Blockchain for CPU? Analyzing Golem's Ethereum Token Sale
Alex Sunnarborg is the founder of Lawnmower, a blockchain investing and market data platform founded in 2015. Here Sunnarborg takes an in-depth look at the imminent crowdfunding effort by Golem, an ethereum-based market for computer processing power, which launches today.
This publication is not intended to constitute investment advice. Please consider all blockchain investments with caution.
Often described as “Airbnb for computers”, the Golem team intends to use the funding to build out a network for renting unused computing power (CPU or GPU cycles) amongst users, existing as decentralized, open-source software on the ethereum blockchain.
As a two-sided market, Golem allows users to either:
- Sell unused computational resources (CPU or GPU time) to others, or
- Buy unused computational resources (CPU or GPU time) from others.
As consumers' and businesses' requirements for processing power differ greatly, Golem looks to take advantage of the varying needs of end users and tasks to create an alternative cloud computing network and market.
As an example of the stark contrast in computational needs, imagine someone who uses their computer primarily for Facebook and email as compared to someone who frequently uses one for work or research related to resource-heavy fields and tasks like high definition image rendering, stock market simulations and back-testing, DNA analysis, machine learning or big data.
The Facebook and email user likely uses their computer’s CPU or GPU to nowhere near their capacity, and could use Golem to sell processing time directly to a user who likely often needs to shop around for additional computational resources during periods of heavy analysis, and currently may turn to services and corporate giants like Amazon, Google, Microsoft and IBM.
The marketplace is anticipated to include individuals, organizations, data centers, and more, and allow for complete flexibility and customization for both buyers and sellers on terms including price, time or task, and amount or percent of resources.
The upcoming crowdfunding is for a token on the ethereum platform called the Golem Network Token (GNT), intended to be core to Golem and “ensure flexibility and control over the future evolution of the project and attributed to a variety of functions in the Golem network following the first major release” (including being the only type of payment permitted for buying computing resources).
The crowdfunding for GNT will last three weeks, or until 820,000 Ether (ETH) is raised. If less than 150,000 ETH are raised in three weeks, the crowdfunding will be deemed a failure and the ETH will be returned to the original owners.
Given the popularity, consumer demand, and speed of successful recent public blockchain asset crowdfunding efforts (especially on ethereum), it’s an interesting time to look at some of the similarities and differences in the structuring of the assets in this new trend or alternative fundraising paradigm.
Instead of birthing a new asset and blockchain with an initial genesis block and the subsequent start of competitive mining like with bitcoin (BTC), litecoin (LTC), or more recently, Zcash (ZEC), 100% of the supply of GNT to ever exist will be allocated and distributed immediately following the closing of the crowdfunding – similar to Digix DAO (DGD), SingularDTV (SNGLS), or First Blood (1SŦ).
As ethereum-based blockchain assets with public crowdfunding projects this year, the demand for these tokens are extremely relevant data points. Further, DGD, SNGLS, and 1SŦ had maximum fundraising caps similar to GNT:
- In March, DGD received $5.5 million of ETH in 12 hours
- In September, 1SŦ received $5.5 million of ETH in 5 minutes
- In October, SNGLS received $7.5 million of ETH in 15 minutes
- In November, GNT will attempt to receive $10 million of ETH.
Given the massive demand and blazing speed for these crowdfunding efforts, GNT may similarly see enormous initial interest and an extremely small time window in which Ether deposits will be accepted in exchange for an initial GNT allocation.
For some additional data points, many ethereum based projects have alternatively opted to raise funds with no concept of a maximum cap:
- In late 2014, ethereum (ETH) itself raised $18.4 million in 41 days
- In late 2015, Augur (REP) raised $5.3 million in 45 days
- In late 2016, Iconomi (ICN) raised $10.6 million in 35 days.
Further crowdfunding attempts in the blockchain space are not only carried out via ethereum. Both Lisk (LSK) and Waves (WAVES) actually compete with ethereum as decentralized app platforms upon which developers can create and deploy software, and each also had their own uncapped public crowdsale earlier this year.
- In early 2016, Lisk (LSK) raised $5.7 million in 28 days
- In mid 2016, Waves (WAVES) raised $16.1 million in 50 days.
During Golem’s crowdfunding, GNT will be available for purchase at an exchange rate of 1,000 GNT for 1 ETH. Given the 820,000 ETH cap, 820 million GNT are available for participants, which will be available for trading immediately after the end of the crowdsale.
This straightforward approach contrasts somewhat with the more creative structure we saw recently in the 1SŦ crowdsale and its moving pricing that awarded early participation with a better exchange rate that may have heavily influenced the extremely early demand and closing. Also of note, 1SŦ were set to be non-transferable for two months following the crowdfunding, delaying attempts to quickly sell the asset after initially buying in.
The GNT generated by ETH sent during the crowdfunding process are set to represent 82% of the overall supply of GNT. Regardless of the level of funding, after the crowdfunding period has ended, additional GNT shall be created and distributed such that 12% of the overall supply will be allocated to Golem Factory GmbH (the company behind Golem, intended to be incorporated in Switzerland), and 6% to team members and early contributors.
Unlike the GNT allocated to crowdfunding participants upon the closing of the sale, the 18% of GNT pre-allocated to Golem, team members, and early contributors is locked (non-transferable) for six months after the creation period ends. Thus, initial selling volume on exchanges will be at the hands of crowdsale participants only.
Along with the raised ETH, the 12% of GNT created allocated to the Factory (a maximum of 120 million GNT, worth 120,000 ETH or about $1.3 million) is intended to supply the company with the funding to develop and execute the overall Golem project, with a slightly flexible budget structure, go-to-market strategy, and development roadmap and timeline depending on the success of the fundraising.
The Golem project is generally designed around four main phases, the “Brass”, “Clay”, “Stone”, and “Iron” Golems – all methodically building upon each other and the overall vision.
Golem has also supplied some detail of their actual expected expenses given the minimum (150,000 ETH) and maximum (820,000 ETH) crowdfunding results, stating that if the cap is reached, they “will be able to finance team of 20 people (most of them developers) for a period of 4 years” (with offices in Zug, Switzerland and Warsaw, Poland).
Given 820,000 raised ETH plus 120 million GNT to the company (worth about $11.3 million combined at current prices), the team of 20 would be allocated 54%, or $6.1 million, of which each would receive $305,000 on average, or about $76,000 per year. However, Golem’s plan to sell or hold ETH or GNT over the life of the project, and questions around the timing and assets they would sell into are currently unknown.
The team behind Golem consists of the founders of imapp (in Poland), a CEO, CTO, and COO (also founders) with masters degrees in economics, math, and computer science from the Warsaw School of Economics and the University of Warsaw, and numerous additional developers from Poland with deep experience in technology and working with ethereum.
The value behind GNT is based on a wide variety of factors, and of course, as Golem clearly sets out, no user should be “submitting ETH to obtain GNT for the purpose of speculative investment”, and rather “is purchasing the functionality of the GNT issued primarily to support the development, testing, deployment and operation of the project”.
Given the inherent opaqueness of the conservative legalese, Golem published a recent blog post entitled “The Economics of the Golem Network Token”. As stated previously, GNT will be necessary to interact with Golem, initially to enable the transfer of value from the buyers of unused processing time to the providers and to software developers, and later (in the Stone and Iron phases) “to assign additional attributes to the token, so that, for example, it is required to store deposits in GNT”.
The Golem team cites the “quantity theory of money”, stating “the general price level of goods and services is directly proportional to the amount of money in circulation”, and since GNT is fixed in volume after the crowdfunding, a growing demand for GNT should increase the number of transactions on the network, and thus the price of the asset over time.
For some more quantitative and data driven metrics, they turn to market size numbers for Software as a Service (SaaS), Infrastructure as a Service (IaaS), and Platform as a Service (PaaS) within the cloud computing sector, stating a combined value of $64 billion ($37.7bn for SaaS on the cloud and $27bn for IaaS and PaaS on the cloud).
As Golem wishes to “create an almost perfectly competitive market for computing and software”, it believes it can take a piece of this pie, starting with “low hanging fruits” like CGI rendering (of which they estimate public cloud rendering as a $200m market), before tackling other promising computing-intensive applications like machine learning and scientific computing.
They continue on to claim these markets are however only the beginning, and “what is really important down the road, are microservices and the ‘true cloud’ functionality, which will allow Golem to offer features similar to what is offered by public cloud services today”.
The Golem team is not shy about its challenges ahead either, and has dedicated another entire blog post and section of their GNT terms to their risks and hurdles as well, including creating a P2P network, defining and verifying tasks, creating a unique transaction system, building a reputation system, handling running untrusted code, potential overall Golem and ethereum software weaknesses, regulatory uncertainty, and more.
Additional questions and concerns I would bring up include ideas such as:
- What happens if the company needs additional funding? As the GNT supply will exist in full after the initial crowdfunding, would Golem try to secure further funding in an alternative fashion like raising money for corporate debt or equity?
- What will the company do with the initially allocated GNT after the six-month lock up? Is there some plan or methodology to evaluate how to allocate the corporate funds and overall portfolio, for example selling GNT over time into another blockchain asset or local sovereign currency?
- Are there any further numbers on relevant market sizes, the current pricing landscape for traditional cloud computing providers like Amazon and Microsoft, comparisons to Golem expectations, or any quantitative projections on early usage, growth, or otherwise?
Computer processor image via Shutterstock
Disclaimer: The views expressed in this article are those of the author and do not necessarily represent the views of, and should not be attributed to, CoinDesk.
Deutsche Bank: Capital Markets Expect Blockchain Impact...
Blockchain Startup Adds UK Cybersecurity Expert to Board