Outlook for AI Agents
During Q4 of 2024, Crypto AI, particularly AI Agents, took Crypto Twitter (CT) by storm. For an introduction to AI agents, refer to our previous piece here. The year kicks off with Trump’s inauguration amidst ongoing policy reforms in the United States. Notably, a $500 billion private-sector investment in AI infrastructure has been announced. As a byproduct, this will undoubtedly draw attention to the Crypto AI vertical, further accelerating its development. It seems the AI Agent Winter is here to stay and OpenAi's release of Operator solidifies this.

The emergence of Agentic Ecosystems
In recent months, agentic infrastructure has seen rapid growth, with leading projects evolving to become the robust ecosystems they are today. Take Virtuals, for example: initially launched as an agentic framework and launchpad, has now evolved into an AI Agent Hub on Base, featuring significant adoption of its framework and a growing community. We believe it's fitting to draw a parallel between these agentic ecosystems and Layer 1s, as agentic ecosystems function as agentic Layer 1s. The similarities between the two are striking—let’s explore them in detail.

Execution engines
Layer 1s feature decentralised execution environments that allow developers to run smart contracts in a trustless manner. Ethereum has its Ethereum Virtual Machine (EVM), while Solana has its Solana Virtual Machine (SVM). Both execution environments provide the foundational layer for deploying and executing applications on their respective blockchain networks. Similarly, agentic frameworks like Virtuals' game framework and Eliza define the rules, resources, and logic that govern how agents interact, operate, and execute tasks within a broader ecosystem.

More importantly, these execution environments provide decentralised applications (dApps) with a standard execution layer that developers can build upon. Similarly, frameworks like Virtuals’ G.A.M.E and Eliza provide developers with a standardised codebase that serves as the foundation of every AI Agent in their respective ecosystems. This is key because it improves the developer experience, making the onboarding process and adoption of the agentic frameworks as easy as possible. If agentic frameworks were the backbone of every agentic ecosystem, then maximising the adoption of said frameworks is the most important mission for the team and the community.

Tokenomics
Layer 1's typically use their native governance token as the gas token, creating a direct link between network usage and token demand. Each operation is assigned a gas cost based on complexity, with users specifying a gas price in the native token. The total fee, calculated as gas used multiplied by gas price, is paid to validators, with some blockchains burning a portion to reduce token supply and enhance value. More importantly, this accrues to the token by controlling supply, hence the term “tokenomics”.

The first agentic ecosystem that embraced tokenomics from the get-go was the Virtuals ecosystem. The Virtuals’ team set up their AI Agent Launchpad in an ingenious fashion, requiring each AI Agent’s token to be paired with $VIRTUAL, Virtuals’ native token. This creates a flywheel effect, where users have to purchase $VIRTUAL in order to acquire any of the AI Agent’s tokens launched from Virtuals Protocol. Essentially, this tokenomic model accrues value to the $VIRTUAL token, akin to Layer 1 governance tokens. The success of Virtuals’ tokenomic model inspired many launchpads to adopt similar designs. Conversely, this speaks to Virtuals’ first-mover’s advantage by having well-designed tokenomics.

Community
The most important asset to Layer 1s is their community. A community of dedicated developers who are willing to dedicate themselves to build out a suite of dApps which then attract users to their chain. It is also important that this group of core developers establish a presence and continuously attract new talented developers to continue innovating. Other than developers, users are equally important because they are the main source of demand for dApps and other services hosted on the chain. However, this relationship is often a chicken and egg problem. Without developers, it would be hard to attract users, vice versa.

On the other hand, a strong community is just as important for agentic ecosystems. One of the most successful examples of a strong agentic community is ai16z’s led by Shaw. Although Shaw is a bit of a controversial individual when it comes to his actions, he has rallied together one of the most robust open-source developer communities in crypto. ai16z adopted a bazaar development model, where open-sourced developers came together and contributed to the Eliza framework in a dynamic manner. Shaw has taken efforts to put together educational content to onboard new developers in the form of detailed livestreams and videos. Not to mention that he also has taken physical trips to multiple foreign countries to engage ai16z open-sourced developers all over the world. The strength of ai16z's community is nothing short of excellent.

Back to basics
Regardless of how sophisticated an AI framework may be or how meticulously designed an agentic protocol’s tokenomics are, the core purpose of the crypto AI vertical remains to foster innovation and create new use cases at this unique crossroads of industries.

What better way to push this vertical forward than to tap on existing web2 AI talents? Yohei Nakajima, general partner at Untapped Capital, stumbled upon AI-generated SVG images which led to the creation of Pippin the unicorn, a family-friendly agentic influencer, as well as a $pippin token. At present, Yohei and his team has built out an agentic framework for Pippin, based on his past work, BabyAGI—the first popular open-source autonomous agent with task-planning capability, boasting 20k GitHub stars and over 70 academic citations. 

While not a direct correlation, CT often views GitHub stars as a measure of a framework's popularity and adoption. Yohei's BabyAGI was a key codebase that ignited the AI Agent trend among web2 developers, drawing significant attention and adoption. Pippin’s framework, which is built on top of BabyAGI, is likely to value-add to the crypto AI vertical. By attracting non-crypto-native talent and leveraging their expertise, the crypto community can break out of its echo chamber, embracing and integrating AI technologies widely adopted by the broader community. The goal remains to push for mainstream adoption of crypto.

*Disclosure: Signum Capital may hold positions in the companies mentioned above and the information provided on this newsletter is for general informational purposes only and does not constitute professional nor investment advice.

Thank you for reading!