The Facebook plan to disrupt the cloud
An alternative for the Libra project
Cemeteries are full of companies that didn’t understand the power game.
Reflecting on my text from last week, I realized that there may be a deeper objective in the Libra project. Marc Zuckerberg cannot be so naive and think that he will be able to impose an alternative currency to sovereign currencies. This is only speculation but can explain Marc Zuckerberg's strong interest in the crypto universe.
From MarketWatch (October 26,2019)
Microsoft was officially awarded a $10 billion cloud contract with the Pentagon on Friday - marking a surprise victory over Amazon Web Services, which was thought to be the frontrunner for the deal.
This deal, called the Joint Enterprise Defense Infrastructure (JEDI) contract, is a 10-year contract to help the Department of Defense move its sensitive data to the cloud. Companies like IBM and Oracle also competed for the contract since the bidding process opened up in July. In April, IBM and Oracle were eliminated from the race, while Google dropped out in late 2018 over concerns that it would be unable to fulfill the terms of the contract.
Microsoft, Amazon, Google, IBM, Oracle, are fighting for this big prize: the Pentagon contract. One name is strangely missing from the call: Facebook.
The main players in the cloud
In 2015, Amazon reported for the first time that AWS achieved $5 billion in annual sales and made it a division in its own right. Amazon's share price has since increased by a factor of 4.5. AWS now generates $36 billion in annual revenue, 13 years after its launch in 2006.
In 2014, Satya Nadella succeeded Steve Ballmer as head of Microsoft and changed the name from Windows Azure (created in 2008) to Microsoft Azure. The cloud is becoming the new center of gravity of computing and Microsoft accepts this fact: Windows is not the universe’s center anymore. Today more than half of Microsoft's cloud runs on Linux. Microsoft's share price has since tripled after stagnating throughout the Ballmer area.
Google only understood very late on the importance of the cloud, although it has the largest number of servers in the world. Its first cloud offer is launched in 2011. With an engineering culture, Google focuses on product perfection rather than sales. However, what works with the individual in the Internet area (spread by word of mouth), does not work with the company where there are more decision-making centres and friction: you need sales representatives. In February 2018, Google announced that its cloud business generated $4 billion in revenue, and in July 2019 it reached $8 billion. These last figures are to be taken with some caution because they include G Suite, which is largely dominant in Google's cloud offering. In November 2018, Google hired Thomas Kurian, an Oracle executive, to lead Google Cloud Platform and give it the sales force it needs.
The cloud concerns infrastructures (hardware), the programming platform (middleware) and programs (software). These three segments are fixed cost industries: a large initial expense (equipment, research) and then minor expenses (maintenance, improvement). The economies of scale obtained at the beginning are then very difficult to move, the positions are relatively fixed, to the advantage of AWS the pioneer largely in the lead then Microsoft (which for the moment does not reveal its turnover).
The size of the market
The market is huge. The digital economy is spreading like wildfire, applications are taking over in many areas (software is eating the world as Marc Andreessen prophesied in 2011). Professions related to the physical world (material processing, transport, trade) are made more productive and more sales-oriented by the integration of artificial intelligence. This artificial intelligence requires massive data (the differentiating factor) that is better processed in a cloud with the right CPU/GPU power.
To measure the size of the market, global IT expenditures ($5 trillion) can be used as a basis. This is already 7 times the global advertising spending, the target market for Google and Facebook. The cloud will "eat" IT because it makes very significant savings, on the one hand, and unleashes the power of coders, on the other:
On the first point, the example of Bank of America, explained by its CEO during its last conference call, is instructive: the introduction of a private cloud has made it possible to reduce the number of servers from 200,000 to 70,000, thanks to virtualization. The servers were previously only used at 1/3 capacity but were clustered by function. Most of its applications now run on 8,000 servers, so there is still progress to make. The bank thus saves $2 billion a year. The major interest of the cloud from a financial point of view is to adjust capacities as closely as possible.
the cloud allows the IT programmer to focus on his mission without worrying about the resources allocated to him, security, compatibility with the company's information system, etc. This release of the programmer corresponds to an ever-increasing demand for specific applications that respond to a function: we no longer buy a car, we use Uber, we no longer buy a parking space, we use Park Now, etc. With the spread of information, we directly seek to respond to the task at hand rather than protect ourselves by owning objects (obsolete information reservoir). It is therefore to be expected that the software share of the economy will increase significantly.
The potential market is therefore well above the current $5 trillion in IT spending. 10% of the world's GDP would not be impossible. That would be $8 trillion. With 20% of the world's GDP, the potential market would be $16 trillion. In other words, the places that are taken today are worth a lot of money because they will be almost impossible to move. We then understand the struggle to obtain the JEDI contract, even if, in the end, it is only a drop in the bucket compared to the potential market.
The strategies of the main actors
AWS: the undisputed leader with probably twice the turnover of the second Microsoft. AWS is committed to innovation in all three segments (infrastructure, platforms and applications), as well as economies of scale to drive down prices. The AWS service is not perfect but it anticipates the customer's needs at a reasonable price (always one step ahead). AWS' argument is that Amazon eats its own cooking: since AWS has solved complex problems for Amazon by allocating the right amount of resources, it also knows how to do so for large administrations, large companies and, a fortiori, SMBs.
Microsoft Azure also offers all three services. Its strategy is to leverage its enterprise customers, won by Office, Windows and its development platforms to deliver cloud services. By detaching the cloud from Windows, Microsoft gives it much greater potential because Linux is the language of servers, or mobility (4 times more smartphones than PCs). Microsoft is positioning itself as a friend of companies, with a mission to make them more productive, even if it is to the detriment of Windows.
Google came late to the market despite having the largest infrastructure and the best engineers. Its problem is primarily a commercial one. Its strategy is to highlight its containment capacity (through Kubernetes, an open source Google project). The official discourse is to allow companies to juggle from one cloud to another and not be locked in (in other words: AWS and Microsoft lock you in). The idea behind the container is to allow AWS and Microsoft customers to switch to Google cloud without any problem and thus hunt on their flower beds. It's a wise strategy for a number 3.
IBM missed the cloud. Being in a free cash flow logic, the company made little investment in infrastructure at a time when places were still to be taken. It still has the platform: it’s not the strong point of the first three that play a vertical integration logic. This explains the acquisition of RedHat for $34 billion (a small amount regarding the potential). RedHat has developed the Openshift technology which is the most important programming platform in the world. Basically, you program on Openshift, the latter takes care of the rest, the programs being contained and sent to any cloud. IBM is targeting the hybrid multicloud, i.e. the programming platform. The argument is to avoid putting yourself in the hands of a single supplier. The first move after the acquisition of RedHat was the merger with Cloud Foundry technology (the number 2 open source platform behind Openshift). The combination of the two undifferentiated Cloud Foundry technologies to the advantage of the largest Openshift.
Finally, Oracle is located upstream of the programming platform. It offers a database service in the cloud. From there, Oracle seeks to do vertical integration (programming platform and applications). The problem is that Oracle is leveraging a technology from the past, with document databases better suited to the amount of data to be processed in the cloud.
Why Facebook does not have a cloud offer
From IT news (July 28, 2018):
Facebook CEO Mark Zuckerberg has said the company has no plans to get into the public cloud business.
The prospect of a Facebook cloud emerged on the company’s Q2 2018 earnings call, during which the platform outlined full-year capex of US$15billion, dominated by data centres.
The company flagged even higher spend in future years.
Investment analyst Anthony DiClemente of Evercore Group quizzed Zuckerberg if Facebook had considered “ways to improve the return on investment of those investments in data centre servers, network infrastructure?”
“What I'm thinking is, in order to perhaps service third-parties to maybe just improve those returns, in the way that other tech and internet companies have in terms of investments in infrastructure,” DiClemente said.
The spin-out of a public cloud from a platform far from illogical given AWS was spun out of Amazon.com’s burgeoning in-house infrastructure.
Zuckerberg responded by saying: “I mean the quick answer is that we're not planning on going in to the cloud services. We're not planning on doing that. We have to build out all this capacity to serve our community. It's a very computationally and resource-intensive set of services that we provide and we need to build that out.
Marc Zuckerberg prefers to abandon a market estimated at between $8 and $16 Trillion to focus on a market of $800 billion. How is that possible?
The first explanation, given by Marc Zuckerberg, is that Facebook's core business requires a large amount of server capacity that the company prefers to reserve for itself rather than rent it to third parties. In this case, why does Google, which invests about the same amount in capacity, not hesitate to rent its servers to companies when it has needs as important as those of Facebook for its search engine and YouTube.
The second explanation is that Facebook, like Google, focused on its own capacity needs, has missed the cloud. Aware that the battle has already been won by Amazon and Microsoft, it prefers not to engage in a losing battle.
Finally, a third explanation, which is not incompatible with the second, is that Marc Zuckerberg has a plan to disrupt the cloud and that he prefers to invest his resources in it.
The Achilles' heel of the cloud
In a brillant post, Alex Danco shows how the Internet has moved from the property economy to the access economy. Information technology before the Internet was structured as a reflection of the property economy. Thus, files that are a pictorial representation of a real workbook with its different folders (hardware possession) give way to direct access to the desired function (complexity being hidden from the user). We no longer hold a cd, we use Spotify. The files that were the central part of the Windows operating system disappear with smartphones. It is more convenient to ask Siri or an app directly.
The Internet is the instrument of the fifth estate, as Marc Zuckerberg explains, because it makes abundant and accessible to all information that cemented the power of those who held it: in the past the church with the copyists, then the aristocracy and the bourgeoisie with the printing press, then commercial companies with the development of capitalism and joint stock companies
Everyone can now work in their own way with information and bring their own unique added value through a particular assembly of it. In the past, to design a computer program, you had to invest in servers, do research and marketing to sell your software. It was a strong barrier to entry. Today we use the public cloud by paying a small subscription, then a number of support functions still by subscription (Stripe for example for payments, Twilio for communication). Companies that used to live on an information monopoly can be harmed by a single creative individual relying on a cascade of third-party applications. Thus, an influencer may have a larger audience than a newspaper that once had a monopoly through the ownership of a printing press and now has to fight over the quality of its articles.
The power that previously rested in material possession is disintegrated by the Internet. The access economy is based on applications (smartphones) or "skills" (Alexa) running on the cloud and which are a form of sorting information answering a specific function. The object is usually multi-functional. The power of possession has been replaced by the efficiency of access.
With abundance, the added value lies in the sorting of information, the one who knows how to organize it has a piece of power. This explains the stock market value of subscription software companies with an interest in a single specific function (as of 27/10/2019):
These stock market values exceed those of many traditional companies selling possession. For example, Slack sells business communication (access) and is worth $12 billion; Avaya, the leader in the sale of business telecommunications systems (ownership), is worth only $1.35 billion.
In this access economy that is becoming the dominant model, no one owns anything and everyone becomes dependent (slave). So there is a real loss of power. If a company is built based on access to different functions by subscription (WordPress for the website, Okta for security, MongoDB for the database, Mailchimp for emails, etc.), it is essential to ensure that the company has access to all the functions required. it becomes dependent on the proper functioning of these applications. It no longer has control over the operations. Possession may be less effective but it helps to keep control. Finally, the fifth estate highlighted by Marc Zuckerberg in his recent speech at Georgetown University is only theoretical: power is actually even more focused on cloud providers who own the infrastructures, so Amazon and Microsoft in the first place. The Achilles' heel of the cloud is that it channels power instead of distributing it, and it is this flaw that Facebook wants to exploit.
The Libra plan
In my articles on subscription, I have discussed the theory of exclusivity:
The possession of an exclusive, rare property gives power and is a sign of that power. No one can encroach on this property or appropriate it. The ownership of an exclusive property distinguishes the individual from his or her peers. The will to power being intrinsic to man, it is not surprising to see this propensity to hold exclusivity. In contrast, what is common, abundant, common is only valued for its utilitarian aspect, so it can very well be rented or borrowed.
This law can be represented as follows:
In the cloud economy, developers create exclusivity but it doesn't really belong to them. They have to subscribe to many services and rely on an infrastructure that is not their own. Their power is limited to the proper functioning of the underlying infrastructure. This is a loss of power and value. To make a comparison, a mobile home is worth much less than a house because it is not attached to a plot of land. Scarcity is in the land, not the house. It's the same with cloud infrastructure. Facebook through the Libra project wants to give back ownership, therefore, power to developers. It is inspired by two crypto currencies that are not yet widely adopted. Facebook wants to perfect them and create an ecosystem of decentralized applications that really belong to their creator.
First of all, the Libra protocol is close to the Ethereum in that it institutes smart contracts, which are indelible computer code. A smart contract is like a notarial deed, set in stone. It is automatically activated in exchange for the crypto currencies paid to buy it. Libra is superior to Ethereum in two respects: on the one hand, smart contracts can be written simply and quickly thanks to the Move language invented by Facebook, more practical than Ethereum's Solidity language. On the other hand, Libra is not really a blockchain because the validation process is carried out trade after trade by a core of 100 validators. This process allows 1 000 operations per second to be processed instead of 20 for the Ethereum. The Libra was designed to be able to have a rapid scale effect: Facebook is in the real world.
Marc Zuckerberg has always dreamed of turning Facebook into an application platform (example of Zynga, the application of social games) but was overtaken by Android and IOS. Libra allows him to realize this dream by copying the Dapp Network. The Dapp Network is a network of applications (mainly games) developed on the blockchain. The interest for the developer is the real ownership y of his application and of the resources committed to develop it. Here is a comparison between the cloud internet and the Dapp Network internet:
The problem of the Dapp Network is liquidity: only 30,000 people use the network and the infrastructure resources available on the Blockchains are low. This is why the cost of development is prohibitive.
The Libra association could develop Dapps (which are smart contracts) based on the Libra protocol. Liquidity would be provided by Facebook (2.7 billion users) as well as the association's 100 members (Uber, Lyft, etc.). The infrastructures, in addition to the 100 members initially could come from Facebook, which has invested at least as much as Microsoft and Amazon.
For Facebook, this would be a way to break up the cloud:
Create an ecosystem of applications, middleware and decentralized infrastructures running on the Libra protocol. This ecosystem would give power back to the players (full ownership of their work), which is now first captured by Amazon and Microsoft but also by other leading players in the cloud.
Have this ecosystem managed by third parties (first by the members of the Libra association and then by anyone who meets Libra's ownership criteria (proof of stake).
Reduce the cost of developing applications thanks to infrastructures provided by Facebook infrastructures as well as other prominent founding members.
Bring a real added value to the ecosystem through a wallet (Calibra) of applications offered to its nearly 3 billion users.
Damn the pawn to Apple's App store and Alphabet's Google play Store.
Justify to the public authorities that Libra ( crypto currency) is just a point of passage for the acquisition of applications developed in Move language.
According to Marc Zuckerberg's leitmotif, give power back to the consumer/creator. Commercial companies are already dislocated by the 5th estate because they are losing the monopoly of information. Libra would reinforce this power by giving the creator a piece of property.
Finally, Marc Zuckerberg may not be so naive with his Libra project. And the defection of PayPal, Stripe, etc. is a good opportunity to refocus the project on smart contracts. The focus on the currency part of the Libra may even have been intentional: asking a lot to get a little, just the reversal of the most important market of the century, the cloud.