Fetch ICO review
by stsoen, Irene Adler, in co-operation with icoventures.net
This review is written exclusively for https://t.me/KryptoSherlock
Official website: https://fetch.ai/
MVP: absent
Team: 33 members + 5 advisors
Brief description: an “intelligent distributed system”
Token ticker/type: FET, ERC-20 that will be converted to the native tokens upon the main net launch in 2019.
Tokenomics and distribution metrics:
Here is the current metrics published on Binance:
The total hard cap of the circulating supply is $60mil, including the offering on Binance. The project has attracted around $16,5mil so far. In comparison with the previous metrics, the overall amount of tokens hasn’t changed but the hard cap has: it has undergone a downsize from “no greater than $100mln” to the current figure as you can see in the image below
Overall, the price of a single FET, with the initial $30mil hard cap in mind, can be calculated as follows: 20% of the tokens equal to 230,599,516 FET with comes down to $0.13/FET bringing the valuation of the project, including the total supply, to $150mil. The Binance offering is listed for $0.05267/FET. It seems that the story of BTT repeats itself here.
The token allocation scheme seemed to be done with not much thought in mind:
Token unlock schedule:
50% of the team and advisors’ tokens are unlocked on year 1 with the release of 25% annually in the next 2 years:
The Foundation tokens are locked for 3 years, the private sale gets a 3-month lock with a gradual release within 3 months after the TGE. All the unsold tokens are locked for 12 months and will be put back in the circulation within 24 months after the TGE.
Social Networks:
as of 02/04/2019:
- Telegram – 7903 members
- Twitter – 2838 subs and 154 likes. The account was created in 2017 but the first tweet is dated January 24th, 2018. There is a compulsory picture with CZ, of course:-)
- Medium – articles have started to appear since December 2018
- YouTube – the first video was published on January 30th, 2018 but there are only 406 subscribers despite an abundance of videos featuring team members
Investors:
Outlier Ventures – young UK-based investment fund that falls closer to a typical VC-fund with an aggressive strategy rather than your typical crypto fund. A rather modest, and somewhat picky, portfolio looks good apart from the fact that technically speaking, they “invested twice” in the team of Fetch (read about it below).
Spark Digital Capital – Asian crypto fund. Seems that only the most successful investments are published on the official website.
Legal Aspects:
Here we can see a textbook separation between the project itself and the Foundation:
FETCH.AI LIMITED Company number 11203766 – registered on February 13th, 2018 in Suffolk, UK with the capital of 100GBP. The board of directors includes Humayun Sheikh, CEO; Toby Simpson, CTO, and Thomas Hain, CSO.
Location: Unit 2 Brickfields Business Park Old Stowmarket Road, Woolpit, Bury St. Edmunds, Suffolk, United Kingdom, IP30 9QS. Right next to it, another company belonging to the CEO of Fetch, ATL Steel LTD, is located. Briefly speaking, it is an industrial neighborhood and ATL Steel does mining, but not the crypto kind as we are used to by now. They mine ores and minerals in quarries. Here is a list of companies with the same CEO. Some of them have a different address but a similar board of directors:
- ATL Steel Limited – registered 23.10.2001
- Gloxal LTD – registered 5.05.2011, dissolved 05.08.2014
- Mettalis Limited – registered 11.06.2012
- EPS Materials Recovery LTD – registered 28.01.2013
- Mettalist Recycling Limited – registered 28.11.2013
- Uvue LTD – registered 24.07.2018, which its directly correlated to Fetch
- Mettalis Investments LTD – registered 1.03.2016
- Metalis Holdings Limited – registered 12.04.2016
- Mettalex Limited – registered 22.07.2016
- NOVUS4 LTD – registered 8.08.2016
The Fetch office is located in Cambridge, same place as UVUE.
The Foundation, FETCH.AI FOUNDATION PTE. LTD. was registered in Singapore on March 28th, 2018. No application to SEC has been filed so far.
The website gets around 500 visitors a day with a bit over than 2000 views. The figure has increased after the news about Binance in December. There is a nice geographical spread instead of your usual visitors from CIS and Asia generating the traffic. The website currently stands at 2080 visitors a day or 62400 a month.
Whois is hidden and the website is hosted by GoDaddy.
GitHub:
The official GitHub https://github.com/fetchai contains 5 repositories with more than 400000 lines of code written. 4 out of 5 repos have been updated recently which cannot be said of the Ledger repo https://github.com/fetchai/ledger with the latest update being pushed on November 27th, 2018. Generally, all the code is written in C++ with some Python here and there. The volume of the code is not very large but it’s sufficient and relevant. The quality of the code is not bad but is sporadic in certain spots:
The history of the development is hidden in the private repos and the team seems to push the public updates in batches. It can be seen as a negative sign as there are no public commits and it is impossible to see who actually is responsible for coding as there are no public members on GitHub.
Technology:
Full Whitepaper: https://www.fetch.ai/uploads/technical-introduction.pdf
There is no point to fully describe another blockchain with a fairly typical token usage. Here are some bullet points of Fetch:
Main technological features:
- Hybrid system with a blockchain and a DAG – for each transaction 2 separate chains are created in order to tackle the system throughput. The nodes interact via the DAG with the sharding feature coming in the future;
- uPoW (Useful Proof of Work) consensus algorithm – all the computational tasks are broken into the smaller packets and could be executed by the nodes of various computational power, i.e. “weaker” nodes can be in charge of recording the transaction on the blockchain while the “stronger” ones could do the actual heavy lifting. Frankly speaking, the description of the mechanism is quite vague.
- Throughput is promised to be around 1 000 000 TPS. There are no technical details though;
- Usage of AI for analysis and to predict the load of the network. AI algorithms are also used to validate the work of the agent nodes.
- Interface – natural language processing.
To conclude, here we have just another blockchain with an attempt to merge it with AI and machine learning. Some projects, like Dfinity, use pseudo-AI with the clear functions, i.e. self-learning, control over the network, patching on the go and interference in some complex situations. Here AI and ML seem to be capable of doing literally everything. In fact, the blockchain is only needed for payments, miner rewards and, in theory, to use in the supply chain. The network is comprised of so-called “agents” that perform certain actions in exchange for a reward. Thus, we see another “all-in-one” project with no measurable and specific targets in mind and a lot of vagueness in the documentation.
The goals set for the project are very ambitious and require a lot of effort to come to fruition. Considering the quality of the technical documentation and the use of methods that are still being developed and tested, the practical realization of Fetch definitely leaves us in a state of doubt. In addition, the team has never released its Security White Paper that was supposed to be out in Q2 2018.
The Team:
The main interest for us lies in the personality of the CEO Humayun Munir Sheikh LinkedIn Twitter
Mr. Sheikh’s first upper management position workplace was with Ososim: a learning technology company that performs simulations of various business processes. Later, he moved on to found Gloxal LTD which was the start of various ventures in the mining industry we mentioned above.
His first IT start-up, ItzMe, was founded in 2015 together with Thomas Hain. The company’s LinkedIn states the use of machine learning and AI in order to easier interact with devices. The website is now defunct and all the social network accounts have been removed.
The most interesting venture is, without a doubt, Uvue LTD. Created in 2015, it has the same registration address and the office location as Fetch.ai. The industry is also similar to Fetch, but here they were trying to create a transportation network for drones and unmanned flying vehicles. The company also attracted investment from Outlier Ventures. Just like in Fetch, the team seemed to want to hop on a blockchain hype train with smart contracts built into Uvue. There is no information on whether the project was successful, but the website states that the team managed to map 600 hectares of fields in the UK. The main website still displays a hiring ad looking for low-level developers.
Looking into financials of Uvue we can see that the company was pretty much unprofitable, and 2017 marked a loss of almost 700,000 GBP. We could speculate that this money was taken out of Uvue to promote Fetch.
Another venture where Mr. Sheikh was a member of the board, NOVUS4, also showed a significant amount of debt as seen in the picture below. However, it seems like by the time Mr.Sheikh joined the company in 2016, the ship had already been sinking for some time. Total losses have increased from 142,000 GBP in 2015 to 808,000 GBP in 2018.
Overall, there is a pattern in all the companies here: most of them have a large revenue, and 3 companies are even profitable with about 1,5mil GBP stocked. The only exception is METTALIS RECYCLING LIMITED that has an astonishing debt of 3,2 mil GBP. Mr. Sheikh, however, was just one of many members of the board and the losses cannot be attributed directly to his managerial abilities. He left the company in 2015 and Mr. Jason Mark Coleman took over the duties with the company continuing to spiral downwards that brought it to the final debt of 3,2mil GBP which included a payment negotiated with one of the creditors as seen in the screenshot below:
We asked Mr.Sheikh directly some of these questions, and this was his reply:
Keep in mind, that we are still waiting for a promised letter from accountants.
Toby William Simpson, CTO LinkedIn GitHub
He has also worked in UVUE and NOVUS4, as well as in Ososim. We would like to note his experience in machine learning during his work in DeepMind: a UK-based company. There is also a video of his talk with Donald Clark: one of the leading experts in AI. A GitHub account was created in 2011 but it is empty.
Thomas Hain, CSO LinkedInGitHubTwitter
Mr. Hain was also a member of the board in NOVUS4 and, together with Mr.Sheikh, a co-founer of late ItzMe. He has more that 15 years of experience at The University of Sheffield. Prior to that, he was a professor at Nagoya Institute of Technology, and a co-founder and a director of Koemei: a company developing a search algorithm for video files. Mr. Hain published many articles on speech recognition and machine learning. His Twitter has no mentions of Fetch and the GitHub is empty.
Jonathan Ward, head of research LinkedIn
Together with Thomas Hain, he developed a Ph.D. studentship course on deep learning of mosquito immunity where Fetch is actively promoted. Having a Master’s in Mathematical Biology and a Ph.D. in computer science makes the topic of the course is more than appropriate.
Troels F. Rønnow, Head of Software Engineering LinkedIn;GitHub;Twitter;Personal Page
A jack of all trades: from solving complex mathematical problems to researching quantum annealing. Before Fetch, Troels worked at Nokia on implementing blockchains in various industries, including healthcare. It is a bit odd, that the Github of such an experienced software engineer is practically empty.
Arthur Meadows – marketing director LinkedIn;Personal Page Worked in a marketing team at UVUE and ItzMe. Even though the experience seems to be vast, Arthur hasn’t spent more than 1.5yrs working for a single company. It seems that he is not directly employed by Cambridge but acts on more of a freelance position at Cambridge Product Management Network.
Ed Fitzgerald, Lead Software Engineer, almost zero commits and some forks of other users’ repos:
These are the GitHubs of a leadership team. 6 out of 10 coders have GitHubs in the same state of despair. It seems that only 2 team members do all the heavy lifting at Fetch:
Nathan Hutton Linkedin GitHub Twitter
GitHub is moderately active within the last 2 years. Not very active online and not much experience. Other social networks are nowhere to be found.
Atilla Bagolu LinkedIn Facebook Twitter Github
Good LinkedIn with a lot of endorsements. All the information correlates to FB posts. Twitter only has 8 posts with the latest ones about Fetch and Binance. GitHub is moderately active within the private repos. His personal page redirects to an ML course.
Advisors:
Melvyn Weeks: a lecturer at Cambridge who specialized in Microecenometrics. He has a good trail online that includes a number of publications in scientific journals.
Monique Gangloff: immunologist and biochemist at Cambridge. She has a large number of publications and a patent in her name. However, we’re not sure about the degree of her direct involvement in Fetch.
Dr. Niall Armes: a well-known and renowned specialist within the field of molecular biology. Just like the previous adviser, Dr. Armes has a number of patents in his name. However, the only connection to Fetch that we are able to see is that, in theory, AI could be used for DNA modeling to help us understand the human body better.
Steve Grand – a roboticist with a lot of experience in the field. In 2001, he led a team that created an artificial robot baby with an idea that it could learn as a human child would. He documented his journey in the book “Growing Up with Lucy”.
Jamie Burke – represents Outlier Ventures, advisor to Seed Token.
Partners and Investors:
MOBI(former Blockchain Mobility Consortium): Fetch.AI is listed on their website and there are re-tweets from the official account. One of the tweets links to an article that points out that Fetch and Bosch have joined efforts to work on creating an autonomous shuttle bus. However, both Fetch and MOBI are not directly named in the article itself. There are a number of articles that Fetch has joined the “IoT-alliance” together with IBM, BMW, General Motors, and XAIN, that work on blockchain solutions for Porsche. It is a bit strange to not see Toyota in the list because the Toyota Research Institurte created Blockchain Mobility Consortium that later evolved into MOBI. MOBI’s social presence could have been developed better though. The main social network accounts only appeared in spring, 2018 and there are some mentions about partnerships with Hyperledger, IOTA and NuCypher.
Blockchain for Europe – contrary to its nice-sounding name, this is not a government-backed organization (EUBlockchain is). It is a consortium of Ripple, NEW, that is experiencing some financial trouble, and Emurgo, a company behind Cardano. They joined Fetch at a later stage.
Warwick Business School – the business school itself is not partnered with Fetch. However, it appears to be a collaboration between several projects that work on improving AI in various fields, i.e. medicine, finance, and urban infrastructure. Nevertheless, a very useful partnership to have for a practical implementation of Fetch.
Uledger – gathered around $500K during its ICO. This start-up has since partnered with its home city of Boise, Idaho to bring blockchain technology into the government data. The project itself doesn’t look bad and the team at Fetch describe them as a company that will help securely collect users’ data from a number of gadgets. In this case, why to have an ICO to build your own blockchain if you have to rely on a 3rd party solution?
Trusted IoT Alliance– another consortium with a “pay-to-join” business model. For $2.5 – $20k a year your logo can be there too. It explains the presence of Monetha and Atonomi as the members of the alliance.
Conclusion:
Not very well thought out tokenomics and a relatively quick unlock of the team tokens in conjunction with a significant financial debt left from the Fetch’s previous ventures, makes you think twice about whether the team will able to practically implement their ambitious goals. The description of the technology is vague and hard to comprehend for a regular user with no scientific background. The reality of our market is that Binance is the only thing that can pull this project to its goal. Despite having an armada of respected advisors, their role and involvement in the projects are fairly questionable with many of them having no experience in AI or blockchain. Most of the GitHub accounts are empty with no publicly available information about the developers and their coding skills. It is true that top developers in the giant IT-corporations prefer to keep their accounts private but this rule shouldn’t apply to every single member of this team. Keep in mind that it is not the team’s first attempt to connect blockchain to a product. It is also unclear why an investment fund decided to put their money in the same team that had already failed in the previous project.
But, let’s face it, you are here to see how many multipliers Fetch will give you upon their (possible) listing on Binance. Our prediction is that there should be some. It is in the best interest of all the parties involved: CZ needs to keep the BNB hype train going, while Fetch will probably cover their debts from the previous ventures as a relatively fast unlock of a large portion of the team tokens tells us. Will it reach the height of BTT/Tron moneymaking machine? Probably not. Everything is in moderation. Especially, in this market.
Final mark: 7/10 (due to Binance mostly)
Subscribe to our channel https://t.me/KryptoSherlock (more similar reviews in English are in the pipeline) and get involved in a conversation at «Sherlock Club»: https://t.me/CryptosherlockClub