![Why is it so hard to get VCs to invest in my Generative AI startup?](https://pic1.calcalist.co.il/picserver3/crop_images/2023/08/07/SkpHeERo3/SkpHeERo3_0_0_2047_1153_0_x-large.jpg)
Opinion
Why is it so hard to get VCs to invest in my Generative AI startup?
By focusing on market needs, technical excellence, and innovative approaches, generative AI startups can position themselves for success in this rapidly evolving field, writes Nir Sabato of Entrée Capital
LLMs and generative AI are probably the biggest tech leap we have seen in many years, and we, like many others, believe they will transform the world in a powerful way. Everywhere you look, this novel technology is gaining traction and drawing much attention.
A Global phenomena?
Just a few weeks ago, during his visit to Israel, Sam Altman said he “strongly believe that now is the best time in history to launch a startup.” By leveraging LLMs, new companies can drive innovation and create significant value over incumbents.
Globally, some Generative AI startups have broken out and are attracting impressive amounts of investment, from enormous Seed funding to significant growth rounds in promising companies like OpenAI, Anthropic, Inflection AI, Cohere, Jasper, Synthesis AI, Stability AI, and others. Venture capital firms around the world are becoming increasingly focused on AI and LLMs.
But beneath these fancy announcements, most Generative AI startups are still battling to raise. Why is this the case?
Generative AI startups' common mistakes
Over the past months, we’ve connected with over 200 early-stage "Generative AI" startups pursuing Seed or pre-Seed investments. Through these discussions, we identified certain recurring factors that, unfortunately, pose challenges in their quest for funding.
Thin layer above the foundational models: One common issue is the lack of deep technical innovation. Many startups rely heavily on existing models without any significant advancements or unique value propositions. This often raises questions about their long-term sustainability and competitiveness.
Limited impact vs. the existing tools: Another challenge is that the potential impact seems marginal on the problem they seek to solve. Today, early stage startups often face direct competition from established players who can easily replicate their offerings (and many of them already try to do so), thus limiting their competitive advantage and attractiveness. e.g. disrupting HubSpot with a GenAI competitor is generally a bad idea as HubSpot can and will add GenAI features and has billions of data points from customers to train its own specific LLMs.
Obvious categories: Furthermore, the market for Generative AI startups is incredibly crowded. With a myriad of companies offering similar solutions, it's hard for startups to differentiate themselves and stand out from the crowd.
Soon-to-be commodity tech: Some of the startups we met planned to train special LLMs or foundational models, with high costs involved. There’s a risk that what you train today gets old tomorrow - and in a year's time you will be better off taking an off-the-shelf model for your application.
What do we look for in a potential investment?
1. First, we look at how well the founder knows their market. Founders who really get their target market and what it needs are often better at understanding the problems their customers have. This deep understanding can give them a special edge in making the best solution. Similarly too, the technical team must have the right experience with AI.
2. We also have a keen interest in startups that are solving big challenges in categories that aren't so obvious. These contrarian areas are where a fresh, creative approach can really set a startup apart and protect it from future competition. In particular, we're really excited about startups that aim for a category where there's no big player dominating. In other words, if you're going after a market that doesn't have a giant that could simply add a similar feature easily, you're more likely to catch our attention.
3. Another key thing we look for is a startup's go-to-market strategy. This plan tells us how the startup will win a slice of the market, attract more customers, and eventually, make a profit. But we're not just looking for any plan - we're interested in innovative strategies, those little secret sauces that make a startup stand out and succeed.
4. Finally, we look for certain characteristics in the founders themselves. We value conviction, intelligence, enthusiasm, truth-seeking, determination, and team spirit. These qualities can make a significant difference in a startup's journey and, in turn, our decision to invest.
To conclude, it's important to remember two key things. First, there's plenty of room in the Generative AI race. We are still in the early days and we firmly believe that many of the great “Generative AI” companies have yet been born. Second, remember that large Seed rounds should not be your goal. As we have learned from the past - it's better to start small, focus on what really matters and provide value to your customers, and only later grow and scale.
Nir Sabato is an Investor at Entrée Capital