The world of AI startups is rapidly expanding, with a multitude of new players and innovative solutions in various industries. According to GlobalData, 3,198 AI startups received $52.1 billion funding across 3,396 VC funding deals during 2022. And while ChatGPT has done for AI what Pokemon GO arguably did for AR - put a tech term on the general public's lips - the AI sector has been skyrocketing for many years.

Several AI companies, including Waymo, OpenAI, UiPath, Automation Anywhere, Zhihu, and SenseTime, received significant investments ranging from $434 million to $3 billion in funding rounds between 2018 and 2021.

Sounds great, yet, Daniil Kirkov at Orbita Venture Studio says there are two things to worry: the unspoken elephant in the room, and the unique challenges of AI startups that require new funding requirements.

Silicon Valley's widely adopted white elephant

While Silicon Valley has produced some of the most successful and innovative companies in the world, depending on which reports you read, Kirikov warns that the industry's lack of success is somewhere between Harvard Business School's 75% failure rate, and the dismal 90% Startup Genome proposed, and usually in the first year.

Having observed many startups from a short distance as a venture capitalist, Kirikov notes that startup ecosystems worldwide have adopted the same approach with varying results, and reportedly with higher failure rates. "Without direct access to the other benefits Silicon Valley offers - resources, mentors, talent, and the network effect created by having so many startups in a relatively small area - it is hard to replicate the original", he adds.

There have been changes in the way VCs select who to fund recently, driven by the global economic crises and the COVID-19 pandemic, "but largely investors are using the same age-old factors when deciding who might be their next winner - and that has to change", says Daniil Kirikov. 

AI startups require new selection criteria

Kirikov suggests looking at a notable study by Harvard Business School professor Paul Gompers and his colleagues, in which they analyzed 1,000 VC investments. The study has set the tone for many pitch decks since, Kirikov says.

Leading the way, 30% of VCs cited the management team as the most critical factor in their investment decisions. A close second at 25% is the market, including size, growth, and competitive landscape. 

Product or service, business model, and financials were cited by a measly 5-15% of VCs as the most critical factors, and that's concerning. Kirikov suggests these are the most crucial questions VCs should be asking when selecting the AI solution to fund. And when these questions are answered, the answers turn those stats on their head, Kirikov maintains.

New deciding factors for a new AI-powered world

"It is critical to dig deeper than simply looking at the problem when assessing the AI startup's solution", says Kirikov. VCs need to start asking questions that go deeper, and assess if the startup is solving an old issue that people were already paying to solve. They should expand on that line of questioning and ask if this issue could have already been solved before the current AI bandwagon. "And could the same solution could have been implemented two, three, or five years ago", Kirikov adds. If the answer is yes to any of these, it may indicate that the startup is not leveraging new technologies or ideas, and others may have already attempted similar approaches without success.

"The inherent risks associated with AI solutions makes it more important to focus on those factors VCs previously felt were less important, or that weren't mentioned at all in the Harvard Business School study", says Kirikov.

One such thing is security. "Security is a vital consideration in the AI startup space, especially as the general public becomes more aware of how companies collect and use their personal data", Kirikov suggests. "As AI technologies become more prevalent and powerful, the potential risks and ethical concerns associated with them increase", he adds.

Kirikov warns that the governments and regulatory bodies are paying close attention to the development and deployment of AI, and startups that do not prioritize security and ethical considerations may face regulatory hurdles and legal challenges. "Similar to the medtech industry, AI startups may need significantly more runway to survive changes in regulation, or to ride out public concerns", he maintains.

Kirikov postulates that the fast-paced nature of AI development, and the resulting boom across the competitive landscape, demand more thorough assessment of the viability of the startup you're considering as a future portfolio company.

In conclusion, Kirikov thinks investors should still focus on every deciding factor, yet, in his opinion, it's time to dig deeper into the problem the startup solves, its business model, traction and growth potential, financials, and risk mitigation. "Don't put so much stock in the team and the market size" he adds. Kirikov is convinced that if such granular assessment becomes the new norm for startup founders and investors, it will propel the success rate we see right across the tech industry.