Funding future generations

Summary

• We uncover sustained growth in venture capital (VC) into startups across Europe, the US, Australia and New Zealand targeting under-fives (known in edtech as pre-K) using novel data science and machine learning methods.
• VC investment into pre-K has more than tripled over the past decade and the UK is leading the way in Europe.
• Special educational needs is one of the most rapidly growing areas of investment, increasing 9x from 2018-2022.
• AI has the potential to radically transform the pre-K sector, with highly personalised learning and automated operations likely to attract significant investment.

Introduction

What are the biggest trends in investment into ventures targeting under-fives? We analysed venture capital (VC) investment data across 33 English-speaking and European countries for the pre-K sector to understand what is attracting investment and how much, before considering where future investment opportunities may lie. 

We took VC funding data from Crunchbase and applied data science and machine learning methods to select relevant companies based on their descriptions, classify them into categories and characterise their investment performance. For data quality and coverage considerations, we limited the scope of our analysis to companies based in Europe, US, Canada, Australia and New Zealand. Overall, we identified 1,439 companies across seven main themes.

Investment in pre-K has grown steadily over the past 10 years

Investment into pre-K has grown 3.4x over the past 10 years, from £201 million in 2012 to £690 million in 2022 (Figure 1). We saw a record year in 2021 during a Covid-19 induced edtech boom that has since levelled off. 

Investment into edtech overall grew by 5x over the period and reached £3.43 billion in 2022. While the growth of pre-K investment was lower in comparison, we find that it represents a significant part of the edtech ecosystem. 

Ventures are now taking advantage of the tailwinds the sector is experiencing, including increasing numbers of dual-career families driving demand for childcare, rapid digitalisation of both the provision of and the operations sitting behind the delivery of childcare, and growing pressure on employers to support employees with caring commitments. 

Mapping the landscape of pre-K ventures

We used data science and machine learning methods to assign a theme and sub-theme to each company in the sector. The resulting landscape of pre-K businesses is visualised in Figure 2, where each circle corresponds to one company and companies with similar descriptions are situated closer to each other. 

Scroll to explore the landscape, noting that some companies will be assigned more than one theme and that some have ceased trading but are included as they feature in the historical trends. 

Content and family support themes attract the most investment

The distribution of investment across the themes over the past five years is shown in Figure 3. We find that content is the most heavily invested theme, having attracted £1.7 billion in funding and the majority of £100 million+ deals (with mega rounds raised by Elemy, Jam City, Age of Learning, and Moonbug), closely followed by family support, with £1.55 billion invested. 

Using our unique sub-theme categorisation (switch the toggle in the interactive chart below to see the sub-themes) we have been able to observe that, within the content theme, 51% of funding has gone towards general content, 34% into play, and the remainder is spread across the more specialised sub-themes of literacy, numeracy and creative. Within family support, 56% of funding has gone towards products and nutrition and 20% has gone into general family support products and services. 

High-growth themes include special educational needs

While special educational needs theme came third in our ranking by investment amount, it has seen the fastest-growing investment over the period 2018-2022 (Figure 4), with the amount invested increasing by 9x against a backdrop of edtech investment growing around 2.5x in the same time period.

This is a promising signal that ventures are catering for children with a wide range of needs and indicates strong potential for this theme as an investable area.

Other high-growth areas include prepartum (which includes preconception and antenatal) and finances sub-themes of family support, which have grown by 8.7x and 6.9x respectively and the operations theme which has grown more than 3x. 

UK leads among European countries

While the vast majority (85%) of VC investment in pre-K from 2018-2022 has been invested in US startups, the UK has led the way in Europe, taking a 60% share of pre-K investment in the region (use the toggle in Figure 5 to see a Europe-only view).

This is in comparison to a 35% share of European edtech investment more broadly, showing a relative over-weighting in the category. We’re optimistic that these findings show there is scope for investment activity in the UK to grow, taking inspiration from what’s happening in the US and building on our existing leadership position to support a vibrant pre-K startup scene. 

What’s next: can AI drive investment into this space?

These trends up to the present day are promising. But where are we heading? As active investors in edtech, Nesta is optimistic about the growing number and variety of startups building in pre-K and we are particularly interested to see the opportunities that will come from a technology that has received an incredible amount of attention in recent months: artificial intelligence (AI). 

Over the last decade, pre-K companies integrating AI, machine learning and related technologies have seen a substantial growth in investment (Figure 6). Our earlier research on parenting tech trends sheds some light on innovative applications of more traditional forms of AI for supporting children and parents. For example, speech detection technology is powering learning-to-read apps such as Ello, language learning with the Buddy.Ai personal tutor and screening for special educational needs as delivered by EarlyBird Education.

Could the recent surge in generative AI tools such as ChatGPT and Midjourney, exhibiting exceptional capabilities in producing text, visual and audio content, boost the investment growth further and open new avenues for innovation and investment in the pre-K space? Very possibly.

For example, it might reduce the costs and effort required to generate captivating text, visual and audio content – content being one of the largest themes in the pre-K ventures landscape. We've already seen the initial applications – and controversies – of generative AI in crafting children’s books, including creating personalised stories with your child as the main character, as by Storywizard.Ai, and generating voice narrations, as by Typecast.Ai. 

Generative AI could also help streamline operations and administrative tasks in childcare settings, for example by expediting writing tasks and answering emails. It could also be used to support time-strapped childcare professionals with writing lesson plans and creating new activities and resources. 

With the technology expected to only become more powerful and easier to use, pre-K companies could leverage it to enhance their products and services, as well as benefit from the current enthusiasm surrounding AI to attract more investment. Nonetheless, everyone should remain mindful of the present limitations and risks of generative AI, especially for vulnerable users.

Nesta will continue to explore this space, with Nesta’s Discovery Hub embarking on a new project to dive deeper into generative AI and our Mission Studio, in partnership with Founders Factory, ideating and spinning out new ventures in this space. 

Nesta Impact Investments will continue to invest in the pre-K sector and other edtech opportunities that narrow education disadvantage. We are keen to talk to any founders who are seeking scale-up funding.

Nesta’s mission goal

Nesta’s fairer start mission is to narrow the outcome gap between children growing up in disadvantage and their peers. Nesta’s thesis is that intervention in the early years has the highest impact on narrowing this gap. Across Nesta’s teams we draw on a range of tools including research, experimentation, data and behavioural science, developing innovation partnerships with government, industry and sector specialists. We back private sector innovation by investing in early stage tech startups through Nesta Impact Investments.

Important information
This document is not an investment recommendation or financial promotion; it is an experimental analysis of general trends around companies working on products and services supporting families of young children and associated venture capital investment. The companies, products and services named within this article are cited purely as examples which help to demonstrate the breadth and depth of products available and their inclusion should not be interpreted as an endorsement by Nesta. We make no representations as to the data security, reliability or otherwise of any of these products or services outlined within and readers should refer to the separate privacy notices for particular providers, as appropriate. None of this content should be construed as parent/carer advice or solutions and readers should make their own decisions as to whether a particular app or service may be suitable for them or their children.

Methodology

Identifying and segmenting pre-K companies

In this study, our approach was founded on a data extraction and machine learning process to provide data-driven insights into market trends. We sourced data from the well-established Crunchbase business information database, enabling us to access data on VC investment. The company information was further enriched by considering companies in the HolonIQ database as well as from our previous research. We carried out an extensive text search in Crunchbase using industry tags and keywords pertinent to the child care industry. A detailed list of these keywords can be found in the linked table.

With this data at our disposal, we turned to machine learning techniques to streamline the identification of relevant companies and their subsequent segmentation into themes. Our initial step involved manually reviewing a subset of the companies as well as applying unsupervised machine learning (clustering) to yield an initial theme segmentation. This allowed us to spot groupings of irrelevant companies to exclude from further investigation.

We then leveraged supervised machine learning to leverage judgments from our manual review and sift out startups of relevance from those that were not. Following this automated sorting, we selected all potentially relevant startups to be labelled by ChatGPT API according to our predefined taxonomy of themes and sub-themes. The prompt that we used can be found here.

To further enrich the analysis, we trained another supervised machine learning model to tag any company text with the themes, using ChatGPT-derived results as the training data. Although given the scale of the data, we could have relied on ChatGPT for all our labelling needs, employing supervised machine learning offers advantages for future work, such as scaling up the labelling to millions of companies as well providing confidence estimates for each assigned label. 

We manually adjusted the assignment into themes and sub-themes, with particular attention to distinguishing separate categories such as Health, and a sub-theme for Family Support: Products and Nutrition.

Our methodology incorporated a comprehensive manual review process. We verified the relevance of a substantial proportion of the identified startups and assessed their assignments to sub-themes, manually correcting certain assignments as necessary. The final outcome of this process can be seen in the landscape visualisation in Figure 2.

Despite the meticulous approach taken, we must note that due to the semi-automated nature of the analysis, occasional errors in the assignment of companies to themes and sub-themes might have occurred. Similarly, our search approach may have missed certain companies. Therefore, we advise interpreting the results as data-driven hypotheses of market trends.

In the subset of AI-related companies (Figure 6) we included pre-K companies that have been tagged on Crunchbase as associated with artificial intelligence, machine learning, speech recognition, natural language processing, predictive analytics or virtual assistant industries.

The countries included in our analysis were Austria, Belgium, Bulgaria, Croatia, Cyprus, Czechia, Denmark, Estonia, Finland, France, Germany, Greece, Hungary, Ireland, Italy, Latvia, Lithuania, Luxembourg, Malta, Netherlands, Norway, Poland, Portugal, Romania, Slovakia, Slovenia, Spain, Sweden, United Kingdom, United States, Canada, Australia and New Zealand.

Acknowledgments

We thank Lisa Barclay, Louise Bazalgette, Alexandra Burns, and Emma Neale for reviewing the article and providing helpful advice. We also thank Siobhan Chan for editing and designing the article, and Grace Fothergill, Milly Butters and Kieran Lowe from the Communications team for their support.