At Adaverse, we’ve always looked for founders who don’t just follow trends, they challenge them. That’s why we invested in NODO, a DeFi protocol that has consistently evolved with the realities of the market. What started as a play-to-earn community pivoted into a prediction market, and now, into something far more ambitious: an AI-powered liquidity engine built for the future of decentralized finance.
In this conversation, we speak with Sowmya Raghavan, co-founder of NODO, about the hard-earned lessons behind each pivot, how autonomous AI agents are redefining market making, and why the next phase of DeFi requires more than just human decision-making: it demands intelligent, on-chain systems that never sleep.
Adaverse: NODO has pivoted several times – from a play-to-earn community to a prediction market, and now to an AI-powered DeFi protocol. What have been the biggest lessons learned from each pivot, and how have they shaped your current vision for NODO?
Sowmya: Each pivot taught us something critical about the real economic drivers in Web3:
1️⃣ Play-to-earn: We learned that speculative hype alone is unsustainable. Without deep utility or embedded value creation, token ecosystems decay. This shaped our belief that yield must be backed by genuine market activity, not just emissions.
2️⃣ Prediction markets: Here we discovered that while engagement is high, the coordination costs and volatility in user engagement made it difficult to build lasting liquidity.
3️⃣ AI-powered DeFi: Through these pivots, we also recognized something fundamental, crypto isn’t like traditional markets. It runs 24/7, relentlessly, while humans need to sleep. This highlighted a simple truth: the future of capital in DeFi can’t be human-managed alone. It has to be driven by intelligent systems that never stop.
That’s why we shifted to building an AI-powered protocol, transforming passive, idle liquidity into actively managed, performance-driven capital. AI is becoming an essential layer of how markets- and soon much of life – will operate. With NODO, we’re placing that future squarely on-chain, where it’s transparent and accessible to all.
Adaverse: Your latest product leverages autonomous AI agents to optimize market making and liquidity provision on decentralized exchanges. Can you walk us through how these AI agents work in real time and what makes them different from traditional quant or high-frequency trading strategies?
Sowmya: Traditional quants and HFT systems often rely on centralized infrastructure, opaque algorithms, and custodial capital. They also typically operate in centralized markets with clear order books and counterparty controls. In DeFi, liquidity provision looks very different: you’re effectively underwriting volatility in an AMM pool, earning fees for taking that risk.
Our AI agents are on-chain execution systems that continuously:
👉🏼 Monitor real-time market conditions like volatility, slippage, and liquidity depth across decentralized pools (e.g., Cetus on Sui).
👉🏼 Decide dynamically when to rebalance liquidity ranges based on predefined strategy parameters (like ΔP_range, δ_trigger) ensuring capital stays “in-range” to capture fees, but avoids excessive churn.
👉🏼 Calculate gas vs profit thresholds, only executing adjustments when the expected additional yield outweighs costs.
👉🏼 Automate position resizing or rotation to new pools based on evolving market data.
Unlike typical bots that just follow signals, these agents integrate strategy intelligence with autonomous execution – they’re managing capital with a continuous risk-return logic, transparently logged on-chain. This is also why they’re more resilient to sudden volatility spikes or manual errors. It’s essentially professional-grade liquidity management, delivered through autonomous, decentralized infrastructure.
Adaverse: Fragmented liquidity and impermanent loss are persistent challenges in DeFi. How does NODO’s AI-driven approach specifically address these issues, and what results have you seen so far in terms of yield and efficiency?
Sowmya: Fragmented liquidity means that on most chains, capital is spread thin across hundreds of pools, many poorly utilized. Impermanent loss (IL) then compounds the problem, eating into returns when markets trend.
NODO’s AI vaults tackle this by:
- Concentrating liquidity intelligently: Our agents keep positions extremely tight around the market price (for example, ±0.10% bands) and rebalance frequently, ensuring high fee capture with minimal idle exposure. In backtests, we’ve achieved >98% in-range time.
- Reducing IL through dynamic range management: Rather than “set-and-forget,” agents adjust bands based on live volatility and pool health, minimizing prolonged out-of-range periods that cause IL.
- Routing capital to pools with the best fee-to-risk ratios: Using real-time data across DEXes, agents prioritize pools that offer the best organic yields, not just artificial rewards.
To evaluate the performance of NODO’s AI-powered strategy, the team conducted a 24-day backtest (April 22 – May 15, 2025) using the HFRR-CLMM strategy on Cetus AMM (Sui). The goal was to outperform static LPs while minimizing impermanent loss and optimizing agent behavior.
Key Metrics
| Metric | CETUS/SUI | DEEP/SUI | USDC/SUI |
| Initial Capital | $1,000 | $1,000 | $1,000 |
| Final Value | $1,404.65 | $1,651.80 | $1,154.73 |
| Total Return | 40.46% | 65.18% | 15.47% |
| Average APR | 1,261.57% | 2,106.89% | 333.05% |
| Max APR (peak moments) | 2,664.93% | 8,079.77% | 820.32% |
| Impermanent Loss | 0.03% | 0.00% | 0.05% |
| Gas Cost (Total) | $180.72 | $189.84 | $189.39 |
| In-Range Time | 99.82% | 98.53% | 99.41% |
Key Insights
➡️ Strategy Optimization Improved Significantly: From 17.65% in-range in early tests to >98% in-range for all pools during this extended window.
➡️ Capital Appreciation > Fee Yield: Most returns were generated from price movement, not trading fees.
➡️ Gas Cost Sensitivity: Transaction fees consumed 29–122% of profits. Gas optimization is a critical tuning parameter.
➡️ Fee Returns Were Minimal: Ranged from just 0.01%–0.06% of initial capital.
➡️ High Frequency Execution: ~3,500 transactions per pool (~150/day), showcasing agentic activity.
Adaverse: NODO has grown a large user base (400,000 users) and partnered with top DeFi projects on the Sui blockchain. What are the main things you’ve done to attract users and build an active community?
Sowmya: We took a very grassroots approach. From the early days, we treated our users not just as customers, but as co-builders. We ran prediction markets that actively involved communities in outcome staking, launched campaigns with referral incentives, and most importantly, maintained a very direct founder-to-user communication channel. Many DeFi users from our recent interviews were invited to test our vaults early, give us honest feedback, and even participate in a private group to stay connected to the core team.
Partnerships have also been key. By forging over 100 partnerships across the ecosystem and joining the Sui Foundation’s incubation cohort, we’ve positioned NODO at the very heart of where liquidity lives. We also integrate with flagship Sui DEXes like Cetus. This lets us drive user adoption natively through existing ecosystems, while offering something uniquely attractive: autonomous yield optimization that most protocols don’t have.
Adaverse: As you expand to new blockchains and integrate with more DeFi protocols, what are the biggest technical or regulatory hurdles you face, and how are you preparing to overcome them?
Sowmya: On the technical side, the biggest challenge is ensuring our agent systems can operate across very different DeFi environments. Each chain has unique liquidity dynamics, gas models, and data latency risks. We’re building modular execution frameworks so our AI agents can plug into Sui today, but easily extend to EVM chains like Arbitrum or Base tomorrow , using chain-specific gas estimators, slippage controllers, and backtested parameter sets.
On the regulatory front, the rise of autonomous agents handling capital introduces new compliance considerations, especially around custody and investor protection. While we’re fully non-custodial, we’re also working on rigorous audit trails and making sure our vaults are transparently structured with clear disclosures on risks like IL and market volatility. Longer-term, we see DAO governance playing a crucial role in aligning incentives and decentralizing oversight.
Adaverse: Looking ahead, what is your long-term vision for NODO in the DeFi landscape?
Sowmya: Our long-term vision for NODO is to build the foundational infrastructure layer for autonomous liquidity and market making across all of DeFi, effectively becoming the “AI engine” that powers capital deployment, risk management, and trading intelligence in a decentralized world.
In the near term, this means scaling beyond our initial AI vaults and Sui deployments to achieve:
- Cross-chain dominance: We’re moving quickly from Sui to multi-chain support, integrating with non-EVM chains like Cardano and deploying cross-chain trading capabilities by Q4 2025. This expands NODO’s reach to more liquidity venues and user segments.
- AI strategy marketplaces & social trading: By late 2025, we’ll launch an AI agent marketplace and a decentralized reputation system where traders, quants, and developers can publish strategies that are executed on-chain by NODO’s infrastructure, bridging retail and institutional capital directly to algorithmic expertise. Verified agent creators and social trading features will help democratize access to sophisticated liquidity strategies.
- Institutional-grade adoption: As we roll out advanced portfolio management and AI-powered risk tools, we’re targeting hedge funds and proprietary trading firms managing upwards of $250M AUM. The goal is to have NODO not just serve retail LPs, but also become a trusted infrastructure provider for professional capital.
But our ultimate horizon goes even further: we envision a permissionless ecosystem where anyone can upload, backtest, and monetize new liquidity strategies. A truly decentralized network of AI agents competing and collaborating to allocate capital, all governed transparently by token holders. In this model, DeFi liquidity becomes adaptive, composable, and hyper-efficient with humans focusing on creative strategy design while autonomous agents handle execution at scale.
Adaverse: As AI agents become increasingly autonomous and start to dominate trading activity on decentralized exchanges, how do you envision the relationship between human traders and AI evolving? Do you think the future of trading will be a contest between algorithms, or is there still a unique role for human intuition and creativity in these markets?
Sowmya: AI is undoubtedly going to handle the bulk of execution, especially in 24/7 crypto markets that never rest. It’s simply more efficient for agents to monitor hundreds of pools, volatility signals, and gas metrics simultaneously than any human could.
But that doesn’t eliminate the human edge. Creativity, intuition, and macro thematic insight still play a massive role. The best opportunities often come from new ideas, spotting market structures, experimenting with alternative incentive designs, or even understanding cultural trends that quantitative models overlook.
We see a symbiosis. Humans will increasingly design the frameworks, define the high-level strategy hypotheses, and let AI agents handle the tactical grind of execution, risk balancing, and optimization. Over time, protocols like NODO are positioning themselves to be the infrastructure that connects these, helping people innovate at the strategy layer, while agents execute on-chain.
Adaverse: Thanks Sowmya, this has been incredibly insightful!
