AI is a broad term, referring to algorithms or software designed to imitate the human brain’s problem-solving and decision-making functions. An AI tool can undertake far more complex tasks than a conventional computer programme, and improve itself continuously as underlying datasets grow.
AI is not a new concept. Machine learning (ML), a sub-category within AI, has been used by many businesses for years. DeepMind, for example, which was acquired by Google-parent Alphabet in 2014, is integrated into both its search engine and Youtube’s video-suggestion algorithm.
Generative AI tools and large language models, such as OpenAI’s ChatGPT and Google’s Bard, have continuously advanced. This progress has created a huge range of potential use cases for businesses and consumers, even those outside of the technology ecosystem and without deep technical expertise.
As the core enablers of its progress – increasing computer power, more robust cloud data infrastructure, and a deepening talent pool – continue to develop, AI is poised for even greater transformation, with the potential for significant disruption.
"...AI is poised for even greater transformation, with the potential for significant disruption."
An overview of AI
AI is a general term covering a broad range of technologies and use cases across many sectors, but it can be broadly grouped into four segments.
Horizontal platforms
Horizontal platforms provide tools and solutions for businesses across sectors, allowing them to integrate AI and ML practices into their day-to-day operations.
Key sub-segments
AI core: Building blocks of AI and ML deployments, including developer tools needed to build and deploy models to production.
Computer vision software: The use of AI and ML to analyse visual data and make meaningful predictions about both the physical world and digital images.
AI automation platforms: Software and services that allow companies in all industries to use AI to automate critical business processes via predictive analytics.
Natural language technology (NLT): NLT uses computational linguistic techniques to learn from communications data and make predictions about the structure and content of language.
Market size
According to Pitchbook estimates, spending on horizontal platforms reached $32.4 billion in 2022, and will grow at a 32.7% compound annual growth rate (CAGR) to $75.7 billion in 2024.1
Example companies
Grammarly - real-time writing assistance
Databricks - simplifying big data analysis
Vertical applications
Vertical applications in AI and ML are tailored to tackle industry-specific problems, and they may not necessarily follow an AI-first approach. In this category, many startups adopt a pragmatic strategy, first developing software solutions to address particular challenges within an industry. Later, they strategically integrate AI and ML components to optimise specific aspects of their products.
Key sub-segments
AI and ML in financial services: Includes technologies that embed AI and ML into existing financial services via advanced analytics, process automation, robo-advisors, and self-learning programs.
AI in healthcare: Includes technologies that leverage AI and ML to improve medicine and care provision.
Consumer AI: Includes technologies that use AI and ML to enhance business-to-customer business models.
AI in IT: Includes enterprise software tools that optimise specific functions typically administered by IT departments, including both back-end and front-end use cases.
Market size
According to Pitchbook estimates, investments in the vertical applications market reached $94.0 billion in 2022, with a 31.2% CAGR out to 2025, resulting in a $212.2 billion market.[2]
Example companies
Indigo - plant microbiome agricultural services
GitHub - software developer platform
Anyfin - loan-refinancing platform developer
Autonomous machines
Autonomous machines are sophisticated systems capable of executing tasks in human-present environments without requiring explicit human control. These machines seamlessly integrate machine learning, computer vision, and comprehensive datasets of the physical world, particularly in the context of navigation.
Key sub-segments
Autonomous vehicle software and design: Software and hardware solutions that provide self-driving or driver assistance capabilities for cars, trucks, and other on-road vehicles.
Intelligent robotics and drone design: Robotic systems and uncrewed aerial vehicles that can operate without human input. AI and ML can be used for the learning, control, and adaptation of robots.
Intelligent robotics and drone software platforms: Operating systems for autonomous robots. These platforms can include fleet management and predictive maintenance.
Market size
According to Pitchbook estimates, the autonomous machines market is $41.1 billion as of 2022 and is forecast to grow at an 18.9% CAGR to $58.1 billion in 2025.[3]
Example companies
Cruise - manufacturer of self-driving autonomous vehicles
Pony.AI - autonomous driving technology developer
Geek+ - robotics technology developer
AI and ML semiconductors
The training and inference processes of AI and ML models demand hardware optimised for maximum computational efficiency. Such hardware requires processing capabilities tailored specifically for AI calculations.
Key sub-segments
AI chips: These types of computer chips attain high efficiency and speed for AI-specific calculations.
Edge AI: Compression algorithms that optimise AI and ML models for deployment within various semiconductor environments and edge devices.
Intelligent sensors and devices: Devices that are optimised to run AI and ML models.
Market size
According to Pitchbook estimates, the AI and ML semiconductor market reached $43.6 billion in 2022 achieving 22.5% growth over 2021. Over half of this total can be attributed to mobile phone application-specific integrated circuits, which is expected to represent a smaller share of the market over time. Pitchbook expects the market to grow at a 21.8% CAGR from 2022 to 2025, resulting in a $72.7 billion market.[4]
Example companies
Nvidia - the market-leading designer of microchips for AI.
Spotlight: generative AI
Generative AI refers to AI systems that produce outputs resembling those of humans. These models use deep learning to create new data such as images, text and audio. Examples include:
generative adversarial networks, such as those used in lifelike images and deepfakes;
text-generation models like ChatGPT, whose output is human-like text;
and music-generation models that create compositions.
AI and ML Investment
Given AI’s significant potential and the fact that early tools such as machine learning have had such a material impact on business operations, some venture capitalists and private equity investors have focused on the sector for many years. However, with recent step changes in technological progress, investment in generative AI businesses is gaining momentum.
According to Pitchbook, investment in businesses incorporating AI and machine learning peaked at $56.9 billion in the third quarter of 2021, as shown in Chart 1, below. Although investment flows slowed from the second quarter of 2022, reflecting the broader market slowdown, AI and ML investment remained robust at $36.8 billion for the same period in 2023.
Chart 1: Private capital invested in AI and ML remained robust in H1 2023
Source: Pitchbook Data, Inc. Capital invested in USD billions. Data includes all AI and machine learning investment by private equity, venture capital, and corporate M&A.
A key feature of AI investment is the high concentration of deal activity in North America, which benefits from a long-established technology ecosystem and strong network effects. While the share of investment has been lower in Asia and Europe, capital invested is still meaningful in these regions.
Chart 2: North America is the most active region for AI investment
Source: Pitchbook Data, Inc. Data includes all AI and Machine learning investment by Private Equity, Venture Capital, and Corporate M&A. ‘Other’ includes Africa, Central America, Middle East, Oceania and South America.
In 2023, generative AI, partly popularised by the launch of OpenAI’s Dall-E and ChatGPT, has become a significant focus for investors. In the first half of 2023, $20.5 billion of capital was invested in generative AI businesses, an amount higher than the combined total of the previous four years, as shown in Chart 3.
Chart 3: Private capital invested in generative AI
Source: Pitchbook Data, Inc. Capital invested in USD billions. Data includes all generative AI deals completed though private equity, venture capital, and corporate M&A.
With AI gaining traction due to better understanding of its use cases and potential end-markets, valuations for generative AI businesses have also increased markedly. The median post-money valuation for a generative AI deal has increased by 39% from 2022 to 2023, reaching $50 million, as Chart 4 details. There is also a clear upside bias as market-leading businesses command premium valuations. In recent months there have been reported fundraising rounds at more than 100x revenue multiples, and businesses pre-revenue and pre-product being valued in the hundreds of millions.[5]
Chart 4: Generative AI valuations have increased markedly
Source: Pitchbook Data, Inc. Post-money valuation in USD billions. Data includes all generative AI deals completed though private equity, venture capital, and corporate M&A.
Transaction examples
Open AI
Developer of an artificial intelligence-based research and deployment platform intended to ensure that artificial general intelligence benefits all of humanity. Key products include ChatGPT (a text based AI chatbot), and Dall-E (an AI image generation tool).
Microsoft invested $10 billion at a $19 billion valuation in January 2023, for a 49% ownership stake with a 75% profit share until $10 billion had been returned. OpenAI had previously raised c.$1.3 billion of capital, primarily from Microsoft.
Inflection
Developer of an artificial intelligence-based platform designed to improve human-computer interaction.
The company raised $1.3 billion of venture funding in a deal led by Microsoft, Nvidia, and Reid Hoffman in June 2023, at a post-money valuation of $4 billion.
MosaicML
Developer of software infrastructure and artificial intelligence training algorithms designed to improve the efficiency of neural networks.
MosaicML was acquired by Databricks in July 2023 for $1.3 billion. It had previously raised $33.7 million in venture funding with investors including LuxCapital and Samsung NEXT.
Hippocratic AI
Developer of a safety-focused large language model intended for the healthcare industry.
Hippocratic AI raised a $50 million seed round at a $200 million post-money valuation in Q2 2023, co-led by General Catalyst and Andreesen Horowitz.
Source (transaction data): Pitchbook, August 2023.
Challenges and risks
AI’s power and potentially transformative benefits bring with it natural risks and challenges.
Data quality and data availability
To work effectively, AI needs large volumes of data to help it make accurate predictions. In the context of private markets, this is particularly challenging given the opacity of the industry and difficulty in obtaining accurate data on private companies.
Interpretability
While AI can be used to interpret swathes of data, it can be difficult to follow the procedure it undertook to reach each conclusion. Many AIs use complex algorithms, far exceeding human capabilities. While this has many benefits, a human’s ability to explain a decision and how they reached a conclusion is invaluable.
Regulation and security
Innovation often precedes regulation, and AI is no different. Regulatory developments are inevitable and necessary given AI’s power. These could materially change how tools are developed and deployed, and, ultimately, determine which AI businesses are viable. Of course, data security and the threat of bad actors must also be carefully managed.
Conclusion
AI represents a sea change in technological development. Although still relatively nascent, AI tools and solutions already offer material benefits to businesses and consumers. A new generation of companies, one that will build even more powerful AI tools, is emerging. Many industries will face material disruption and will be forced to adapt.
"A new generation of companies, one that will build even more powerful AI tools, is emerging..."
The AI landscape is becoming increasingly competitive. There are already new, open source large language models, built with very small teams, that are proving more capable than ChatGPT and Bard on many benchmarks. AI businesses that emerge as market leaders will likely have proprietary datasets on which they can train their models, providing intellectual property and a defensible competitive advantage. Additionally, user experience matters. AI tools will not be embraced by businesses and consumers alike if they are not easy to understand and use.
The rise of AI has undoubtedly generated immense interest and excitement in the investment world. However, this increased interest has also raised concerns about the potential for speculative investments and the formation of a market bubble. As competition intensifies, entry multiples are likely to rise further, making it essential for investors to exercise caution and prudence.
In order to navigate this dynamic landscape, diversification in investment strategies becomes crucial. The best way to gain exposure to AI is through a diversified portfolio, which spreads the risk across various assets and industries. This approach can help investors mitigate the impact of any potential market downturns or overvaluation in specific AI-related sectors.
Moreover, specialist investors who focus solely on the AI domain can enjoy unique competitive advantages. Their deep industry networks allow them to access exclusive investment opportunities and stay well-informed about the latest developments in the field. Additionally, specialist investors benefit from deal pattern recognition, gained through repeat investments in AI ventures.
Endnotes
5. The generative AI landscape: Top startups, venture capital firms, and more , CB Insights, 2023.
Important disclosures
This material has been prepared by Titanbay Ltd and its affiliates (together, “Titanbay”) and is provided for information purposes only. This document is directed at professional investors and qualified investors who have sufficient knowledge and experience to understand the risks of investing in private market investments.
This material should not be construed as legal, tax, investment advice or an invitation, general solicitation, recommendation, an opinion regarding the appropriateness or suitability of any investment strategy, or offer to buy, sell, or hold any investments or securities offered on or off the Titanbay investment platform. The views, opinions and estimates expressed herein constitute personal judgments of certain members of the Titanbay team based on current market conditions and are subject to change without notice. This information in no way constitutes Titanbay research and should not be treated as such. Any forecasts, figures, opinions or investment techniques and strategies set out are for information purposes only, based on certain assumptions and current market conditions and are subject to change without prior notice.
All information presented herein is considered to be accurate at the time of production unless otherwise stated and has been prepared from sources Titanbay believes to be reliable. No representation or warranty or guarantee, express or implied, is given as to the truth, accuracy or completeness of the information or opinions contained herein and material aspects of descriptions contained in this material are subject to change without notice. No reliance may be placed for any purposes on the information or opinions contained in this material. Titanbay is not responsible for any error or omission in this material, nor do we accept liability for any losses arising from its use. Non-affiliated entities mentioned are for informational purposes only and should not be construed as an endorsement or sponsorship of Titanbay.
Investments in private placements and private equity investments via feeder funds in particular, are complex, highly illiquid and speculative in nature and involve a high degree of risk. The value of an investment may go down as well as up, and investors may not get back their money originally invested. Investors who cannot afford to lose their entire investment should not invest. Past performance, including simulated performance, is not a reliable indicator of future performance. For private equity investments via feeder funds, investors will typically receive illiquid and/or restricted membership interests that may be subject to holding period requirements and/or liquidity concerns. Investors who cannot hold an investment for the long term (at least 10 years) should not invest.
Titanbay Ltd is an Appointed Representative of Brooklands Fund Management Limited which is authorised and regulated by the Financial Conduct Authority with firm reference number 757575. Copyright Titanbay 2023.
Comments