Divining the Future of AI

Watch AI venture trends, including companies Nvidia invests in...

When Silicon Valley Bank failed the weekend before March 13, 2023, I predicted that venture funding would slow down, and it did. AI venture funding dropped 38% from Q1 to Q2 in 2023. But, you cannot keep a good investing idea down, and funding has rebounded, both in terms of dollars and number of deals. While much of the focus is on generative AI, there are companies doing (and trying to do) almost anything you can imagine with AI. This article will focus on recent venture funding of AI, Nvidia’s role in some of this funding, and highlight a few interesting companies. We will also discuss the concept of “moats”, or lack thereof around some of these companies.

Executive Summary

• AI venture funding has rebounded in Q3 after bottoming out in Q2.

• The generative AI theme remains strong, but other themes such as AI-driven drug discovery and autonomous vehicles, etc. are represented.

• Nvidia has invested in a significant number of venture deals, becoming a force in the space.

• So far, while many patents have been filed for AI, we have not seen any “moats” from enforcement of those patents. This favors larger companies.

Just the Facts: AI Funding Has Rebounded

AI funding has rebounded since the second quarter, to 823 deals and $12.3 billion through September 15 (see chart), and we still have almost two weeks to go in September. This is close to the peak number of quarterly deals in Q3 of 2021, 983 deals. Clearly, this is a bit of a frenzy, so caveat emptor. There are plenty of anecdotes out there of companies adding larger language models (LLM) to their offerings, even if they do not make sense for the problem being solved, because that is what investors want to buy.

Where is the Money Being Invested?

The top non-debt deals were telling, and each at least $100 million. They are not all about generative AI, which is refreshing given the media frenzy. Themes included generative AI computing platforms, autonomous driving, AI-driven drug discovery, data-oriented firms, as well as AI for the defense industry and a bit of robotics. Here are the top non-debt deals:

1. Conigital raised GBP 500 million (approximately $626 million) in their series A! They automate commercial and industrial fleets (driverless AI, including retrofitting) in closed environments, like seaports, airports, truckyards, and railyards. https://conigital.org

2. Databricks raised $500 million in their Series I for a $43 billion valuation! They simplify data storage through their “lakehouse” data lake structure. https://www.databricks.com

3. Hugging Face raised $235 million in their series D for a $4 billion valuation. The “home of machine learning” facilitates collaboration around AI, running AI, including LLMs, and provide access to specialized GPUs/AI processors in the cloud. https://huggingface.co

4. Helsing raised Euro 209 million (approximately $224 million) in their series B. They are a defense AI company working with governments and industries. https://helsing.ai

5. Imbue raised $200 million in their series B. Their goal is to build AI systems that can reason and code. https://imbue.com

6. Genesis Therapeutics raised $200 million in their series B. The company focuses on AI-powered drug discovery for patients suffering from severe disorders. https://www.genesistherapeutics.ai

7. AI21 Studio raised $155 million in their C round. They provide API access to their Jurassic-2 and task-specific large language models. The startup competes with generative AI powerhouses like OpenAI and Anthropic. https://www.ai21.com

Nvidia’s Role in Venture Funding

Nvidia invests both directly and through their venture arm, NVentures, which was founded in 2021 and officially invests in early stage venture and venture capital. NVentures has reported doing seven deals for a total of $570.4 million so far this year. Just in the last few weeks:

• Enfabrica’s $125 million Series B for. They design networking chips for AI workloads.

• San Francisco-based Imbue’s $200 million Series B that valued the AI research lab at $1 billion. (See above, #5)

• Israel-based AI startup AI21 Labs’ $155 million Series C. (See above, #7)

• New York-based Hugging Face’s $200 million round. (See Above, #3)

Do you detect a theme? Watch this space.

Where Are the Moats?

As an investor, one looks for “moats”, or something that can prevent a competitor from simply copying what you are doing. It could be a big first-mover advantage, access to big capital, trade secrets, an amazing company culture, a founder cult of personality, 10x better technology, or be legal protection, such as a patent.

The number of artificial intelligence-related patent applications in the technology industry was 16,917 in Q2 2023, versus 22,685 in Q1. Sounds like the lawyers are doing great, but I have not seen a news article recently regarding AI patent enforcement or a cease and desist order. It remains unclear if these patents are actually creating a “moat” to protect the intellectual property of the AI companies in question. That leaves the rest of the potential moats above, which may not be present in all companies…

Especially now that generative AI large language models (LLMs) are available though application programming interfaces (APIs), the moats for using AI in applications, websites, etc. have dried up and filled in. for better or worse, this will give the advantage to better funded, faster companies in many cases rather than slower, more thoughtful companies, potentially even those that develop (marginally) better technology. There will always be a place for technology that is 10x better.

What I know is that a lot of companies without moats means that many of the 823 venture deals done this quarter may not be so hot five years from now, but obviously some will, whether through acquisition, IPO, or continued growth. It is an amazing time for AI, but they cannot all be winners.

News of the Week

Apple CEO Tim Cook on Creating a Clean Energy Future [CBS]

Exxon Knew about Climate Change a Long TIme Ago [WSJ]

Elon Musk Calls for AI Referee [NY Post]

Copyright © 2023 by Alec Crawford         

#artificialintelligence #venturecapital #innovation #STAYblog #ElonMusk $NVDA