Lost for words? We asked a robot to write this article.

The recent explosion of progress in the field of artificial intelligence, fueled in part by the exponential growth of computing power as predicted by Moore’s Law, has sparked a renewed interest in the ethical implications of such technology and its potential impact on the global economy. Philosophers, politicians, and entrepreneurs alike are debating the potential effects of AI in areas like employment, finance, logistics, academia, art, and healthcare. Many are also beginning to think more deeply about the responsibility of AI developers to ensure that their systems are fair and transparent, and to consider the potential risks and benefits of AI for different groups of people.

As we stand on the brink of this profound technological shift, it is worth considering the cautionary tales of science fiction authors like Asimov (Isaac), who warned us of the potential dangers of unrestrained artificial intelligence. In his famous "Three Laws of Robotics," Asimov proposed a set of ethical guidelines for the creation and use of AI, underscoring the need for careful consideration of the consequences of our actions.

Many feel these advancements place us in close proximity to The Singularity, the hypothetical future event with potentially dire consequences in which artificial intelligence will surpass human intelligence. At the same time, the promise of AI is immense, with the potential to greatly expand commerce, quality of life, and the frontiers of human knowledge and capability.

As a leading technology advisor and investor, it is both our desire and responsibility to carefully analyze the potential impacts of these developments on the global economy and the organizations that comprise it. In this research report, we will explore the latest developments in AI technology, discuss how businesses today deploy/monetize it, provide insights and recommendations for businesses and investors looking to capitalize, assess potential risks and rewards of investing in AI-related businesses and industries, and discuss the main players in the landscape for this exciting and transformative technology.

"When greater-than-human intelligence drives progress, that progress will be much more rapid.”
-Ray Kurzweil, Director of Engineering at Google

Introduction

Eric Crowley

Adam Segall

partner

Analyst

Jeffrey Cohen

Research assistant

TEAM

Note: Images generated by consumer AI platform NewProfilePic

Introduction to AI

AI, or artificial intelligence, is a broad term that refers to the ability of a machine or computer program to exhibit intelligent behavior. There are three main types of AI, which are distinguished based on their level of intelligence and capability:

Artificial General Intelligence (AGI): Hypothetical concept of a machine that has the ability to understand or learn any intellectual task that a human being can. This would require the machine to have a broad range of abilities and be able to adapt to new situations and tasks.

Artificial Narrow Intelligence (ANI): Used to describe artificial intelligence that is focused on a specific task or set of tasks. This type of AI is currently the most common, and examples include virtual assistants like Siri or Alexa, or self-driving cars. Most references to “AI” in this report typically pertain to ANI.

Artificial Superintelligence (ASI): ASI is a hypothetical concept that refers to a future AI that would be much more intelligent than any human being, and potentially even surpass human intelligence in every domain. This type of AI is often discussed in the context of the potential risks and benefits it could bring.

“First we build the tools, then they build us.”
-Marshall McLuhan, Philosopher

While AGI and ASI are still largely theoretical, the AI systems that we see in the world today are mostly examples of ANI. Within the realm of ANI, there are many different subsets with different applications, unique characteristics, and different techniques to achieve their goals. Some of the major subsets include:

Machine learning: This is a type of AI that involves training algorithms on large amounts of data to make predictions or take actions based on that data. Machine learning algorithms can be supervised, unsupervised, or semi-supervised, depending on how they are trained.

Natural language processing (NLP): This is a type of AI that focuses on enabling machines to understand, interpret, and generate human language. NLP algorithms are used in a wide range of applications, from language translation to voice recognition.

Computer vision: This is a type of AI that focuses on enabling machines to see and understand the world around them. Computer vision algorithms are used in applications such as image recognition and object detection.

Robotics: This is a type of AI that involves the use of machines to perform tasks that are typically done by humans.

Deep learning: This is a type of machine learning that involves the use of neural networks with many layers of processing. Deep learning algorithms are able to learn complex patterns in data and make predictions or take actions based on that data.

Within these subsets, there are many different models that are used to perform different tasks. For instance, neural networks, designed to mimic the structure and function of the human brain, are often used for tasks such as image recognition and natural language processing, while Large Language Models (LLMs) are trained on a massive amount of text data in order to produce human-like text. Many of the tools referenced in this report are examples of LLMs, and these models are just a few examples of the many different approaches that are being used to develop AI systems that are increasingly powerful and intelligent.

One of the most powerful and well-known of these AI systems is GPT-3, which is an LLM and serves as the core of the wildly popular ChatGPT by OpenAI as well as the generative AI content platform Jasper. GPT-3 is capable of generating human-like text and performing a wide range of natural language tasks, such as translation, summarization, and question-answering. Many of these tools fall within the realm of “generative AI,” a type of artificial intelligence that is focused on generating new content, such as text, images, or music. ChatGPT, which reportedly reached millions of users in the first few days, is hailed as a potential disruptor for legacy search engines like Google due to its impressive ability to interpret complex queries and provide responses instantly and in almost any format.

Recent progress in the space has been substantial and rapid, with the sophistication and accuracy of these models increasing exponentially over the recent years and months. Key factors driving this progress include exponential leaps in computational power and the use of open source technology and software, which allows AI researchers and developers to share their work and collaborate on new ideas. Many of the biggest strides in AI have been made by companies and organizations that are committed to open source, such as Google, Facebook, Stable Diffusion, and the OpenAI consortium. These organizations have made a significant investment in open source AI technology, and have contributed a great deal of their own research and development to the community. By doing so, they have helped to accelerate the development of AI and make it more accessible to a wider range of researchers and developers.

Output from ChatGPT, using the command

"Write a swashbuckling sea shanty about the promise of AI"

Previously dismissed as gimmicks, the current generation of AI tools will dramatically reduce the time to produce content, resulting in more content at lower costs. This dynamic opens up the potential for substantial changes in how media, content, professional and customer service firms operate.

AI has become a crucial part of many businesses' strategies due to its ability to automate tasks and processes, improve decision-making, increase efficiency and productivity, or simply provide a product/service not available anywhere else. The potential for AI to transform legacy businesses and drive economic growth is undeniable, and by effectively implementing AI, businesses can gain a competitive edge in their industries and improve their bottom line. Investors, on the other hand, can benefit from investing in companies that use AI in their operations, as these companies are likely to experience strong growth, high profitability, and sustainable competitive advantages. Some examples of how businesses currently leverage or are completely built around AI today include:

Marketing & Content Creation: AI is increasingly being used in the field of marketing and content creation. For example, some companies are using AI-powered tools to analyze customer data and identify patterns and trends to create personalized marketing messages, or content that is more likely to resonate with a specific audience. In addition, generative AI can be used to create new, original content, such as articles or social media posts, at scale. This can be useful for companies that need to produce a large volume of content quickly and efficiently. Some examples of companies using AI and generative AI in this way include Jasper.ai, Persado, and MarketMuse. Bessemer Venture Partners recently predicted that 50% of online content will be generated or augmented by AI, up from 1% today. Will search engines begin to become swamped by AI-generated content drowning out human voices?

Computer Science/Programming: In the computer science field, businesses are using AI to develop innovative solutions to complex software problems. One example of this is Github Copilot, which uses AI to assist developers with in-line code recommendations and creation while a developer works. The AI technology behind Github Copilot can automatically generate code based on the developer's inputs with correct syntax and in markdown, making it easier and faster for them to complete their projects.

Art/Creative: In the art and creative sphere, AI is being used to develop new forms of artistic expression and creativity. For example, the Midjourney and DALL-E 2 platforms use AI to generate unique pieces of art based on the user's inputs. The AI technology behind Midjourney can create abstract art, portraits, and other styles of art based on the user's preferences, allowing individuals and businesses to commission unique and personalized pieces of art. Platforms like Lensa are able to train off pictures of individuals and create highly stylized headshots, with some even taking in user prompts to recreate portraits in the styles of the old masters.

"AI has the potential to be more disruptive than the internet. It will change the way we do business, the way we live and the way we relate to one another."
-Ray Kurzweil, Director of Engineering at Google

A key consideration for the viability of these next-generation tools is that they can be easily accessed through a browser or an app with simple commands, which will rapidly accelerate adoption. There are also several different revenue models used by these businesses, each with its own pros and cons. Some of the most common revenue models for AI businesses include the following:

Subscription-based: This model is the most prevalent today and involves a standard recurring SaaS fee. It can be a good way to generate predictable, recurring revenue, but it may be difficult to attract or retain customers if the product or service is single/limited use (i.e., “novelty”) or not perceived as valuable enough to justify the ongoing cost. Businesses facing this dilemma are experimenting with ways to generate recurring value or switching away from the subscription model altogether. Given the consumer focus, many of these businesses fall within the universe of Consumer Subscription Software (CSS), which is one of GP Bullhound’s main areas of expertise. Please see our CSS reports for information on the nuances between enterprise and consumer subscription software.

Pay-per-use/pay-per-output: This model involves charging customers for each time they use the AI product or service, either on a flat fee or based on the time/computational resources consumed. This model lacks the stability of the subscription model but can have extremely high margins given the low cost of computing resources. Launches of new products or product updates also serve as additional opportunities to charge users within a subscription timeframe that would only charge users once. Shortcomings include limited revenue potential and customer commitment, potential for high fees on a per-transaction basis, and the need to constantly update/offer new things to keep customer interest.

Licensing: This model involves charging customers a one-time, recurring, or usage-based fee for the right to use the AI product or service, chiefly through integration in an existing offering. This can be a good way to generate a large amount of revenue upfront and transfer risk to the licensee, but upside potential is generally more limited.

Proprietary Data: In this model, the AI company would collect large amounts of data, such as customer information or market trends, and use its AI tools to process and analyze the data. The company could then sell this information to other businesses or organizations like healthcare/pharmaceutical companies, governments, or other AI developers, who could use it to make better-informed decisions or improve their products and services. This model can be highly profitable for the AI company, as the value of data is increasing and there is a high demand for high-quality, actionable data.

The most effective revenue model for an AI business will depend on a variety of factors, including the nature of the product or service, the target market, and the business's overall goals and objectives. However, as the technology continues to advance and become more sophisticated, it is likely that AI will disrupt even more industries in the future. For example, AI could be used for:

Legal: Automate document review and contract analysis at a large scale. Platforms like ndaOK already exist for some simple workflows, but we anticipate it will be some time until they are widely tested and accepted within the legal field for more complex tasks.

Journalism: Automate the process of gathering, analyzing, fact checking, and reporting news. This could lead to cost savings for news organizations, but could also raise concerns about the accuracy and reliability of AI-generated content.

Manufacturing/Industrials: Improve the efficiency of processes (e.g. for predictive maintenance and quality control) and logistics (e.g. for route optimization and real-time tracking).

Accounting/Finance: Identify accounting irregularities and whistleblow in cases like FTX and Enron. Quickly produce written reports by analyzing financial statements and craft predictions for future performance.

Media/Entertainment: Automate the creation of media, with examples of movie screenplays or games generated in ChatGPT, including descriptions of characters and locations, being loaded into Midjourney to create detailed and realistic images for storyboarding and costume design. We believe that AI technology will eventually be able to fully render movies and audio/music in the near future.

At GP Bullhound, we are always evaluating the next generation of technology. Testing the tools and then gradually working them into our normal workstreams has demonstrated the power of the current crop of ANI tools. The output is powerful, but more importantly, the consumer interface is easy to understand for users without machine-learning proficiency. We also see an opportunity for what we describe as “meta-models,” or AI models/technologies that can bridge the gap between existing but functionally separate models (i.e., take the output from one and reformat/load it into another automatically).

EBIT-AI – the coming impact of ai tools on digital businesses

Given the potential for disruption, we recommend that non-AI businesses today take steps to AI-proof their businesses to ensure continuity and sustainability. A business can take several steps to ensure that it is prepared for the potential impacts of AI technology that may include:

Educating employees about AI technology and its potential impacts on the business

Developing an AI strategy and roadmap

Investing in AI-related research and development

Building strong partnerships with AI vendors and industry experts

Implementing ethical and responsible AI practices

Training employees on telltale signs of AI-generated content from potential customers or partners

Developing unique and marketable tones/perspective/empathy/forward-thinking ability to differentiate from generic AI imitations – it will be harder to differentiate human produced versus AI-produced content in the next six months. These abilities will all become critical for creators to avoid producing copycat content.

By taking these steps, a business can effectively AI-proof itself to the best of its abilities. But if you are a business directly operating in the AI space, consider taking additional steps to ensure an advantage over other AI-based competitors that are unique to and separate from the standard points of differentiation for general software companies.

It is worth noting that the open source nature of many core AI technologies will greatly increase the competition in the space. One example of how an AI company can create an economic moat is by developing proprietary training data. Training data is a critical component of many AI algorithms, and the quality and quantity of the training data can greatly impact the performance of an AI model. By developing proprietary training data that is unique and high-quality, an AI company can make it difficult for competitors to replicate the performance of its AI models. We anticipate that sourcing high-quality and proprietary training data will be the next big arms race in AI. The race to scale will be incredibly important for AI companies as user feedback will improve the tools, which will pull in more users, creating a virtuous flywheel.

Other examples of how an AI company can create an economic moat include developing unique AI algorithms or technologies or gaining a deep understanding of a specific industry or use case and tailoring the technology accordingly. By creating an economic moat, an AI company can protect its market share and profits, and position itself as a leader in the industry.

Research/Consulting: Companies are using AI-powered tools to automate the process of collecting and analyzing research data (including NLP/chatbots for surveys) and identifying patterns and trends. This can help researchers to make faster and more accurate discoveries, and can also enable them to process larger amounts of data than would be possible manually. In addition, AI can be used to generate new hypotheses and test them using data, helping researchers to explore new ideas and potentially make breakthroughs in their fields. Some examples of companies using AI in research include Enlitic, Berg Health, and BenevolentAI. GP Bullhound also recently advised Arca Blanca, an AI/data science consultancy, on its acquisition by Artefact, a premiere digital strategy and innovation firm.

Gaming: Latitude’s AI Dungeon generates fictional stories based on user prompts for co-authored adventures. Voyage takes this concept to the next level in immersive games.

Communications: Companies are increasingly using AI to refine and enhance the tone and content of their communications. One example of a communications company using AI is Volograms, which uses AI algorithms to create immersive and interactive 3D experiences for users.

Personal Coaching/Therapy/Relationships: Companies are using AI-powered chatbots to provide personalized advice and support to users to simulate a conversation with a human, allowing users to discuss their thoughts and feelings in a non-judgmental environment. Replika is an AI companion that has reached over 10 million users who spend hours and send dozens of messages on the platform per day. Generative AI can also be used to create personalized content, such as personalized affirmations or meditations, based on a user's specific needs and goals. Some examples of companies using AI and generative AI in this way include Woebot, Relationup, and Stellena.

'Humanoid android in corporate board room giving presentation to female shareholders, realistic, photorealistic, HDR, 4k', an image generated by Midjourney

'Robot stock traders on the floor of the New York Stock Exchange', an image generated by Midjourney

'US founding fathers but they are robots drafting the constitution in style of revolutionary war era painting', an image generated by Midjourney

'Scientists in lab coats standing around a computer in a dark room illuminated by the light from the computer with a robotic Vitruvian man on the screen', an image generated by Midjourney

OutsAIzed Returns

As AI technology continues to advance and become more widely adopted, there are many potential opportunities for both public and private market investors to capitalize on its growth and integrate it into their investment strategies. They have the potential to generate outsized returns, and strategies or portfolios with allocations for AI are likely to outperform and place capital allocators well above competitors or benchmarks in both private and public markets.

One way for investors to do this is to invest in companies that are working on developing AI technology. This could include companies that are focused on AI research and development for use by others, as well as those in other industries where AI is a core component of their product or service. Previously mentioned points of differentiation to pay attention to when investing in AI companies include proprietary data and algorithms with superior performance.

In addition to directly investing in AI-related companies, investors can also look for opportunities to invest in companies that are using AI to drive growth and improve their operations. For example, a company in the healthcare sector that is using AI to develop new medical treatments or improve patient care could be a good investment opportunity. These examples of highly sustainable “economic moats,” or enduring competitive advantages that come from AI, make for a core component of any compelling investment thesis.

Private market investment in the AI industry has been on the rise in recent years, as more and more companies look to capitalize on the growth and potential of this technology. Venture capital and private equity firms have been investing heavily in AI-related companies, providing them with the capital they need to develop and scale their technology while realising significant returns in the process.

M&A activity in the AI space has also been increasing, as larger companies look to acquire smaller AI startups to gain access to their technology and expertise. This has led to a number of high-profile M&A deals in the AI industry, with companies like Google, IBM, and Apple all making significant acquisitions in this space.

There also exists significant potential for capital allocators to integrate AI in their own processes for investing in non-AI businesses or industries. For instance, AI algorithms can be used to analyze large amounts of financial data and identify trends and patterns that may not be visible to human investors. This can help investors to make more informed and profitable investment decisions.

But this new frontier is not without its fair share of risk, as many speculate AI may be entering a “bubble,” or a situation where the price of an asset becomes greatly inflated due to high demand and speculation. Bubbles often form when investors become overly optimistic about the potential of an asset, and are willing to pay increasingly high prices for it. This can lead to a rapid increase in the price of the asset, which may not be sustainable in the long term. Eventually, the bubble will typically burst, leading to a sharp decline in the asset's price and significant losses for investors who bought at the peak of the bubble. Investors should ensure implementation of continuation of proper risk management protocols and diversification to insulate themselves from potential downside given the highly speculative nature of the industry at this time.

The current AI revolution is different from past investing bubbles for a number of reasons. First, the growth and adoption of AI technology is being driven by real and tangible advances in the technology, rather than speculative hype and over-optimism. AI technology has made significant progress in recent years, and its potential to transform a wide range of industries is becoming increasingly clear.

Second, the AI market is still relatively small, with global spending on AI technology estimated to reach $190 billion by 2025. This is significantly smaller than other technology markets, such as the internet and mobile industries, which have reached much larger scales. This suggests that there is still significant room for growth in the AI market, and that it is not yet at the point of saturation or over-investment.

Finally, the AI industry is still in its early stages, and many of the companies operating in this space are still working on developing and refining their technology. This means that there is still a high level of uncertainty and risk in the AI market, and investors are carefully assessing the potential of individual companies before making investment decisions.

While there are risks and uncertainties in any investment market, the current AI revolution does not appear to be a bubble from an investing standpoint. Instead, it represents a unique and potentially transformative opportunity for investors (the information in this article is provided to you for informational purposes only and should not be regarded as an offer or solicitation of an offer to buy or sell any investments or related services that may be referenced in this article).

“This is by far the fastest-moving technology that we’ve ever tracked in terms of its impact and we’re just getting started.”
-Paul Dougherty, Chief Technology Officer at Accenture

Notable AI/ML private placements

Landscape of prominent AI investors

Sources: GP Bullhound Insights and Pitchbook (16 December 2022)

Source: GP Bullhound Insights

US Generative AI Market Size, 2022-2030 ($bn)

Source: Grand View Research

Issues and implications of AI use

Like any technology, AI is not without its limitations and potential issues. First, one of the most well-known limitations of AI is its tendency to exhibit bias. This can manifest in a number of ways, but the most pernicious forms of bias are sexual and racial bias. For example, AI algorithms that are used in the criminal justice system have been found to be disproportionately harsher on people of color and individuals who are poor. Similarly, AI algorithms used in the hiring process have been shown to have a bias against women and people from certain racial backgrounds. This bias is often a result of the data that is used to train these algorithms, which is often biased itself. This can lead to unfair and discriminatory outcomes, which can have serious consequences for the individuals affected. To combat bias in AI, there are a few potential approaches:

Use diverse and representative data sets: One of the key causes of bias in AI algorithms is the data that is used to train them. To reduce bias, it is important to use diverse and representative data sets that accurately reflect the populations that the algorithms will be used on.

Regularly evaluate and audit algorithms: Another approach to combat bias in AI is to regularly evaluate and audit algorithms to identify potential sources of bias. This can include testing algorithms on diverse data sets and using human moderators to review the results.

Use fairness metrics: To further combat bias in AI, researchers and developers can use fairness metrics to measure and evaluate the bias of algorithms. These metrics can help identify and mitigate sources of bias in AI algorithms.

Foster diversity and inclusion in the tech industry: Finally, to combat bias in AI, it is important to foster diversity and inclusion in the tech industry. This can include initiatives to attract and retain a diverse workforce, as well as creating an inclusive and equitable workplace culture.

“The real question is, when will we draft an artificial intelligence Bill of Rights? What will that consist of? And who will get to decide that?”
-Gray Scott, Futurist, Philosopher and Artist

Another major drawback of AI is its potential to spread misinformation. This is particularly concerning in the age of social media, where AI algorithms are often used to spread false or misleading information at a rapid pace. This can have serious consequences, as people may believe and act on this misinformation, leading to widespread confusion and even harm. AI algorithms can also be used to create fake news and propaganda, which can be difficult to detect and can have serious implications for society. To combat this issue, there are a few potential approaches:

Use fact-checking algorithms: One way to combat the spread of misinformation using AI is to develop and use fact-checking algorithms. These algorithms can be trained to identify false or misleading information and flag it for review by human moderators.

Educate the public: Another approach to combat the spread of misinformation using AI is to educate the public about the dangers of misinformation and how to spot it. This can include providing resources and training on how to identify fake news and propaganda, as well as the importance of verifying information from multiple sources.

Regulate the use of AI in the media: To further combat the spread of misinformation using AI, governments and regulatory bodies can implement policies and regulations to limit the use of AI in the media. This can include requiring AI algorithms to be transparent and accountable, as well as imposing penalties for the use of AI to spread misinformation.

Encourage media literacy: Finally, to combat the spread of misinformation using AI, it is important to encourage media literacy among the public. This can include initiatives to teach critical thinking skills and the importance of evaluating the credibility of information sources.

One of the other key issues with the use of AI in education and research is plagiarism. AI algorithms like ChatGPT are able to synthesize new material in an exceptionally natural voice, making it difficult to detect when plagiarism has occurred. This can undermine the integrity of research and lead to the spread of incorrect or fraudulent information in addition to the formation of poor work ethics in students and researchers. To combat this issue, there are a few potential approaches:

Use AI algorithms to detect plagiarism: One way to combat the use of AI for plagiarism is to develop and use AI algorithms that are specifically designed to detect copied or AI-generated text. These algorithms can be trained to identify copied text and flag it for review by human moderators.

Educate researchers and students: Another approach to combat the use of AI for plagiarism is to educate researchers and students on the importance of avoiding plagiarism and the potential consequences of engaging in this behavior. This can include providing resources and training on proper citation and research practices.

Implement strict policies and penalties: To further combat the use of AI for plagiarism, institutions can implement strict policies and penalties for plagiarism to deter individuals from engaging in this behavior. This can include imposing fines or other penalties for plagiarism, as well as requiring researchers and students to undergo training on plagiarism detection and prevention.

AI is still in the very early stages but is evolving at an unprecedented pace. Only recently, very few people outside of Silicon Valley had heard of AI, but today millions are using consumer-ready AI tools like Jasper, Lensa, and Dall-E. While claims of the arrival of AI have happened frequently over the past 20 years, this time it feels different. Only time will tell if it is truly a new age of AI, but we continue to monitor the AI industry and are prepared to advise our clients on this rapidly evolving sector.

Disclaimer

No information set out or referred to in this communication shall form the basis of any contract. The issue of this insights report (the “report”) shall not be deemed to be any form of binding offer or commitment on the part of GP Bullhound or any of its affiliates or subsidiaries. This report is provided for use by the intended recipient for information purposes only. It is prepared on the basis that the recipients are sophisticated investors (so-called “professional clients” in the meaning of Annex II of Directive 2014/65/EU on markets in financial instruments, or their equivalent elsewhere) with a high degree of financial sophistication and knowledge. This report and any of its information is not intended for use by private or retail investors in the UK or any other jurisdiction where access, use or availability of this information would be unlawful.

This report does not provide personalized advice or recommendations of any kind. You, as the recipient of this report, acknowledge and agree that no person has nor is held out as having any authority to give any statement, warranty, representation, or undertaking on behalf of GP Bullhound in connection with the contents of this communication. Although the information contained in this report has been prepared in good faith, no representation or warranty, express or implied, is or will be made and no responsibility or liability is or will be accepted by GP Bullhound. In particular, but without prejudice to the generality of the foregoing, no representation or warranty is given as to the accuracy, completeness or reasonableness of any projections, targets, estimates or forecasts contained in this report or in such other written or oral information that may be provided by GP Bullhound. The information in this report may be subject to change at any time without notice.

GP Bullhound is under no obligation to provide you with any such updated information. All liability is expressly excluded to the fullest extent permitted by law. Without prejudice to the generality of the foregoing, no party shall have any claim for innocent or negligent misrepresentation based upon any statement in this report or any representation made in relation thereto. Liability (if it would otherwise but for this paragraph have arisen) for death or personal injury caused by the negligence (as defined in Section 65 of the Consumer Rights Act 2015) of GP Bullhound, or any of its respective affiliates, agents or employees, is not hereby excluded nor is damage caused by their fraud or fraudulent misrepresentation.

This report should not be construed in any circumstances as an offer to sell or solicitation of any offer to buy any security or other financial instrument, nor shall they, or the fact of the distribution, form the basis of, or be relied upon in connection with, any contract relating to such action. The information contained in this report has no regard for the specific investment objectives, financial situation or needs of any specific entity and is not a personal recommendation to anyone. Persons reading this report should make their own investment decisions based upon their own financial objectives and financial resources and, if in any doubt, should seek advice from an investment advisor. Past performance of securities is not a guide to future performance and the value of securities may fall as well as rise. In particular, investments in the technology sector may be subject to frequent fluctuations. The information contained in this report is based on materials and sources that are believed to be reliable; however, they have not been independently verified and are not guaranteed as being accurate. The information contained in this report is not intended to be a complete statement or summary of any securities, markets, reports or developments referred to herein.

This report may contain forward-looking statements, which involve risks and uncertainties. Forward-looking information is provided for illustrative purposes only and is not intended to serve as, and must not be relied upon as a guarantee, an assurance, a prediction or a definitive statement of fact or probability. Actual events and circumstances are difficult or impossible to predict and may differ from assumptions.

Any and all opinions expressed are current opinions as of the date appearing on the documents included in this report. The information contained in this report should not be relied upon as being an independent or impartial view of the subject matter, and for the purposes of the rules and guidance of the Financial Conduct Authority (“the FCA”) and of the Financial Industry Regulatory Authority (“FINRA”), and of the Commission de Surveillance (du Secteur Financier “CSSF”), this report shall not be viewed as a research report and is considered marketing communication and a financial promotion. Thus, in accordance with COBS 12.2.18 of the FCA Handbook, its contents have not been prepared in accordance with legal requirements designed to promote the independence of investment research and it is not subject to any prohibition on dealing ahead of the dissemination of the report.

The individuals who prepared the information contained in this report may be involved in providing other financial services to the company or companies referenced in this report or to other companies who might be said to be competitors of the company or companies referenced in this report. GP Bullhound, through its investment banking and asset management departments, does and seek to do business with companies which are, or may be covered in this report. As a result, both GP Bullhound and the individual members, directors, officers and/ or employees who prepared the information contained in this report may have responsibilities that conflict with the interests of the persons who access this report.

GP Bullhound and/or connected persons may, from time to time, have positions in, make a market in and/ or effect transactions in any investment or related investment mentioned in this report and may provide financial services to the issuers of such investments. The information contained in this report or any copy of part thereof should not be accessed by a person in any jurisdictions where its access may be restricted by law and persons into whose possession the information in this report comes should inform themselves about, and observe, any such restrictions. Access of the information contained in this report in any such jurisdictions may constitute a violation of UK or US securities law, or the law of any such other jurisdictions. Neither the whole nor any part of the information contained in this report may be duplicated in any form or by any means. Neither should the information contained in this report, or any part thereof, be redistributed or disclosed to anyone without the prior consent of GP Bullhound. GP Bullhound and/or its associated undertakings may from time to-time provide investment advice or other services to or solicit such business from any of the companies referred to in the information contained in this report.

Accordingly, information may be available to GP Bullhound that is not reflected in this material and GP Bullhound may have acted upon or used the information prior to or immediately following its publication. However, no person at GP Bullhound (which includes its members, directors, officers and/or employees), may undertake personal transactions in financial instruments of companies to which this report relates, without receiving prior clearance from the GP Bullhound Compliance Officer or nominated delegated. In addition, GP Bullhound, the members, directors, officers and/or employees thereof and/or any connected persons may have an interest in the securities, warrants, futures, options, derivatives or other financial instrument of any of the companies referred to in this report and may from time-to-time add or dispose of such interests. GP Bullhound Corporate Finance Ltd and GP Bullhound Asset Management Ltd are private limited companies registered in England and Wales, registered numbers 08879134 and 08869750 respectively, and are authorised and regulated by the Financial Conduct Authority. Any reference to a partner in relation to GP Bullhound is to a member of GP Bullhound or an employee with equivalent standing and qualifications. A list of the members of GP Bullhound is available for inspection at its registered office, GP Bullhound 78 St. James's Street, London SW1A 1JB.

For US Persons: This report is distributed to US persons by GP Bullhound Inc. a broker-dealer registered with the SEC and a member of the FINRA. GP Bullhound Inc. is an affiliate of GP Bullhound Corporate Finance Ltd. All investments bear certain material risks that should be considered in consultation with an investors financial, legal and tax advisors. GP Bullhound Inc. engages in private placement and mergers and acquisitions advisory activities with clients and counterparties in the Technology and CleanTech sectors.

In addition, the persons involved in the production of this report certify that no part of their compensation was, or will be, directly or indirectly related to the specific views expressed in this report. As such, no person at GP Bullhound (including its members, directors, officers and/or employees) has received, or is authorised to accept, any inducement, whether monetary or in whatsoever form, in counterparty of promise to issue favorable coverage for the companies to which this report may relate.

In the last twelve months, GP Bullhound or an affiliate is or has been engaged as an advisor to and received compensation from, or has invested in the following companies mentioned in this report: Arca Blanca.