Recap · August 4, 2025 · 40:35
Anil Lulla: How Crypto & AI Intertwine, Building Delphi, and More | TG Podcast
The Signal
Delphi Digital CEO Anil Lulla tells the story of building a research firm into a three-part crypto organization, then making AI a first-class area of work. The conversation ranges from early DeFi investments and Axie to the practical places where crypto could matter for AI, and to Delphi’s belief that the two fields will collide again.
Key Takeaways
- 01
A global crypto hive mind
Lulla describes Delphi as Research, Ventures, and Labs, organized around people who can compound one another’s work rather than operate as isolated crypto participants.
- 02
Consulting funded the start
Before the fund, Delphi earned through consulting for early protocols including Synthetix, Compound, Aave, Balancer, Lido, and Axie.
- 03
AI became an internal priority
The team’s curiosity after ChatGPT and the Situational Awareness paper led Delphi to build AI expertise rather than outsource its thinking.
- 04
Open systems are the overlap
Lulla sees crypto as potentially useful for open-source models, decentralized data and compute, and broader access, while acknowledging AI can progress without it.
- 05
Retail may get private access
He expects crypto-native platforms to be among the faster routes for retail access to private AI-company exposure before IPOs.
On the Record
“I think like playing it multiplayer is like the biggest hack.”
“The goal at Delphi is everyone's just working on what they want to work on.”
“I think AI doesn't need crypto, but we need AI to have crypto baked into it.”
“By the end of next year, 100%.”
The Breakdown
Three arms, one organization
Lulla introduces Delphi as Delphi Research, Delphi Ventures, and Delphi Labs. The organization began as a research shop in 2018, but he describes its real asset as a global hive mind: people who have already made money in or outside crypto and want to work inside a wider group where their information and impact can scale. The three units allow that shared research, investing, and building to reinforce one another.
Bootstrapping through early DeFi
Traditional investors were not eager to fund a crypto research firm, so Delphi supported itself with consulting. Lulla says Synthetix was among the first DAOs to pay the firm for risk-management work. That work led into early DeFi protocols including Compound, Aave, Balancer, Lido, and eventually Axie, where Delphi became especially visible. Consulting was therefore both a way to survive and a route into protocol relationships.
The fund’s first conviction tests
He recalls trying to call capital for an early Thorchain investment around six or seven cents, only to have the investor ignore the calls before a Twitter-driven drawdown. Delphi had done the work and viewed the fall to roughly three cents as either a generational entry or zero. The anecdote captures the uncertainty and small-team pressure of the early fund, when a missed call could determine whether conviction became a position.
Why Delphi went deeper on AI
Delphi’s model lets people follow curiosity, and ChatGPT’s arrival created a visible internal spark. A partner’s reading of the Situational Awareness paper sharpened the group’s sense that people were underestimating AI’s pace. Rather than delegate that analysis, Lulla says Delphi built an AI team and research platform over roughly nine months to find investments and products inside and outside crypto. The investment effort followed the research rather than substituting for it.
The crypto-AI intersection
Lulla does not claim AI requires crypto. He argues instead that crypto people should want AI systems with open-source and decentralized components, rather than a wholly centralized future. He names inference, training, and projects such as Grass as areas worth examining, while keeping the distinction between a technical possibility and an investable thesis. For him, the decentralization argument is a preference about the future’s architecture, not a shortcut to returns.
Access before an IPO
For the audience’s practical question, Lulla says the attractive AI companies often remain in secondaries or SPVs, which are difficult for ordinary buyers. His expectation is that crypto and retail platforms could provide unusually direct access before public listings. Delphi plans to publish more of its thinking through its intelligence site rather than treat the intersection as a one-line ticker call.
Past winners and current teams
Asked for Delphi’s best investment, Lulla names Axie, saying it returned more than 1,000x, and also points to Thorchain’s move from cents to above $20. Today he highlights investments such as Ethena and Jito, stressing that backing unusually strong teams matters. He ends optimistic that another AI-crypto season arrives, even if he cannot name the precise timing.
Distilled from the episode transcript · Counterparty Recap Desk
