The Web3 Cloud Paradigm Beyond the Decentralisation Paradox
This article explores the Web3 cloud paradigm beyond the decentralisation paradox alone and describes why Web3 companies appear to be reliant on centralised cloud providers and may remain to be so for the time to come.
As a part of a potential solution, the article suggests that Web3 start-ups, mid-sized projects and even billion-dollar companies may restructure their cloud needs through new Web3-based solutions so to further decentralize their infrastructures[1]. Ultimately, these enterprises may save both their (decentralised) faces and their valuable resources.
Introduction
The importance of Web3 and blockchain technology for decentralising control over economies and societies have been readily described. To say the very least, Web3 projects are set on a path to creatively disrupt many aspects of our lives.
A16z argued that where an industry is fueled by private capital and margins are of secondary concern, new projects tend to start (and remain) in the cloud. In these fast-moving industries, start-ups prioritize immediate development over longer-term cloud cost efficiencies and cost savings[2]. A16z coined this concept the ‘cost of cloud paradox’.
Web3 companies appear to be falling for that very same cloud seduction and may continue to be temped for the time to come.
Web3’s Unquenchable Thirst for Cloud
Crypto winter or not, the Web3 industry is roaring and building. Private capital investment in blockchain and Web3 has been growing exponentially[3]. While developing and evolving, projects heavily focus on attracting new human resources and ultimately developing bound-breaking products and services. Meanwhile, margins and profitability are not of immediate concern to new projects. Indeed, Web3 start-ups need time to mature.
These circumstances dictate that Web3 companies run their business by using cloud providers — a sector that is moreover growing at exponential levels as well[4].
First, the rapid expansion and evolution of the Web3 market[5] will likely lead to increased demand for cloud solutions to store all new data that is generated.
For instance, the explosion of alternative layer-1 networks[6] indicates that the newer blockchain entities are producing great amounts of data compared to the first generation of blockchains. For instance, Algorand’s entire blockchain history is appraised at almost 2 TB (split between main, beta, and testnet amassed over two years) while Bitcoin’s 375 GB is bundled over eleven years.
Another example is Solana. Solana’s estimated data requirement is sitting at around a petabyte (1000 TB), which makes it rather unviable[7] to expect that nodes will store this data locally.
Moreover, in the near term, a myriad of layer-1s focused on scalability and speed will continue to emerge, expecting to be generating ‘disproportionate amounts of data’ as a function of improved block space and transaction throughput[8].
The continued exponential increase of blockchain infrastructures will require new storage solutions to store both live and historical data. Choosing reliable, convenient hyperscalers may be the obvious one.
Decentralised but Dependent
Secondly, new cloud infrastructure is necessary to avoid current and future Web3 applications being overly dependent on centralised services.
By nature, Web3 projects seek to be decentralized, they evolve rapidly and are globally distributed. But paradoxically, those same Web3 companies often rely on a handful of centralised companies to help roll out Web3 products and services globally.
Indeed, unlike the name suggests, decentralized protocols may be centralized in where their nodes are validated[9]. This is problematic from an ideological point of view as Web3 ought to be characterized by ‘less trust and more truth’[10]. Rather, nearly 70 percent of Ethereum nodes worldwide run on hosted services, of which over half are on AWS[11].
Besides ideological matters, dependence on a single cloud service provider may bring financial issues, either in terms of legal damage or in the light of potential damage to the network’s reputation.
Cloud insurer data[12] shows that during 2021, overall, the major cloud providers experienced an outage of at least thirty minutes — every three weeks. Such structural issues are problematic certainly if companies (either centralized entities or decentralized applications) heavily rely on a single service and cannot (optimally) operate during server downtime.
This dependency could also generate financial risks. Robinhood was fined almost $60 million and was ordered to pay $13 million more in restitution to thousands of clients due to outages[13]. Binance stands to face legal claims over outages, with users seeking international arbitration[14].
Crucially in the light of Web3, such outages on crypto platforms also undermine trust and prevent growth[15] even before (trustless) technology can do its job. Indeed, ‘can’t be evil > don’t be evil’[16], yet the question is to which extent decentralization is a reality when its underlying cloud is not.
The Cloud-Paradox Argument
However, the choice of Web3 for (hyperscaler) centralized cloud providers appears to be inherently intertwined with the very cost of cloud. Indeed, a third reason why new solutions are needed is that the very cost of cloud and the dominant position of big providers can be labelled as a market failure.
Indeed, what has been coined the ‘cloud-paradox’ for SaaS companies holds that although choosing for cloud infrastructure is cheap early on, it becomes a lot more costly down the road[17] — certainly when looking at cloud costs relating to the total revenue generated[18].
It is to be expected that this applies vis-à-vis to Web3 as well. As shown, blockchains specifically need a lot of aggregate computing power to support the nodes, making computing power costs by its very nature one of the critical costs that maturing Web3 companies experience.
Moreover, the well-known cloud-services providers own 75 percent of the cloud market[19], while cloud adoption is still growing rapidly[20]. That makes them hyper-dominant too.
Crucially, the hyperscalers, representing a combined five trillion-dollar market cap, are buffeted from the competition. Hyperscalers have high-profit margins[21], driven in part by running their own infrastructure, enabling ever greater reinvestment into product and talent, while their share prices are boosted[22].
The hyperscaler dominance may cause unease for blockchain and crypto enterprises. Web3’s ideology and raison d’etre may stand at unease with the Amazons and Googles as the centralized and powerful entities of the Web2 era.
Last but not least, the hyperscalers’ hyper-profits will be mirrored on Web3’s balance sheet — if not today then tomorrow once the project matures and requires more computing power.
A Paradigm Shift?
Both Web3 and existing cloud providers will need (to offer) new cloud server infrastructure solutions for several main reasons.
The rapidly evolving Web3 industry needs new cloud solutions to meet its increasing demand. Decentralized protocols may be centralized in where their nodes are validated, which is problematic from an ideological point of view.
Besides, dependence on a single cloud service provider may bring financial issues and it can undermine trust. And although cloud is cheaper early on, it becomes a lot more costly down the road. The lack of competition and high costs are market inefficiencies looking to be solved.
Yet despite the ample room to create new cloud solutions for a new market, there remains a mismatch between the needs and wants of Web3 on the one hand and those of existing cloud providers on the other.
Existing cloud providers that challenge the hyperscalers are experiencing difficulty in fully accessing the multi-trillion dollar Web3 market[23] due to the lacking infrastructure that connects them to Web3.
Web3 companies, on the other hand, may be located either offshore or there where they do not have access to traditional financial services needed to ‘plug and play’ with the cloud hyperscalers. Importantly, Web3 companies may also simply wish to pay exclusively with cryptocurrencies, which makes onboarding them rather tricky, if not impossible.
Web3 as the Solution to Web3
Dr. Gavin Wood, one of the creators of the Ethereum blockchain, famously defined Web3 as ‘less trust, more truth’, implying that trust in the intentions of others and the centralization of power due to big data can be replaced with transparency and irrevocability built into blockchain technology.
In the spirit of Web3, it might be worth considering Web3-based alternatives to counter the centralised power of existing cloud providers
Indeed, Web3 companies are jumping in to decrypt the mismatch, either by bridging it or by offering completely new solutions. Projects like ICP, IPFS, Swarm and Arweave are building (partial) solutions to the mismatch.
Time will tell whether Web3 start-ups, mid-sized projects and even billion-dollar companies will be able to restructure their cloud needs so to further decentralize their infrastructures. They may surely consider Web3-based alternatives to ultimately save both their (decentralised) faces and their valuable resources.
Connect with the author on LinkedIn: https://www.linkedin.com/in/alexandra-overgaag/
— — — — — — — — — — — — — — — — — — — — — — — — —-
[1] This article is meant for educational purposes and does not offer investment advice.
[2] a16z, Sarah Wang & Martin Casado, The Cost of Cloud, a Trillion Dollar Paradox, see, https:// a16z.com/2021/05/27/cost-of-cloud-paradox-market-cap-cloud-lifecycle-scale-growth-repatriation-optimization/.
[3] CBInsights (2022). State of Blockchain, p. 11. Global blockchain funding surged over 700% year-on-year to reach over $25 billion in 2021.
[4] The cloud infrastructure services market is growing in a nearly exponential manner, at 34% per year. See Synergy Research Group, Huge Cloud Market Still Growing at 34% Per Year; Amazon, Microsoft & Google Now Account for 65% of the Total, see,
https://www.srgresearch.com/articles/huge-cloud-market-is-still-growing-at-34-per-year-amazon-Microsoft-and-google-now-account-for-65-of-all-cloud-revenues.
[5] As mentioned in Pantera Capital (2022), Blockchain Letter, The Next Mega-Trade, see, https://panteracapital.com/blockchain-letter/the-next-mega-trade/, the number of people using crypto is closely following that of the Internet in the 1990s — eight years for each to hit 100 million users. The development of the internet serves as a narrow analogy for the development of Web3.
[6] 21Shares (2022). The State of Crypto, p. 21. Layer-1 blockchains specialize both in settling transactions and launching Web3 or decentralized applications. In this layer, blockchains also provide immutable data hosted on each network’s server or node. Most layer-1s rely on variants of Proof of Stake mechanisms for settlement and network security purposes.
[7] Infra 12, p. 60.
[8] Ibid., p. 64.
[9] For instance, in September 2021 Coin Rivet revealed data from the Block Logic‘s validator app that one single Amazon data center server accounted for 37.45% of active Solana stake, making this AWS server to have a ‘kill-switch stake’ in the network. See, https://coinrivet.com/it/amazon-hosts-37-of-Actively-staked-sol-could-this-be-a-solana-kill-switch/.
[10] Dr. Gavin Wood, one of the creators of the Ethereum blockchain, famously defined Web3 as ‘less trust, more truth’, implying that trust in the intentions of others and the centralization of power due to big data can be replaced with transparency and irrevocability built into blockchain technology, see The Father of Web3 Wants You to Trust Less, https://www.wired.com/story/web3-gavin-wood-interview/. Web3, in the context of Ethereum, refers to decentralized apps that run on the blockchain and that allow anyone to participate in the network without monetising their personal data, see Web2 vs Web3, https://ethereum.org/en/developers/docs/web2-vs-web3/.
[11] Ethereum Mainnet Statistics, see https://ethernodes.org/networkType/Hosting
[12] Parametrix Insurance (2022), 2022 Crypto Confidence Report: Building Trust to Drive Growth.
[13] Robinhood to pay $70 million for outages and misleading customers, the largest-ever FINRA penalty, see, https://www.cnbc.com/2021/06/30/robinhood-to-pay-70-million-for-misleading-customers-and-outages-the-largest-finra-penalty-ever.html.
[14] Binance is facing an unconventional new legal claim over May 19th outage, see, https://www.theverge.com/2021/8/19/22631213/binance-lawsuit-downtime-may-outage-options-leverage-bitcoin
[15] Infra 7, most crypto users have experienced an outage on a platform they use which inhibited access to their funds, while a majority of users only somewhat trust crypto platforms, exchanges and wallets. This lack of trust is preventing growth and is, paradoxically, keeping new audiences away from the fundamentally trustless technology that blockchain is.
[16] Chris Dixon mentioned that ‘the only API you should trust is one controlled by a blockchain. Can’t be evil > don’t be evil’, see, https://twitter.com/cdixon/status/1461723922336915459.
[17] Infra 1.
[18] Ibid. The cost of cloud contributes heavily to the total cost of revenue (COR); based on benchmarking public software companies, a16z found that contractually committed spend on cloud averaged 50% of COR while, crucially, actual spend as a percentage of COR is typically even higher since companies are often conservative when sizing cloud commit size. Crucially, companies are often conservative due to fears of being overcommitted on spend, hence it has been argued that committed spend is typically ~20% lower than actual spend, ‘elasticity working in both ways’. Flexibility of cloud products that fit a company’s need is hence key.
[19] Being Amazon.com Inc., Microsoft Corp., Alphabet Inc. and Alibaba Group Holding Ltd, according to, Gartner Says Worldwide IaaS Public Cloud Services Market Grew 40.7% in 2020, see https://www.gartner.com/en/newsroom/press-releases/2021-06-28-gartner-says-worldwide-iaas-public-cloud-services-market-grew-40- 7-percent-in-2020.
[20] Deloitte Insights (2022). The Future of Cloud-Enabled Work Infrastructure, p. 3 and infra 3.
[21] AWS’ operating margin in percentage of net sales is currently sitting at 35.3%, see https://ir.aboutamazon.com/news-release/news-release-details/2022/Amazon.com-Announces-First-Quarter-Results-f0188db95/.
[22] Infra 2.
[23] See for the latest figures, https://coinmarketcap.com.
[24] Infra 1. And to any other asset connected to the internet, such as API servers, message ques, Proof-of-Work miners, IoT gateways and devices, file servers, databases and so on.
