by 2100)
However, we are nowhere close.
Limiting warming to 1.5°C is possible within the laws of chemistry and physics but doing so would require unprecedented changes.
– Jim Skea of Imperial College London and an IPCC co-author
By “unprecedented changes,” Skea means essentially two things. First, the world needs to start cutting emissions. Instead in 2017, the world’s emissions a new record high. Second, we need to reduce those emissions very quickly. (Quartz)
Many researches have been done to analyze the impacts of a warming of 1.5°C and above.
[NOT for recommendation purposes; more for tracking purposes]
这件事已经想了有一阵子。昨天具体构思后，简单地做一个 5G 股权指数 (index)，并按着指数做这么一个组合，以 $1000／股为起始面值，以 10 月 22 日作为开始日期。
组合很简易，局限于在美股交易的股票。一开始纳入了 7 只，比重有些 arbitrary（更像是根据我现有的持有比例来的）。
这个组合的第一阶段主要是 5G 基础或者说 enabler；等 2-3 年后将调整为侧重因 5G 而爆发的应用。
成分股中分为两类 – 核心 (Core) 和 其它 (Others)，将主要跟踪核心成分股，目前 4 只（高通 康宁 Verizon AT&T）。
说一下主要成分股的主要的非 5G 因素。
高通 – 主要收入继续受到智能手机出货量压力和低价压力
康宁 – 主营业务之一的显示屏单价压力，LCD 需求增长放缓，新技术的份额扩大
Verizon 和 AT&T – 传统业务，媒体业务，尤其是 AT&T 的收购和债务
英特尔 – 服务器 data center 领域要接受 AMD 的挑战，再观察一下准备移到 Core
暂时有约 20% 的现金未分配留着添加调整。
留一个截图
