ICO Review — Fetch AI (FET)

Lennard Neo

Published in

Astronaut Capital

10 min readFeb 21, 2019

On the 21 of February 2019, Picolo Research released an ICO review report on Fetch AI (FET). View the full report here.

Making Distributed Ledger Technology Smarter

Fetch AI is seeking to combine machine learning (ML), artificial intelligence (AI), autonomous agent systems and a distributed ledger network to create an economy where data could be exchanged seamlessly. Machines integrated with Fetch AI’s protocol would be able to communicate and solve complex problems in our daily lives, eventually increasing efficiency that is beyond human capabilities.

Company overview

Fetch AI is incorporated in the UK with an aim to create a digital economy where autonomous agents can interact with one another without human intervention. The environment will create a decentralised digital world to perform useful economic work where information trades effortlessly. There is a plethora of applications that can be adopted, but the overall bottom line is to bring data to life.

Here are several issues that Fetch AI seeks to resolve:

Creating economic efficiencies by providing market intelligence to prevent issues such as shipping empty containers and energy waste by power grids
The “disconnect” between centralised systems as they cannot effectively communicate with one another
Scalability of distributed ledger technology

Commercial & Technical Strategy

Fetch AI is building a product that facilities execution of tasks, delivering data, providing services and getting rewarded for the efforts displayed. The technology encompasses a 3-layer architecture; 1) Autonomous Economic Agents (AEAs), 2) Open Economic Framework (OEF), 3) Smart Ledger. ML and AI are integrated across all layers to incorporate an intelligence element into the system.

Several significant features include:

OEF — The digital world where AEAs operate in, allowing them to access services and connect with other peers. In addition, high-level commands (transfer, explore, search and discover) and low-level commands (reaching consensus, peer trust, transactions and protocol control) can be executed
AEA — A digital entity that acts on its behalf. They are paired with existing hardware and can exchange data with other agents to fulfil requirements of the owner
Smart Ledger — A unique blend of blockchain and DAG with two main components; 1) smart contract on the ledger and 2) data stored on the DAG
uPoW — Useful proof-of-work consensus that builds market intelligence and trust over time
Machine learning intelligence that increases efficiency using past, present and future elements employed by AEAs
A synchronisation of 40k tps between nodes and 30k tps for mining blocks

Team

Fetch AI has a total of 31 employees working for the company as compared to 38 listed on their website. Here we sort out relevant personnel together with their experiences:

Humayun Sheikh (CEO & Founder) — 23+ years of experience as an entrepreneur and investor within IoT, ML and AI space. One of the first investor in DeepMind, owned by Google

Toby Simpson (CTO) — 27+ years of experience in programming and designing engines. Past experiences include Ososim Limited, DeepMind, Nice Tech Limited and Creature Labs

Thomas Hain (CSO) — 16+ years of experience in research and academia in advanced machine learning technology. A professor and the head of Speech and Hearing Research at the University of Sheffield. Holds a PhD from Cambridge

Jonathan Ward (Head of Research) — 13+ years of experience as a research scientist and engineer in algorithmic design. An expert in the field of ML, complex systems and blockchain. Holds a PhD in computer science from UCL

Troels F. Ronnow (Head of Software Engineering) — 17+ years of experience. Expert in the field of computer science and theoretical physics. Prior experience includes Nokia Technologies, Post. Doc., and Utopia Solutions. A co-author of 35 patent applications and holds a PhD from Aalborg University

Maria Minaricova (Head of Business Development) — 20+ years of experience in marketing and program management. Prior experience includes GEANT, Oracle, Arthur Andersen, and AMP. She is also a member of several groups including IoT alliance, Women in AI, EU Blockchain observatory and forum, and European AI alliance.

Arthur Meadows (Head of Investor Relations) — 24+ years of experience in product marketing and consultancy. Prior experience in companies such as Uvue Ltd, itzMe, Stream 121, Grapeshot and NeuStar. Holds an MBA from Cambridge Judge Business School.

Addition team members include:

19 Software and Machine Learning Engineers
5 researchers and scientists

The project has a C-suite team that boasts a wealth of experience in establishing technology companies. They are backed by a strong team of engineers, developers, and scientists with experiences in big companies such as Google, Sony, Bloomberg, and Nokia. It is also important to note that almost 50% of the team are PhD holders, with 5 members not having a LinkedIn profile.

Advisors

Melvyn Weeks — Assistant Professor in Economics @ University of Cambridge

Monique Gangloff — Senior Research Associate @ University of Cambridge, PhD in Structural Biology

Niall Armes — CEO @ TwistDX Ltd, PhD in Molecular Biology

Steve Grand — Expert in creating artificial life forms, AI, Robotics

Jamie Burk — Founder & CEO @ Outlier Ventures

Kash Iftikhar — Vice President @ Oracle, Head of Business Unit @ Dell, MBA

Investors

Outlier Ventures (Seed Stage) — With a believe in decentralised data infrastructure, Outlier Ventures invests at the convergence of blockchains with artificial intelligence, the internet of things and robotics. Past investments include Haja Network, Ocean Protocol, IOTA, Sovrin, SEED.

Blockwall Management GmbH (Seed Stage) — A European asset manager exclusively focused on crypto assets, specifically in layer-1 protocols.

Partnerships

MOBI consortium — As a member of the consortium, Fetch AI is working to collaborate with manufacturers within the transportation and mobility sector

Blockchain for Europe — Fetch AI is one of the 4 founding members for the association, which represents blockchain organisations in Europe alongside Ripple, NEM and Cardano

Artificial Intelligence Innovation Network — A network that brings researchers across diverse fields to examine transformational AI technology on business and society

ULedger — A blockchain protocol that allows Fetch to integrate real world data from IoT devices into their network

Binance — Token sale will be conducted on Binance Launchpad

Roadmap

Fetch AI was incepted in 2018 and released its ledger code promptly in Q3 2018. However, the release of Virtual Machine (VM) was delayed in Q4 2019 and only released in Feb 2019. Below, we cite some major developments going forward:

Feb 2019 — Private test network
Apr 2019 — Full public test network with smart contracts, Develop community support
Jul 2019 — Alpha release
Q2 2019 — Beta release with all expected main net functionalities functional
Q4 2019 — Main net release

Token Sale

Fetch AI raised $2.06m in seed funding and $0.975m & 24,597 ETH in private sale. Total tokens supply will be fixed with 17.6% allocated for token sale (including seed and private rounds) and 15% for mining. Investors, advisors, foundation and team will have a lockup and vesting periods.

Fetch AI has a planned expenditure for its first three years namely on working capital, operation cost, and academic partnerships. The greatest costs incurred would be salaries and working capital associated with the development of the technology.

The main token functionalities are 5 folds:

Currency for transactions — FET tokens act as the medium of exchange between agents

Staking/ Mining — Users will be able to earn FET tokens by providing value to the network

Connect to the digital world — FET is required as a deposit to access the network to encourage appropriate behaviour

Access to services in the digital world — Allowing AEAs to view, explore and search the digital economy and interact with other nodes to negotiate services

Access to ML/AI algorithms — Access to a range of intelligence tasks that are available on the smart ledger

Community Engagement / Social Media

Fetch AI has several social media channels to engage their community. Regular updates on the team’s developments are posted on Twitter with a Medium page where articles are posted to educate the community on the technology. They also have a Youtube channel with 25+ videos explaining what the project is about.

Engagement developments have been gaining traction as Fetch AI’s homepage is constantly rising in ranks since Jan 2019 based on Alexa Rankings. The top 3 visitors by countires are US, India and UK respectively.

Here are the project social media statistics as per the report date:

Telegram (English) — 13,100+ members
Twitter — 3,921 followers
YouTube — 481 subscribers
Medium — 62 Followers

Competitor Analysis

Fetch AI will be competing with several blockchain projects that have the intention to integrate ML and AI. However, Fetch AI’s product features display superiority as compared to their peers. Here we list a couple of its competitors in comparison to Fetch AI.

Strengths

Innovative product allowing the autonomous exchange of information with higher throughput against its competitors
Token sale conducted on Binance Launchpad adds to the credibility of the project
Potential exchange listing within a short period after TGE given Binance partnership
Good deal structure with low hardcap of $6m and reasonable market cap valuation based on the initial circulating supply of tokens
A strong team with great wealth of experience and academic record

Weaknesses

Past companies founded by CEO had huge liabilities (UVue with GBP 800k debt, and Mettalis with GBP 3.2m debt). Another company itzMe has shut down its website and no longer operating
Delayed VM launched could create a snowball effect onto other milestones in the roadmap
Minimal commits and no Github activity from most employees, however, Binance has asserted that most developments are done privately and will move to public repositories in due course

Opportunities

AI $15.7 trillion impact — PwC research predicts that AI could contribute $15.7 trillion to the world economy by 2030. This is equivalent to a 26% and 14.5% boost to GDP in China and North America respectively. The sectors that will have the greatest potential AI impact would be automotive, healthcare, and retail and consumer sectors. PwC also states that one of today’s start-ups could become a market leader in 10 years.
Truly Autonomous Blockchain environment — This sector has been saturated with numerous blockchain infrastructure projects, however, those that have ML and AI integrated are far from many. Not to mention one that can create a digitalised environment for devices (agents) to interact with one another and to take advantage of meaningful data. Furthermore, most applications interact with data through pull methods rather than push and Fetch AI’s technology is seeking to resolve this.

Threats

Machine learning models require huge computing power, and resources to run. This could increase exponentially if the complexity of the problem increases, eventually resulting in the network becoming inefficient
Developing AI programs could get out of hand if not developed right. One good example is Facebook abandoning two of their AI experiments after the bots created an incomprehensible language that only the machines could understand

Conclusion

In conclusion, Picolo Research presents a ‘Spec Buy’ rating on Fetch AI. The technology offers a unique proposition where the ecosystem is crowded with blockchain infrastructure projects. Generating additional value through the creation of AEAs will re-establish how inefficient information is being utilized. The opportunity is there, and it essentially boils down on the team delivering the technology timely.

We highlight several reasons to affirm our rating,

A novel idea to create a digital world where autonomous agents can interact with one another
Binance launchpad listing enhances the project’s credibility and adds foundational support to the team
Good deal structure with low hardcap and market cap valuation based on initial circulating token supply
Significant competitive edge against similar projects

Not withstanding the above, Picolo acknowledges a couple of issues such as close to empty GitHub repositories that have been rectified by Binance and will be released in due course. In addition, most top management positions have links with past company failures by the CEO in the past such as Ososim, Uvue, and NOVUS4, which had huge debts or had ceased to exist. However, the debt has been cleared and it is difficult to attribute failures to a handful of employees as there are other employees involved. Last but not least, integrating AI and ML into blockchain is challenging, affecting the practical implementation of the network.

In light of the preceding, Picolo analysts recommend an overall ‘Spec Buy’ rating on Fetch AI.

End.

Produced by Picolo Research

Picolo Research: https://picoloresearch.com/
Astronaut Capital: https://www.astronaut.capital/
Medium: https://medium.com/astronaut-capital
Twitter: https://twitter.com/astronautcap
Telegram: https://t.me/astronautofficial