Dear readers,
Generative AI (genAI) continued to advance rapidly last month. Foundational models such as the Meta Llama 3 8B and 70B models were released in April, while French company Mistral.ai was speculated to be raising funds at a valuation of over $5 billion. Both companies require large amounts of capital to finance expensive hardware, such as graphic processing units, which explains the high levels of funding required. Additionally, interesting to note that both companies are open-source.
While there is no doubt about the technological advancements in this field, the economics of these models still need to be found to ensure that LLM providers can generate robust cash flows in the long term. Several companies, including Cohere, Stability AI, and Inflection, are struggling, and public investors are increasingly expecting clear guidance on this point. Meta's share price dropped after revealing a significant uptick in capex needs, despite the technological success of the Llama 3 so far. We offer some insights on Mistral's economic engine in this edition.
At the application level, genAI applications continue to receive significant funding, both within software and hardware business models. Robotics, in particular, received widespread attention, with Figure, a company building human-like robots, raising $675 million in Q1, the highest funding last quarter. Vertical applications also remained a bright spot, as smaller LLMs enable companies to leverage their datasets in a cost-efficient way. In total, venture capitalists plowed $2.2 billion into genAI startups according to The Information's genAI database.