Chain Venturer: Spencer Farrar of Theory Ventures

📢 Sponsor | 💡 Telegram | 📰 Past Editions

Happy Weekend 🙋🏻‍♂️,

Welcome to Chain Venturer, a series of intriguing conversations with crypto investors. This week, we have Spencer Farrar from Theory Ventures.

Theory Ventures is a data fund that focuses on investments in data infrastructure, data pipelines, machine learning, AI, and crypto. Farrar dedicates his expertise to the intersections of crypto and AI, balancing his efforts evenly between these two pivotal areas.

The firm operates under the belief that the traditional IPO market has become less accessible and efficient over time, leading to prolonged periods before companies go public and diminishing investor returns. They view the rise of crypto tokens as a potentially more efficient alternative to traditional IPOs, noting the high cost of IPOs and the broad accessibility that token offerings provide to a global online audience.

If you would like to participate in a future episode, DM me here.

-Marco

In case you don’t know…

I’m an investor at Primitive Ventures. We back founders who believe that crypto represents the new socioeconomic primitive that will redefine individual sovereignty and financial paradigms.

If you’re building an interesting project → DM me here.

If your project is too early in the lifecycle, I also run a syndicate, you can check it out here.

Anyway, enjoy this week’s conversation.

Conference season sooon

Spencer Farrar of Theory Ventures

Spencer Farrar is a Partner at Theory Ventures,  an early-stage venture capital firm that invests $1M-25M in software companies that leverage technology discontinuities into go-to-market advantages.

Farrar began his career in structured finance at Velocity Financial ($VEL). By 2021, he had advanced to a Partner role at Bienville, an asset manager overseeing $4B in assets. Prior to joining Theory Ventures, he served as the Chief Strategy Officer at a startup focused on developing derivatives trading infrastructure.

Farrar graduated from the University of California, Berkeley, where he studied Cognitive Science. During his time at Berkeley, he distinguished himself as an NCAA Champion, while earning multiple academic honors and receiving a bronze medal for Team USA.

Here’s my conversation with Spencer Farrar.

Quick takeaways:

• Spencer’s introduction to the crypto world occurred in 2016 during a drive over San Francisco's Bay Bridge with friends discussing Bitcoin and Ethereum. This conversation sparked his curiosity, leading him to invest in Ethereum and deeply explore the crypto rabbit hole.

• Theory adopts a concentrated investment approach with a portfolio limited to 12-15 investments. This strategy is based on rigorous analysis, aiming to simplify management and optimize returns.

• Spencer views all crypto networks as decentralized databases, emphasizing their unique functionalities such as privacy and computational guarantees despite higher transaction costs compared to traditional databases.

• Spencer notes a division between monolithic and modular architectures within the industry. He has observed significant traction toward both and attributes unique strengths to each approach.

• Spencer is cautiously optimistic about the intersection of crypto and AI, particularly in areas like computational integrity and managing royalties for AI model contributors.

What was the defining moment that drew you into the world of crypto?

Farrar's entry into the crypto world began during a casual drive over San Francisco's Bay Bridge in 2016 with friends, including an engineer and a lawyer. The conversation that day focused on the potential of Bitcoin and Ethereum, sparking Farrar's interest due to his existing curiosity about cryptocurrency.

His friend, the lawyer, was particularly enthusiastic about Ethereum, highlighting its ability to execute smart contracts and automate agreements without intermediaries, solely through code. This idea appealed to Farrar.

The defining moment: Inspired by this discussion and further personal research, Farrar decided to invest in Ethereum, purchasing it when its price ranged around $20. This investment marked the start of his deeper exploration into the crypto world.

Farrar then spent much of his weekend during his undergraduate study learning about crypto projects. He’d spend hours reading online forums, reviewing investment decks and networking with founders in San Francisco. While at Berkeley, many of his peers left school to work on projects like Cosmos. One of Farrar’s friends at Berkeley was Sunny Aggarwal, the co-founder of Osmosis Labs. This environment and his journey into trading and investment laid a strong foundation for his future in the crypto space.

Farrar's fascination with the crypto world deepened as he navigated his early career in structured credit. His dedication became so consuming that it even drew the attention of his workplace's IT department, a moment that underscored his need to shift towards crypto full-time. Recognizing his passion and the potential of the burgeoning field, Farrar took a significant step by transitioning to a role that allowed him to manage capital for a New York City hedge fund, specifically focusing on cryptocurrency investments. This decision not only marked a serious commitment to his new career path but also positioned him at the forefront of financial innovation, eventually leading him to become an early investor in groundbreaking projects like SUI and to establish connections with key figures in the industry, such as Evan Cheng.

What is Theory Ventures?

Theory Ventures is a data-driven venture capital firm that focuses on maintaining a highly concentrated investment strategy. Farrar joined Tomasz Tunguz at the firm's inception, shortly after the two successfully co-invested in SUI, which laid the foundation for their strong professional relationship.

The philosophy behind Theory Ventures is shaped by a critical analysis of current market trends. Many venture capitalists adopt an index approach, spreading investments across a broad portfolio with the hope that one will yield significant returns.

However, through rigorous analysis, including a Monte Carlo simulation conducted with a Stanford professor, Theory Ventures concluded that a portfolio limited to 12 to 15 investments would improve portfolio management and optimize return profiles. This strategy allows them to acquire a significant ownership stake in each company they invest in, enhancing potential returns.

Theory Ventures is a data fund that focuses on investments in data infrastructure, data pipelines, machine learning, AI, and crypto. Farrar dedicates most of his time focusing on crypto and AI, balancing his efforts evenly between these two pivotal areas.

The firm operates under the belief that the traditional IPO market has become less accessible and efficient over time. With the huge influx of capital into venture capital, software businesses are staying private longer leading to prolonged periods before companies go public and diminishing investor returns. They view the rise of crypto tokens as a potentially more efficient alternative to traditional IPOs, noting the high cost of IPOs and the broad accessibility that token offerings provide to a global online audience.

In summary, Theory Ventures is a highly concentrated, thesis-driven firm that commits substantial resources to deep research, having published thousands of pages of research in the previous year alone. They focus on specific areas that align with their investment thesis, aiming to maximize their influence and returns in a carefully curated portfolio.

Can you elaborate more on what you mean by decentralized infrastructure as a database?

Farrar explains that all crypto networks function as databases. Given Tomasz Tunguz's extensive experience with traditional databases, such as his investments in Snowflake and DuckDB, Theory applies a similar analytical approach to understanding crypto databases' unique functionalities and utilities.

Farrar notes that while the cost to transact on platforms like Ethereum has dramatically decreased—achieving a tenfold improvement in recent years—it remains substantially higher than traditional databases like PostgreSQL. Even with advancements in blockchain technologies like Sui and Solana, which allow transactions for a fraction of a cent, the costs are still a thousand times higher than using Web2 databases such as BigQuery or AWS Aurora, the latter being 1000x cheaper than the most efficient crypto databases today.

However, Farrar emphasizes that the higher transaction costs of crypto databases do not necessarily detract from their value. These platforms may offer unique advantages such as privacy guarantees or computational guarantees. For instance, using zero-knowledge proofs to verify transactions in a decentralized manner could be crucial for specific use cases or utilities. Therefore, Theory Ventures analyzes these platforms as potential databases with distinct functionalities.

Farrar points to innovations like Cribl that enable midstream processing, an architectural advantage that can greatly reduce log analysis costs. This means it processes data between the source (where it's generated) and the destination (where it's ultimately stored and analyzed). This company achieved rapid growth by optimizing how data is handled before it reaches traditional databases like Splunk. These types of architectural innovations are what Theory and the team look for as an initial wedge into the market.

This efficiency is analogous to what he sees in decentralized databases operating across multiple computers globally, potentially reducing costs and improving data handling efficiency. In an ideal scenario, these decentralized systems could provide back-end services without being public-facing, offering Web2 developers a reliable alternative with enhanced security and privacy that traditional databases might not provide.

In summary, Farrar views decentralized databases as a technical infrastructure and a strategic asset that can offer significant advantages over traditional Web2 databases in terms of cost, security, or privacy.

What are your high-level thoughts or observations on why developers are increasingly interested in the Move language and Sui's Move VM?

Farrar finds Move appealing primarily because of its simplicity and excellent developer experience. Unlike traditional Web3 development environments, Move uses an object model that is more intuitive for developers accustomed to such frameworks. In contrast, other smart contract languages, such as Solidity, often require developers to represent objects and data through complex mappings.

Move is still a nascent language, and while it may take some time to achieve widespread adoption, Farrar believes it stands a strong chance of becoming the preferred language for smart contract development. He contrasts this with the Ethereum Go client, which is not designed modularly and involves a steep learning curve due to its intricacies.

Farrar is excited about Move's prospects because it streamlines the learning and development process, potentially speeding up the adoption curve. However, he acknowledges that reaching critical adoption is a significant challenge for any new technology, framing it as the central issue that Move needs to address.

Do you lean more towards a monolithic or modular approach in the context of the increasingly fragmented on-chain ecosystem? If you favor modularity, how do you propose to aggregate the different pockets of liquidity?

Farrar articulates his views on the evolving landscape of blockchain technology, noting a division between monolithic and modular architectures within the industry. He has observed significant traction toward both and attributes unique strengths to each approach.

Drawing from his firsthand experience, Farrar reflects on his time working with Solana during its early stages and at its first hackathon. He appreciates Solana's performance despite initial challenges, such as compute limits and occasional network downtimes. He favours it for his transactions, considering it one of the most efficient networks available.

In discussing the modular approach, Farrar considers the complexities developers face when integrating multiple components within fragmented ecosystems. He highlights the potential of roll-up-centric models to facilitate smoother interactions and liquidity management across various platforms. A project that has captured his interest is Initia, which allows for flexibility in using different virtual machines (VMs) like EVM, Move, and WASM, underpinned by a unified infrastructure. This setup enables composability and native communication among different roll-ups, potentially addressing the issue of liquidity fragmentation.

Farrar explores the concept of shared infrastructure as a viable solution for enhancing interconnectivity between different systems, particularly through mechanisms like shared sequencers or using the Inter-Blockchain Communication (IBC) protocol, as seen with projects migrating to Cosmos for its robust asset flow and communication capabilities.

Moreover, he acknowledges the security concerns of new blockchain configurations, suggesting validating validator sets could be a strategic response, as demonstrated by platforms like dYdX.

In summary, while Farrar sees merit in monolithic and modular frameworks depending on the specific use cases and utility, he remains particularly intrigued by the innovative potential of modular systems to revolutionize how blockchain architectures are constructed and operate in synergy. His balanced view underscores the belief that both paradigms will continue to coexist and evolve, each contributing uniquely to the blockchain ecosystem's diversity and resilience.

What are your thoughts on the intersection of crypto and AI? Do you think the concepts and sectors at this intersection are substantive or overvalued?

Farrar expresses cautious optimism about the intersection of crypto and artificial intelligence, noting that while many networks have been launched with strong narratives around this convergence, the practical utility remains largely unproven. However, he identifies two areas within this overlap that he finds particularly compelling:

Firstly, computational integrity is a significant area of interest. Farrar believes in a future where multiple agents perform tasks autonomously on behalf of users. The challenge lies in ensuring the authenticity and accuracy of these tasks. Blockchain technology could potentially play a crucial role in verifying that the computations and responses from these agents are executed correctly and yield accurate results. This application of blockchain could become increasingly important as we rely more on autonomous systems.

Secondly, Farrar highlights the ongoing development and commercialization of AI models. He points out the potential for blockchain to manage and distribute royalties to contributors who refine and enhance AI models. The current landscape sees many AI models being used and retrained, often in legally grey areas, to create more specialized and efficient versions. Blockchain could offer a structured way to compensate contributors fairly, ensuring they receive royalties for their work, especially as these original models contribute to training subsequent iterations.

However, Farrar remains skeptical about the viability of distributed computing networks in the long term. While there are short-term opportunities for those with high-performance computing resources to gain through token incentivization, he anticipates that larger hyperscalers will eventually dominate this space due to their vast resources and distribution capabilities.

In conclusion, while intrigued by the theoretical applications of blockchain in ensuring computational integrity and managing intellectual property rights within AI development, Farrar remains circumspect about the current execution and the long-term viability of some of the emerging business models at the intersection of crypto and AI. He calls for a clearer understanding of who the customers are and the specific applications being run on these decentralized networks, particularly those incentivized by token economies.

Rapid Fire Questions

1. What's one piece of content every aspiring investment professional should read/watch?

   • The best way to learn how to be an investor is to actually be in the market and make mistakes.

2. What’s your biggest investment mistake?

   • Having a great thesis, not sticking to it, not seeing it through, and being shaken out by market volatility.

3. What’s the most underrated use case of crypto?

   • Stablecoins.

4. What’s your most contrarian view in crypto right now?

   • I could see a world in which Solana is the market lead for the next few years.

5. What’s the biggest risk that the crypto space is facing?

   • Without regulation or diligent risk analysis, there's potential systemic risk with some of these financial primitives in crypto. Designing next-generation financial products without properly underwriting them, or even worse, explaining the risks transparently to users can be detrimental to user adoption

If you enjoy reading this issue, please consider subscribing. It takes 1 minute of your time but it would mean the world to us 🙇

Disclaimer: All the information presented in this publication and its affiliates is strictly for educational purposes only. It should not be construed or taken as financial, legal, investment, or any other form of advice.