Hello and welcome to our quarterly LP report on Davidovs Venture Collective. We have just closed our 7th quarter, which means we are approaching our 2-year mark soon. When we began our journey in September 2021, we had only planned our investment strategy for 4 years. Since our rolling fund has an evergreen nature, it's time to conduct a major performance audit, adjust our strategy for the next 2 years, and plan for another 2-year cycle after that.
Our initial investment strategy was developed in the midst of the 2021 bull run, but we planned for a more constrained environment and this turned out to be a good call. The portfolio of the 55 companies we invested in over 7 months looks very resilient, with only 2 out of 55 being written off (about 3.1% of the fund’s claimed capital). The rest have a staggering average of 18 months runway, meaning that almost every company we invested in was either able to make it to their next round of funding or raised so much that they could skip it altogether. It’s still very early to judge what kind of returns we might be looking at, but this is how our team’s performance numbers compare to our own projections from 2021 [this part is available in the confidential version of this report].
In addition to reflecting on our seven quarters of work, we accomplished a lot of exciting things this quarter:
We also worked a lot on our firm and collective’s long-term vision and strategy beyond the rolling fund and are planning to host a call sometime in September to discuss it with the community. Please dial in, we need your feedback and contribution!
Now to our signature quarterly futuristic long read and the State of VC report:
Large language models (LLMs) are truly pivotal. As Yuval Noah Harari said, “Language is the operational system of humanity.” Thus, generative AI powered by LLMs is not just about chat interfaces and the ability of the machine to speak to a human but about creating a universal connector for machines to access anything related to humanity.
We have come close to a moment when humans create autonomous AI agents capable of performing duties of an employee, making decisions, and interacting with both real and virtual colleagues through text. It also becomes possible to automate managerial decisions using AI techniques such as reinforcement learning and Markov Decision Graphs.
So if we can potentially create virtual employees and virtual managers, why can’t we create virtual corporations? Here’s where we come to the concept of autonomous enterprise and technologies aiming at each level of autonomy.
In an autonomous enterprise, defined by Gartner as “a style of business partly governed and majority-operated by self-learning software agents, that provides smart products and services to machine-customer-prevalent markets, operating in a programmable economy,” not only the routine operations are automated but also the business decisions defining those.
While the leaders aiming to future-proof their businesses need to factor this concept into a long-term strategy cycle, an autonomous enterprise is not the endpoint but rather a multi-step journey. It is defined by the nature of the business, its current level of AI adoption, as well as its short- and long-term goals.
Similar to levels of vehicle autonomy adopted by developers of self-driving cars, here’s our attempt to define the scale of the autonomous enterprise, the criteria for each level, and solutions to power them.