Special thanks to Sarthi and Sarang for the review and feedback. ❤️
It’s August. Half the year is gone, which makes it a good time to look back at what the smart money called in January: a16z’s partners, Galaxy’s researchers, and Y Combinator’s Request for Startups all pointed at the same short list: stablecoins, RWA tokenization, payments, agents and AI, and privacy.
Some of those markets are already taking shape. Others still need the infrastructure, product, or distribution to catch up, and that’s where the opening is.
The question now is simple: what do you build?
Build stablecoin on- and off-ramps
Stablecoins (mainly USD-pegged) have made the cross-border transfer leg far easier: dollars move globally in seconds to minutes, 24/7, onchain at low network cost. In most real flows, this is only the middle step: users still need easy access to stablecoins on the way in, and local ways to spend, withdraw, or invest on the way out.
The weak thesis is that stablecoins will simply replace existing cross-border providers because they are faster and cheaper. In mature corridors that edge is often overstated. Firms like Airwallex already route transfers over local rails such as ACH, SEPA, and UPI for fast, low-cost settlement. When the end goal is local fiat, FX, compliance, and on/off-ramp costs remain.
The bigger opportunity is what happens around the transfer: products that let users receive, hold, and invest digital dollars, then spend or withdraw locally through bank accounts, cards, QR payments, real-time payment networks, or cash-out agencies.
The hard part is getting local execution right, such as licenses, banking partners, FX liquidity, compliance, and payout networks. These are slow, market-specific, and hard to copy.
Stablecoins made global dollar movement much better. The next opportunity is making that money locally useful.
Build the agentic economy
The agentic economy needs two things underneath it: a way to know an agent (who it is and how good it is), and a way for it to pay. We’re still early. The bottleneck has moved from intelligence to identity, and right behind it, to settlement.
Circle co-founder Sean Neville calls the first part “Know Your Agent”: signed credentials that link an agent to its principal, its constraints, and its liability. It’s KYC for agents. But identity is only part of it. The other question is how good an agent is: its track record, and whether its work can be verified. Early implementations are already here: onchain agent registries, wallet-native agents transacting in USDC, ERC standards for trust-minimized agents, and toolkits that pair identity with built-in payments and fraud controls. Until a common identity standard emerges, merchants block agents at the firewall. No incumbent owns this yet. ERC-8004 is a starting point: onchain registries for an agent’s identity, reputation, and verified work, for building this missing layer of trust.
The second is payment, and it’s further along. x402 and MPP use the HTTP 402 “Payment Required” code: an agent calls a payable endpoint, pays in stablecoins, and gets the requested service back. No account, no human, no checkout. Any server becomes a merchant. These payment protocols settle at sub-cent cost with no protocol fee, and any developer can price an endpoint. The protocol is the easy part; it’s open and free. But the businesses on top are thin: some exist, few are good. Agent wallets and spending limits, service listing and discovery (e.g., mpp.dev/services), reliable agents built to these standards.
The agentic economy runs on identity and payment. Identity has no standard; payment has protocols but no winner yet. Both are wide open. Go build the ones that work, and you’re laying the floor the whole agent economy will stand on.
Build privacy into the product
Many businesses will not move meaningful activity onchain without strong privacy controls. They need to protect commercial relationships, transaction amounts, positions, payroll data, supplier terms, and customer information.
Privacy is one of the core product requirements for finance.
Ali Yahya highlights why it can also become a durable advantage: assets can move across networks, while private information and transaction metadata are far harder to move safely. Even when a transaction hides its direct contents, surrounding information can reveal patterns about timing, counterparties, and behavior.
Builders can create valuable products here: private settlement systems, confidential business payments, selective disclosure tools, privacy-preserving identity, compliance systems that reveal required information only to authorized parties (see Oracle’s confidential computing), privacy-preserving AI (see venice.ai), and secure workflows for tokenized assets.
The best privacy products will give users control without forcing them to become cryptographers. They will feel straightforward, trustworthy, and familiar.
Build for customers
Crypto has spent a decade building capabilities: global settlement, programmable money, digital ownership, transparent markets, software-native finance. The next wave comes from applying them to a customer’s real problem.
So start with a painful workflow, and the people who live with it. Find where money gets stuck, where compliance stalls, where cross-border breaks, where software forces manual workarounds. Then use crypto only where it’s a real advantage, not because it’s crypto.
The winners will make cross-border payments feel local, institutional operations feel fast, agentic commerce feel safe, private finance feel usable. Their users won’t know or care which chain, token, or consensus is underneath. They’ll care that it’s faster, cheaper, safer. Safe is the one a builder can’t trade off, and it settles best on the most secure, most neutral, most battle-tested layer there is. That’s Ethereum.
Crypto built the rails. The industry gets built on top.
Sources and further reading
15 Reports: The 2026 Crypto Outlook Roundup - Enterprise on Chain
Stablecoins in Asian Cross-Border Payments - IOSG Ventures
Big Ideas 2026: Part 3 - a16z
The Missing Infrastructure for AI Agents: 5 Ways Blockchains Can Help - a16z crypto
Demystifying Evals for AI Agents - Anthropic
Thanks for reading. It’s time to build! 🥤
Riely and the Geode Labs
Have thoughts on this issue? Reply or DM me - I read everything.
Want to support Local Ethereum? Reach out about sponsorships.
Riely
Riely is the Editor in Chief of Local Ethereum, covering Ethereum and crypto adoption stories from around the world. Based in Berlin, she covers stories from India, Argentina, Poland, Taiwan, Serbia, and beyond, with a focus on how decentralized technology intersects with local culture, economics, and politics.
Published August 26, 2026 · 5 min read
Local Ethereum
Enjoying the read? Get this in your inbox.
In-depth Ethereum ecosystem coverage: regional deep dives, monthly roundups, and global analysis. Twice a month, free.
Subscribe, it's free
