In an exclusive interview with The Coin Headlines, Varun Datta, Founder and CEO of early-stage digital finance infrastructure-focused venture firm, Truth Ventures, shared his thoughts on a wide array of topics, ranging from Web3 infrastructure, its convergence with AI, DeFi adoption, and stablecoin use-cases.
Datta’s investment philosophy is rooted in conviction-led investing. He has also provided strategic guidance, technical expertise, and ecosystem access to help companies scale from concept to adoption. The interview follows below.
A lot of Web3 infrastructure was built between 2021 and 2024 that nobody is using at scale yet – how do you distinguish genuine infrastructure from infrastructure that was built ahead of demand that may never arrive?
Being early is not the same as being necessary. Genuine infrastructure solves a structural problem: settlement, liquidity, interoperability, compliance, custody, or access. Speculative infrastructure often starts with the technology and then searches for a market to justify it.
Real infrastructure usually has evidence of demand before it achieves scale. Many institutions still use manual processes, scattered systems, or costly intermediaries. The issue is clear, and there’s already a willingness to pay. Infrastructure built for imagined demand works differently. Its usage relies on incentives, and the product needs ongoing support and stories. Plus, there’s little cost when it fades away.
I also look at whether the infrastructure becomes more valuable as more participants join and whether its utility grows alongside the market. Strong infrastructure compounds. Each new application, asset, or institution adds value for users, products, and ecosystems. Weak infrastructure relies too much on one ecosystem or story. It gets stuck in that space and confuses isolated actions with real progress.
The key test is if adoption is slowed by outside issues like regulations, distribution, liquidity, or interoperability, or if there’s simply no real need. Durable infrastructure fixes ongoing issues. It adds clear economic or operational value and supports various products, institutions, and market cycles. If activity disappears when token incentives end, or its success depends entirely on one application, ecosystem or narrative, the demand is likely subsidized rather than sustainable.
Everyone is talking about AI and Web3 converging – where is the real opportunity and where is it mostly narrative?
The real opportunity lies where AI needs capabilities it cannot provide on its own: identity, ownership, payments and verifiable execution. AI can make decisions, but Web3 infrastructure adds economic power. It lets autonomous agents own assets, make payments, access services, and prove their actions without needing a single platform or middleman.
The strongest use cases are therefore not “AI on a blockchain” in the abstract. They are applications where decentralised infrastructure addresses specific issues. These include machine-to-machine payments, proof of data origin, controlled access to digital assets, and good coordination among autonomous agents. In these cases, Web3 is not being added for branding; it provides the trust, transaction and accountability layer that AI lacks.
Programmable payments, verifiable credentials, digital ownership, and clear execution can give agents a portable economic identity. This lets them work across platforms without needing a separate deal with every provider. The real opportunity is therefore less about decentralising AI models and more about building the infrastructure that allows autonomous systems to participate credibly and independently in an economy.
The convergence turns mainly narrative when projects just add a token to an AI product, put model outputs onchain, or call a centralised app an “agent economy.” In many cases, blockchain adds cost and complexity without removing a meaningful constraint.
The test is simple: does Web3 allow the AI system to do something it could not otherwise do credibly, securely or independently? If not, the convergence is largely narrative. If it enables genuine autonomy
AI agents need to transact, hold assets, and execute autonomously – what infrastructure gaps need to close before that is actually possible at scale?
The biggest gap is not giving agents the ability to transact. It is giving them the authority to do so safely. An agent can execute instructions and interact with a wallet. However, to operate at scale, it needs a reliable framework for identity, permissions, and accountability. Every agent needs to be able to prove who authorised it, what it is allowed to do, how much it can spend and under what conditions its actions can be stopped or reversed.
The second gap is financial infrastructure designed for autonomous actors rather than humans. Agents need programmable wallets and real-time settlement. They also require access to various asset types. Also, they should be able to transact across platforms without needing separate integrations for each party. That also requires better interoperability between chains, payment systems and data ecosystems. An agent economy cannot scale if every transaction is confined to a different ledger or dependent on a closed ecosystem.
The third gap is trust. Agents need clear data, open execution records, and checkable decision trails. This helps institutions see not just what an agent did, but also why it did it. Security and recovery are equally important. If an agent is compromised, exceeds its mandate or makes a faulty decision, there must be clear controls for limiting exposure, revoking access and assigning responsibility.
The infrastructure challenge is therefore not autonomy alone, it is bounded autonomy. AI agents can work effectively at scale when they can act on their own. This requires rules that are programmable, verifiable, and enforceable. Until identity, settlement, interoperability and accountability mature together, most agent activity will remain experimental rather than economically trusted.
DeFi TVL has not meaningfully recovered to its 2021 peaks despite significantly better infrastructure – is this a demand problem, a trust problem, or a product problem?
It is primarily a product problem, reinforced by trust, but the premise also deserves scrutiny. Did the 2021 TVL peak represent genuine adoption in the first place? A large part of that capital came from very high yields, token incentives, leverage, and speculation. Much of the liquidity was not loyal to a protocol, or even to DeFi itself; it moved wherever returns were highest. When those incentives disappeared, and the risks became more visible, the capital left. Today’s lower TVL may therefore say less about DeFi’s failure to recover and more about how much the 2021 peak overstated durable demand.
The infrastructure is unquestionably better now: transactions are faster, cheaper and more reliable. But much of DeFi still offers products that are difficult to understand, operationally complex and designed for users already comfortable with crypto. Infrastructure can reduce friction, but it cannot manufacture product-market fit.
Trust is also a constraint, with users exposed to smart-contract, oracle, governance and liquidity risks. For institutions, legal certainty, custody, compliance, and accountability are just as important as technical performance.
The real question isn’t why TVL hasn’t returned to 2021 levels. It’s whether DeFi can draw in and keep capital without paying users to join. Extended growth will happen when it provides better access, pricing, transparency, or efficiency than traditional options.
Which DeFi primitive – lending, DEXs, derivatives, yield – do you think is closest to genuinely replacing its traditional finance equivalent, and which is furthest away?
DEXs are closest, although they are nearer to replacing the structure of traditional markets than their scale. Blockchain-native assets combine trading, liquidity, and settlement in one clear system. This cuts down the need for brokers, exchanges, clearing houses, and custodians. That is more than digitising an existing process; it is redesigning the market structure itself.
The remaining barriers are institutional. They include fragmented liquidity, inconsistent execution, regulation, and market protections. DEXs have therefore proven the model, but they have not yet built the conditions required to absorb mainstream capital markets.
Derivatives sit next. Onchain perpetuals are in high demand. They provide continuous trading, clear collateral, and quick settlement. But replacing traditional derivatives markets will require deeper liquidity, more sophisticated margining and risk management, reliable pricing and legal certainty around netting and default. The trading product is strong, but the market infrastructure isn’t ready to replace the traditional option widely.
Lending is furthest away because most DeFi lending is still overcollateralised. Borrowers often have to put up assets worth more than the loan. This makes it more like automated secured financing rather than traditional credit. Replacing traditional lending requires identity, credit assessment, private but verifiable financial data, legal enforceability and recourse when borrowers default. DeFi will support traditional lending until it can offer safe, undercollateralised credit.
I would treat yield differently. Yield is not a standalone primitive; it is the return generated by lending, trading, staking or taking risk. The real question is whether that yield comes from sustainable economic activity rather than leverage, token emissions or temporary incentives.
Stablecoins are DeFi’s clearest consumer success story – what is the second one, and how far away is it?
Stablecoins are DeFi’s clearest consumer success story. The second is DeFi lending. Stablecoins solved digital dollars. DeFi lending solves digital credit.
Protocols like Aave have already proven that transparent, permissionless lending markets can operate at scale, with billions of dollars borrowed and lent on-chain. Today, the primary users are crypto-native individuals and institutions, but the underlying infrastructure is mature.
The next step is abstraction. As wallets improve, tokenised real-world assets grow, and regulation becomes clearer, users will increasingly access DeFi lending through fintech apps rather than directly interacting with protocols.
I think we’re around three to five years away from DeFi lending becoming as invisible – and as impactful – as stablecoins are today
Non-dollar stablecoins keep launching and failing to gain traction – what needs to be true for a yen, euro, or AUD stablecoin to genuinely compete with USDC at scale?
A non-dollar stablecoin does not need to beat USDC everywhere. It needs to become indispensable somewhere. Stablecoins follow the currency hierarchy below them. The dollar leads in global trade, reserves, collateral, and crypto liquidity. A euro, yen, or AUD token isn’t just competing with USDC; it’s also up against the dollar, which is the main currency for settling transactions in the onchain economy.
The most credible path is to own a specific economic corridor or market where the underlying currency already matters. A euro stablecoin could grow in European capital markets and treasury flows. A yen stablecoin might expand through Asian trade settlements. An AUD stablecoin could thrive in regional commerce, payroll, or commodity-linked activities. The demand has to begin with real economic obligations denominated in that currency, not with the hope that users will switch simply because another token exists.
It also needs a complete liquidity network. This includes reliable banking access, solid on- and off-ramps, tight spreads, market makers, exchange pairs, collateral utility, and links with wallets,
payment providers, and DeFi protocols. Issuance is the easy part. The tricky part is making sure users can earn, borrow, trade, settle, and account in that currency without constantly converting back to dollars.
That is the real threshold for competition. A non-dollar stablecoin matters when it shifts from being a token for a local currency to the main financial system for an economy or corridor. Until then, USDC retains the advantage because it isn’t just a stablecoin; it is part of the dollar’s wider network effect.
