The Next Bull Run's Battleground Is Hiding in Two Unexpected Asset Classes

In the summer of 2017, while building ChainLit to decode ICO whitepapers for university clubs across Bonn, I noticed a recurring pattern. The projects that survived the subsequent crash weren't the ones with the loudest Telegram groups or the most aggressive marketing. They were the ones that solved a real human need—whether it was remittances, identity, or access to savings. Today, as we sit at the edge of another cycle, with BTC halving behind us and ETH ETF inflows steadying, the same pattern is emerging. But the battleground has shifted. The next bull run will not be won on the floors of DEXs or in the liquidity pools of copycat L2s. It will be won in two asset classes that most retail traders still ignore: tokenized Real-World Assets (RWAs) and AI-governed autonomous agents on-chain. These two categories are not just narratives—they are the institutional and philosophical mirrors of what blockchain was always meant to be: a bridge between code and human trust. I've learned this the hard way, through a decade of community building, bear markets, and boardrooms. Today I want to show you why these two classes are the real infrastructure for the next wave, and why buying the hype around generic infrastructure tokens might leave you holding empty bags. Let's start with RWAs. When I partnered with Deutsche Bank's digital assets desk in 2024, I designed a crypto literacy program for senior executives. Their first question was always: "How does this add transparency without violating our privacy obligations?" This is the core insight that most crypto natives miss. RWAs—bond tokens, property deeds, invoice financing pools—solve the institutional trust problem not by making everything public, but by making the audit trail verifiable while keeping individual positions private. Using zero-knowledge proofs, a fund manager can prove that the tokenized treasury bills backing a stablecoin are held in a regulated custodian without exposing the custodian's full ledger. This is not theory. As of Q1 2025, over $18 billion in tokenized US Treasuries are live on public chains like Ethereum and Polygon, with BlackRock's BUIDL alone commanding $5 billion. The growth rate is 40% quarter-over-quarter, and it's accelerating because the underlying user need is real: Asian pension funds want dollar exposure without the hassle of SWIFT; African fintechs want T-bill yields for their mobile money float. The technical layer is now ready—we have compliant oracle networks, instant settlement, and regulatory clarity in Singapore, Switzerland, and the UAE. Based on my audit experience with over a dozen RWA protocols, the key risk is not smart contract bugs but liquidity fragmentation. Most RWA tokens trade on niche DEXs with thin order books. Until they hit the order books of Binance or Coinbase, the volume will remain institutional-only. That transition is the alpha. Now the second class: AI agents on-chain. This is where my work on the "Human-Centric AI" initiative comes in. In 2025, I led a global summit in Frankfurt with 1,000 participants debating how to embed ethical constraints into smart contracts that autonomous agents would execute. The thesis is simple: the next million users on-chain won't be humans hitting "swap" buttons—they will be AI bots executing complex workflows. Imagine an agent that monitors Dutch electricity prices, buys energy tokens when below €0.05/kWh, and sells them to a factory's smart contract when demand spikes. Or a DAO-governed agent that automatically rebalances a treasury between ETH and staking yields. These agents need deterministic execution environments, which only blockchains provide. The technical challenge has been oracle reliability and proof-of-machine-cost—how do you verify that an agent actually consumed computational resources? Recent advances in zk-SNARKs for agent execution proofs are solving this. Projects like Autonolas and Fetch.ai have seen their on-chain agent interaction counts double every month in 2025. The contrarian angle here? Most of these agents today are glorified Telegram bots stealing MEV. But the infrastructure is getting real: EigenLayer restaking now secures agent heterogeneity, and Coinbase's Base chain is already seeing 15% of its transactions originate from non-human wallets. The cultural translation work I did with bankers taught me that institutions will embrace AI agents far faster than they embraced DeFi—because agents reduce human error and fit into existing compliance workflows. But let's step back and apply a stress test. The buzzing narrative right now is that the next bull run will be driven by Layer 2 scaling, or by a new monolithic L1, or by the next DePIN hardware token. I am deeply skeptical. My analysis of the data availability (DA) layer hype reveals a critical flaw: 99% of rollups today generate less than 50 transactions per second. They don't need dedicated DA from Celestia or Avail—they can post to Ethereum mainnet at negligible cost post-Dencun. The DA narrative is a marketing ceiling, not a technical floor. The real bandwidth constraint is not data—it's user onboarding and regulatory clarity. RWAs and AI agents directly tackle those constraints. They bring off-chain trust and off-chain intelligence into the chain's scope. Here's where the contrarian angle becomes uncomfortable for most traders. The first wave of RWA tokens (like ONDO, Pendle's yield tokens, and tokenized treasuries) have already rallied 5x from their lows. Some call it a bubble. But look at the fundamentals: the Federal Reserve is pivoting to rate cuts, which will drive demand for yield-bearing tokens. The price of ONDO is still trading at a discount to the net asset value of the treasuries it represents. That's not a bubble—that's mispricing due to illiquidity. Similarly, AI agent tokens like OLAS and FET have retraced 60% from their peaks because the market realized that agent utility was overhyped. But during that retrace, the number of active agents on their networks quadrupled. This is classic "useful but undervalued" territory. My experience during the 2017 ICO crash taught me to ignore prices and watch usage. The usage signals in these two classes are screaming. Yet the risks are real. For RWAs, the regulatory sword hangs over every token. In the US, the SEC has not yet issued guidance on tokenized securities; any project that labels its token as a "utility token" for a security asset is playing with fire. Based on my work with Deutsche Bank, the only safe structure is to issue tokens under Regulation S or as digital notes from a licensed bank. Most small projects don't have that legal budget. For AI agents, the risk is ethical and operational. An agent with a poorly written smart contract can lock funds forever. During my summit, we simulated a scenario where an agent's price-feed oracle was compromised—the agent bought high and sold low, losing 30% in five minutes. Until we have standardized agent audits and fail-safes, retail users should not hand over custody. Community is the only chain that cannot be broken—and that applies to the communities that build these assets. So what's the takeaway? The next bull run's battleground is not a chain war. It's a war of trust translation. The chains and tokens that survive will be those that translate real-world value (yield, property, legal claims) and real-world intelligence (automated decision-making) into verifiable, composable on-chain assets. The maximalists will keep fighting over TPS and finality. The pragmatists—and the institutional bridges we build—will win the next cycle. I've seen this pattern before, and I'll bet on it again.