Despite global declines last quarter, venture capital grew modestly in Europe
Seed-stage deal size ramped up at the end of 2023, while early- and late-stage declined
Corporate venture capital activity ticked up in the fourth quarter, thanks to AI
Capital-intensive large language models have fueled a boom in foundational technology investments
Global venture capital funding fell 15% quarter over quarter in the final months of 2023, hitting the lowest level since early 2020 (see Figure 1). Following slight growth between the second and third quarters, investments in the US and China declined in the fourth as the countries navigated continued economic uncertainty.
Despite overall declines, there were some bright spots: Funding in Europe grew modestly at 7% quarter over quarter, thanks to an influx of investments in artificial intelligence (AI) and fintech start-ups. The US secured nearly half of worldwide venture funding primarily due to significant generative AI funding rounds. Semiconductor start-ups were similarly successful, buoying investment in China.
While the average size of early- and late-stage deals dipped by 9% and 16% quarter over quarter, respectively, seed-stage emerged as the most robust funding stage, growing by 17% (see Figure 2). Given the lower barriers to entry in the application side of generative AI, more startups are sprouting, and investors are taking advantage of the opportunity.
Deals involving corporate venture capital (CVC) accounted for 20% of all funding in 2023, a steady share despite the overall funding downtrend. Both CVC-funded deals and the number of CVC investors ticked up in the fourth quarter of 2023, due to strong participation in AI-focused funding rounds (see Figure 3).
Indeed, generative AI start-ups defied 2023 investing trends with an upswing in funding. Seemingly resilient to investor hesitance and macro uncertainty, the sustained interest in generative AI is a testament to the technology’s transformative potential across industries. Investments in companies that build foundational technologies, including large language models (LLMs), surged last year and show no signs of stopping (see Figure 4).