A lot of 2023’s new unicorns have been generative AI startups. Here’s what to expect from the sector next year, according to Accel.

• A lot of 2023’s new unicorns were generative AI startups, according to a new report from Accel.
• Around 60% of all new $1 billion SaaS and cloud startups have been “GenAI native.”
• Investors have pumped $18 billion into GenAI startups so far this year, up nearly five times on 2022.

This year’s new unicorns have been dominated by generative AI startups.

Since the launch of ChatGPT-3 in November 2022, venture capitalists have poured billions of dollars into young startups working on the game-changing technology. As the hype surrounding generative AI grew, high-profile investors rushed to back, in some cases, weeks-old startups like Mistral.

According to the latest Euroscape report from Californian investor Accel Partners, the US has already produced nine generative AI startups worth more than $1 billion this year, while Europe and Israel have produced three. Despite a backdrop of low venture capital funding, widespread layoffs, and valuation slumps in the technology sector, the new high-value companies have emerged.

Investors poured a record $18 billion into generative AI startups in 2023, nearly five times the $3.9 billion invested in 2022.

According to Accel’s report, roughly 60% of new unicorns in the SaaS and cloud sectors globally were described as “GenAI native.” In the United States, where companies like Anthropic, Jasper, and Runway have exceeded $1 billion in valuation this year, the proportion was 75%. This figure fell to 45% in Europe and Israel, home to newly minted unicorns such as Synthesia and AI21.

Both regions also have a slew of promising unicorns in the generative AI space. According to Accel’s report, these companies were involved in three key verticals: application, infrastructure, and foundational models.

According to Phillipe Botteri, partner at Accel, Gen AI is a “fundamental trend” that is redefining the potential of software, and he expects “more to come.”

According to the report, France has emerged as a hotbed for generative AI, with French startups closing the top three largest funding rounds in Europe, including Mistral’s massive $113 million seed round.

However, Europe has a long way to go before catching up to the United States in terms of later-stage funding. While generative AI startups in Europe received $900 million across the continent’s seven largest deals, their counterparts in the US received a whopping $14.1 billion — including OpenAI’s massive $10 billion round, according to the report.

Nonetheless, the EU and Israel compensate with a larger pool of AI talent, producing 50% more AI publications than the US, according to the report. According to a Stanford University report, researchers from the EU and UK contributed to 15% of all AI journal publications from 2010 to 2021, compared to 10% in the US.

Generative AI will become a commonplace tool.

“Generative AI will unlock new verticalised applications built with smaller and dedicated models and industry-specific workflows,” according to the report published by Accel. These startups, such as Harvey and Hippocratic AI, typically require niche datasets to train their models. According to Accel, sectors such as healthcare, legal, and drug discovery have the greatest potential for using AI for specific use cases.

The report also predicted that generative AI-driven media creation will become more common as tools for synthetic video, image, and voice generation become more widely available. Startups such as ElevenLabs, Synthesia, Stability AI, and Runway have risen to prominence for providing these services, and the report predicts that they will soon become mainstream across personal and business use cases.

According to the report, enterprises will also rush to incorporate generative AI more seamlessly into their automation tools. AI is expected to streamline enterprise activities ranging from document management and communications mining to content creation. Microsoft, UiPath, and Celonis, among others, “are leveraging proprietary AI and third-party large language models (LLMs) to address a broader range of enterprise use-cases,” according to the report.