Weekly Briefing Note for Founders

31st July 2023

This week on the startup to scaleup journey:

  • Europe VC funding down 60% in 1H23
  • AI is transforming how VCs operate

Europe VC funding down 60% in 1H23

As widely predicted, European VC dealmaking was down significantly in 1H23. According to Pitchbook, deal value slumped 60.8% compared with the same period a year ago and was down 34% compared to 2H22. The drop relative to 1H22 was particularly stark as 1Q22 was the peak of the dealmaking frenzy across Europe that began in 2021. The reasons for the decline, largely attributed to the macro environment, are now well understood: High interest rates, high inflation, a closed IPO exit window and difficult fundraising conditions for VCs. The leading indicator for recovery will be public markets, which have shown some early signs of life in recent months, especially in the US. Private markets often lag public markets for a few quarters. In terms of sectors, Software has been the most heavily hit, with deal value dropping by 71.8% YoY in 2Q23, more than any other sector. In fact, software went from representing 40.2% of deal value in 2022 to 29.3% in 2023. But one bright spot highlighted by Pitchbook is AI: "Despite this trend, startups within the software segment that are linked to AI may see significantly more dealmaking in future quarters given the recent investment boom in the space."

Where do we go from here? The key takeaways from the VivaTech event held in Paris a few weeks ago shed some light. More than 150,000 attendees, 450 speakers, 2,400 startups, and 2,800 exhibitors attended, together with many of Europe's biggest VCs and LPs. Listening to LPs, the primary capital providers of the venture ecosystem, is always illuminating. They reiterated their focus on pushing capital towards VCs that have a proven track record, are highly confident in their established investment theses, and do not adjust practices drastically based on near-term market conditions. This perspective sets the tone for the entire industry and underlines why founders must adopt a back-to-basics mindset. In terms of hot sectors, Generative AI was covered extensively given recent developments and deals taking place. Investment in this rapidly emerging sector hit €1.0B across Europe in 2022 and is already past the €0.5B mark in 2023, according to Pitchbook. In recent weeks, France-based Mistral AI secured €105M in one of Europe’s largest-ever seed rounds, having been founded only four weeks prior. Meanwhile, UK-based AI-generated video creator Synthesia obtained €83.6M at a €845.6M pre-money valuation, 4.5x the valuation during its previous funding round in July 2021. Software companies without an AI strategy are going to find it ever harder to grab investor attention over coming quarters.

Investors at Vivatech reiterated that capital is available, but securing it quickly is more difficult for founders. As we highlighted last week, the decrease in both capital investment and valuations represent a 'reset' to the longer term trend, not an existential collapse in VC. Deal terms have clearly shifted in favour of investors, and we know that once more competition is intense between startups seeking capital across all financing stages. The bar in selection and due diligence has risen significantly. We see this almost daily as startups that are struggling to raise capital reach out to us for a second opinion on their approach. Our 'campaign review' incorporates the key considerations highlighted by investors at VivaTech: Do the founders have a deep understanding (real-life experience) of the problem they are trying to solve? Is the solution needed right now? Is it essential? Can product/market fit be sustained? And crucially, given that execution is now just as vital as the product, is there a clear plan to scale and is it working? But the twist is that it's as much to do with the 'fit' of the investors being targeted as the proposition itself. In 2021 and early 2022, many investors strayed 'off-piste' in terms of both sector and stage - then found themselves in difficulty. Finding investors that have strong alignment in their core investment theses - as well as immediately deployable capital - has once more become a vital element in any funding strategy.

AI is transforming how VCs operate

In 2021, Ali Tamaseb, General Partner at DCVC, published his acclaimed book, Superfounders. His findings, based on a manual study of tens of thousands of company data points, revealed a series of remarkable insights into what makes startups successful. Only a few years later, AI and data-driven approaches are now revolutionising VC decision-making. Insights that once took thousands of hours of research are now reaching the fingertips of VCs in seconds. This is rapidly changing how investors assess investments, from deal sourcing and screening, due diligence, post-investment management and, ultimately, to exit strategy. Dr. Andre Retterath, Partner at Earlybird Capital, recently published the Data Driven VC Landscape report. This analysis looks at how the VC community is embracing the latest tools in AI and machine learning to enhance investment efficiency, optimize portfolios, and unlock higher returns. The report concludes that using AI and rethinking the approach to allocating capital is no longer a ‘nice to have’: it’s a necessity to remain competitive as a VC. And many others agree. Gartner stated recently that, "By 2025, more than 75% of VC investor reviews will be performed using AI and data analytics".

The first big win for investors adopting AI is efficiency. Notoriously time poor VCs can decrease the time needed to perform a broad range of tasks, increasing their output quality, and allowing smaller teams to achieve more. AI also improves effectiveness: Returns in venture capital follow a power law distribution. This means there is a very high cost to missing outliers. Luckily for VCs, machine learning models have been found to outperform human investors in screening. By using an automated data capture & screening system to monitor and triage through news, social media, and other cutting edge research, VCs are getting a higher deal coverage and lowering their miss rates. A recent Sifted article described how data-driven sourcing is helping European VCs use various 'signals' about talent and companies to find and engage with startups before the competition, stating; "That’s different from methods that VCs have traditionally used to scout founders, such as personal referrals." This is also leading to a higher level of inclusiveness; "As a result, with more and better informed deals, the wide disparity in funding seen across the world can begin to narrow. Less than 20% of investors are women, and female founded companies represent 2% of global deal volume; despite women investors and founders being shown to outperform the broader market. Data-driven initiatives help reduce bias and make better investment decisions."

At the core of the tech stack for VC investors are the startup research databases, such as Crunchbase, CBInsights, and Dealroom, to name just a few. Vendors are working quickly to figure out how to further embed AI and data analytics into their platforms. Pitchbook, one of the highest-ranked according to an in-depth study, has recently launched the 'Exit Predictor', a new tool that leverages machine learning and their vast database of information on VC-backed companies, financing rounds, and investors. This capability provides objective insights into startups’ prospects of a successful exit. The primary component underpinning the tool is a classification model developed by PitchBook's Institutional Research Group that predicts the probability that a VC-backed startup will ultimately be acquired, go public, or not exit due to failure or becoming self-sustaining. Key data points leveraged include; Company details (e.g. Patents, industry, employee count, news coverage, number of acquisitions); Deal activity (e.g. Maturity, fundraising frequency, average deal size), and Active investors (e.g. Investor track record, number of crossover investors, number of investors). The model was initially trained on 46,000 observations from startups with known outcomes. The algorithm was then tested on more than 11,000 out-of-sample observations and accurately predicted success (M&A and IPO) versus no exit outcomes at a rate of 75%, much higher than what most VCs could honestly claim to be their hit rate from 'manual' methods.

Happy reading!

Subscribe to our mailing list

Stay informed. We will email you when a new blog post is published.

* indicates required

To subscribe to our Newsletter click here