IPOs in 2020 and the Generative Tech Market: Exploring the Intersection of Investment and AI

IPOs in 2020 and the Generative Tech Market: Exploring the Intersection of Investment and AI

In 2020, the IPO market saw a mix of success and disappointment. While only 25% of companies ended the day trading lower than their IPO price, over 25% of companies witnessed a significant increase, trading more than 50% higher than their IPO price. This phenomenon, known as the "IPO pop," has become a common occurrence in recent years. Out of the 61 IPOs in the US in 2020, the median company experienced a pop of 20% on the first day.

However, despite the excitement surrounding IPO pops, it is important to note that these companies could have raised significantly more capital if their IPOs were priced at what the market valued them. In fact, in aggregate, the 61 companies that went public in 2020 raised $6.7 billion less than they could have. This discrepancy arises from the desire of institutional investors to acquire stocks at a lower price in order to maximize their returns on investment.

While the IPO market showcases the investment side of the tech industry, there is another growing trend that is revolutionizing various sectors - the emergence of generative technology. The generative tech market encompasses a 5-layer tech stack, with each layer contributing to the development of AI models for specific applications.

At the core of this tech stack are general AI models, such as GPT-3 for text, DALL-E-2 for images, Whisper for voice, or Stable Diffusion. These models have the capability to generate outputs in broad categories like text, images, videos, speech, and even games. However, to capture more nuance and specificity, specific AI models are trained on more narrow and specialized data. These models excel in tasks such as writing tweets, ad copy, song lyrics, or generating e-commerce photos and 3D interior design images.

One layer above specific AI models lies the hyperlocal AI models. These models are specialists in their respective fields, capable of producing outputs tailored to individual preferences. For example, a hyperlocal AI model can write a scientific article in the style preferred by Nature, create interior design models suited to a specific person's aesthetic, or even write code in the particular style of an individual company.

While proprietary and trusted data can offer a certain level of defensibility, it is important to note that data network effects tend to plateau over time. Competitors can find similar datasets and produce models that may not be as good but are difficult for customers to distinguish from the original. In the near future, AI writing will become indistinguishable from human writing, and most people will enjoy music and lyrics written by AI. Therefore, relying solely on data network effects may not be a sustainable strategy.

To create a strong defensibility and capitalize on data network effects, it is advisable to focus on the hyperlocal layer. This layer benefits from proprietary and trusted data, providing an opportunity to differentiate from competitors. However, the generative tech market is evolving rapidly, and the API layer or Generative OS plays a crucial role in accessing and switching out AI models as needed. This layer allows for the creation of countless applications catering to various needs, commodifying AI models to a certain extent.

In this fast-paced market, speed is of the essence. Product speed, fundraising speed, and sales speed are key factors that can determine success. Launching a feature quickly and allowing the model to learn over time is often more effective than spending excessive time hunting for specific data to build the perfect model. It is important to get the product in the market, observe what works and what doesn't, and address any customer discomfort. Keeping a close eye on competitors and borrowing the best ideas can also contribute to success.

Aggressive sales strategies can help embed the product in customers and create network effects, ultimately strengthening defensibility. Finding an investor who is willing to sprint alongside the company can provide the necessary support and resources for rapid growth.

In conclusion, the IPO market of 2020 highlighted the desire for institutional investors to acquire stocks at a lower price, while the rise of generative technology in the tech market presents opportunities and challenges. By understanding the dynamics of IPOs and leveraging the different layers of the generative tech market, companies can navigate this evolving landscape. Prioritizing speed, embracing sales strategies, and finding the right investor can contribute to success in this competitive industry.