Despite generating substantial revenue, few generative AI startups are profitable due to the high upfront investment required to build flagship models. Big tech investors like Google, Amazon, and Nvidia are playing the long game, viewing generative AI investments as strategic bets. However, if these startups cannot overcome the challenges they face, there is a possibility that the investment bubble could burst.
Key takeaways:
- Investments in generative AI startups are being consolidated into a smaller number of early-stage ventures, with 225 startups raising $12.3 billion from VCs in the first half of 2023.
- Early-stage startups, such as Elon Musk’s xAI and China’s Moonshot AI, were the clear winners, raising billions of dollars. However, the fate of many generative AI businesses remains uncertain due to potential legal and regulatory challenges.
- High-quality training data is becoming harder and more expensive to obtain, and the process of training models is also becoming more costly. For instance, OpenAI’s GPT-4 cost $78 million to train, while Google Gemini’s price tag was $191 million.
- Despite generating substantial revenue, few generative AI startups are profitable. For example, OpenAI, which is reportedly generating around $3.4 billion in revenue, could end up losing $5 billion this year. This suggests that investors in generative AI are playing the long game, viewing these investments as strategic bets.