Several AI startups developing applications based on large language models (LLM) are rapidly increasing sales and initiating a new race to commercialize cutting-edge technologies. Their rapid growth has attracted the attention of investors willing to invest hundreds of millions of dollars in developing consumer AI products.
Image source: Alex Shuper / Unsplash
Investors are betting on companies like Cursor, Perplexity, Synthesia, and ElevenLabs. These startups are building applications based on powerful generative AI (GenAI) models from OpenAI, Google, and Anthropic. They are helping to drive broader adoption of rapidly evolving technologies in both consumer and enterprise settings.
According to analytics platform Dealroom.co, the volume of funding for startups developing AI applications will amount to $8.2 billion in 2024, up 110% from 2023. This investment frenzy indicates high interest in AI tool developers, who can attract hundreds of millions of dollars amid rapid growth in demand.
AI search startup Perplexity raised $500 million in December in its fourth funding round of the year, tripling its valuation to $9 billion. The company is now in talks for another round at a significantly higher valuation, according to sources. Meanwhile, legal AI startup Harvey raised $300 million in February.
Startups building apps for software developers have also attracted increased investor interest. Companies like Reflection AI, Poolside, Magic, and Codeium have raised hundreds of millions of dollars in 2024 to develop technologies aimed at improving programmer productivity. In January, Anysphere, maker of the Cursor programming automation tool, raised $105 million at a $2.5 billion valuation. Investors are interested in the company at a valuation of $10 billion or more, sources said. The three-year-old startup has already achieved annual recurring revenue of $200 million.
Image source: Dealroom.co and Flashpoint
AI startups have also benefited from increased competition in the LLM market, which has driven down the cost of processing queries and generating AI responses. This has allowed them to use the LLM infrastructure without having to build their own expensive AI models, speeding up time to market.
Bret Taylor, chairman of OpenAI and co-founder of Sierra, noted that the company has changed its AI models at least five to six times in a short period of time due to the rapid pace of change in the industry. Sierra, a startup that develops AI-powered customer support agents, was founded in February 2024 and reached a valuation of $4.5 billion in October of that year. According to him, using a two-year-old AI model today is comparable to driving a car from the 1950s – that’s how quickly technology becomes obsolete.
According to an analysis of payments data from fintech company Stripe, the largest AI companies are achieving millions of dollars in sales within their first year of operation. This is happening significantly faster than startups in other tech industries, demonstrating the power of applied AI to quickly create sustainable business models.
But it’s hard to gauge how robust the customer base of AI startups is, or how sustainable their current revenues will be. With all the hype around AI, early adopters are coming in fast, which can skew growth numbers by not guaranteeing long-term subscriptions. Some investors are avoiding the race for the hottest apps, fearing that even the best ones are just service wrappers for existing AI models.
There is a risk that such startups will be displaced if a larger company with a large user base decides to replicate their functionality. Hannah Seal, a partner at venture capital firm Index Ventures, which has invested in legal AI assistant Wordsmith, points out that many of these AI startups have not yet gone through a full annual subscription renewal cycle. So the level of customer churn remains unknown and could significantly impact future growth.