Menu #16: The Machines Are Onchain

Near Protcol
Never Split the Difference
Web3 Dinner Club: 25th June
The Agent Economy
Brand spice
Chef’s Note

This is The Menu: the UK Web3 operator’s weekly briefing.

What founders, investors, and builders are actually discussing behind closed doors.

The industry is starting to look less like a series of disconnected experiments and more like a system finding its shape.

In this issue:

Signal, served weekly.

The dinner club is proudly sponsored by Novel Labs.

A multi-award-winning London storytelling studio building the brands of the future in AI, blockchain, and emerging technologies.

Best known for the $100m expansion to the Bored Ape Yacht Club, The Mutant Cartel World.

If you’re a startup or scale-up building a brand and looking for real go-to-market impact from those who have repeatedly built unicorns and category kings as VCs and founders... ask for an intro at the table.

NEAR is a Layer-1 blockchain. Same category as Ethereum or Solana, but ask the team what it is today, and you'll get a different answer.

Founded by two engineers with backgrounds in AI research, NEAR was always built to scale. Its sharding technology, Nightshade, splits the network into parallel processing lanes so transactions stay fast and cheap regardless of demand. No congestion. No eye-watering fees.

What sets it apart in 2026 is its direction of travel.

Most Layer-1s are financial rail;, they exist to move value. NEAR has repositioned itself as the "blockchain for AI": infrastructure where AI agents can hold identity, verify data, and transact autonomously across chains, without a human approving each step.

It's a subtle but significant shift.
From a platform for developers to a backend for intelligent systems.

Most blockchains are built for transactions.

NEAR is building for something else entirely: a world where AI agents act, pay, and coordinate on your behalf across every chain without you needing to touch a bridge, a wallet, or a swap interface.

That pivot started quietly in 2023. By 2026, it has become the defining bet of the protocol.

The identity shift

NEAR no longer leads with "fast, cheap Layer-1".

Its own homepage now reads:
"AI serves as the front end to interface with users and carry out intent, while our blockchain acts as the backend to handle identity, trust, and data."

This isn't marketing. It's architecture. NEAR has spent the last two years building the infrastructure stack to back that claim up, and the pieces are now visible.

NEAR Intents

The most important building block is NEAR Intents, a multichain coordination layer that changes how users and agents interact with blockchain entirely.

Instead of manually executing each transaction step, approve, bridge, swap, confirm, users simply express what they want to achieve. The protocol handles the routing, sequencing, and settlement automatically.

For humans, that's better UX. For AI agents, it's something more significant: the ability to operate across chains autonomously, coordinating liquidity and value without human oversight at each step.

Universal Send

The most tangible product of that vision landed in late May 2026:
Universal Send.

Scan a QR code. Choose your token. Confirm once.

The system automatically handles the cross-chain conversion and routing; the recipient gets the asset they want, regardless of what the sender started with. No manual bridging. No failed transactions from wrong network selection.

Crucially, it's confidential by default. The sender, receiver, amount, and routing path are not exposed on public explorers. In a world where on-chain payment history is permanently visible, that's a meaningful design choice.

The AI token rally

Markets have started noticing. NEAR led the AI token rally in May 2026, rising 34% in 24 hours as the network upgrades landed, outperforming the broader AI coin category at a moment when NVIDIA's earnings were driving renewed confidence in AI infrastructure plays.

A strategic partnership with NVIDIA for sovereign AI infrastructure and TVL surpassing $1.8 billion have reinforced the narrative.

Why this matters for operators

The deeper story isn't about price.

It's that NEAR is building the plumbing for an economy where AI agents need to pay each other, across chains, at scale, privately, and without human sign-off on every step.

That's not a crypto story. It's an infrastructure story. And the teams building AI-native products in 2026 will care about which blockchain can actually support autonomous, multichain agents, not just which one has the lowest fees.

W3DC:
Most Layer-1s are still competing on throughput and fees. NEAR is competing on a different question entirely: what does a blockchain need to look like when the primary user isn't human?

Chris Voss’s Never Split the Difference sounds like it should be a hard-edged book about winning negotiations.

And in one sense, it is.

Voss was a lead FBI hostage negotiator, so the stories are fast-paced, tense and often more gripping than a business book has any right to be.

But the lesson is surprisingly simple:

Most people are not beaten in negotiation because they lack clever tactics.

They are beaten because they do not listen properly.

The book argues against the old idea that negotiation is just compromise. Meet in the middle. Split the difference. Everyone gives something up.

Voss takes a different view.

Good negotiation starts by making the other person feel heard. That means listening carefully, naming what they might be feeling, asking better questions, and resisting the urge to rush toward your own answer.

The tools are practical.

Mirroring encourages people to say more.

Labelling helps surface fear, hesitation or resistance.

Calibrated questions such as “How would that work?” or “What would need to happen?” move the conversation forward without forcing a fight.

And the idea of getting someone comfortable enough to say “no” is useful, because it gives them control and often opens up a more honest conversation.

For founders and operators, this matters.

You negotiate all the time.

With investors.
With clients.
With suppliers.
With your team.
With yourself.

The best part of the book is that the lessons are not just useful for boardrooms or hostage situations. They apply to parenting, relationships, sales calls, hiring, difficult feedback and everyday work.

At its core, this is not really a book about tricks.

It is a book about understanding people.

W3DC: The strongest negotiators are not always the loudest, smartest or most forceful. They are often the ones who listen long enough to understand what is really being said.

A curated, seated dinner for a small group of builders working in crypto, AI, and frontier tech.

One table. No pitches. No panels. No ego contests.

Just the kind of conversation that doesn't show up in your LinkedIn feed. The relationships that move capital, talent, and ideas in Web3 don't start at conferences.

They start at a handful of dinners with the same people, repeated over time.
Seats are limited by design.

Proudly sponsored by Novel Labs.

Most conversations about AI and crypto focus on price action or speculation.

On June 11, Coinbase launched Coinbase for Agents, a product that lets AI assistants like ChatGPT and Claude connect directly to a Coinbase account and execute trades, manage portfolios, and pay for digital services autonomously, within limits the user sets.

At launch, agents can trade spot crypto and derivatives. Equities, prediction markets and commodities are coming next.

The payment layer underneath it is called x402: a machine-to-machine protocol that lets agents pay for research APIs, real-time data feeds, and on-demand compute without a human logging in, checking out, or approving the transaction.

That last part is what matters.

This isn't just automation. It's a structural shift in who or what is transacting.

AI agents can't open a bank account. They can't pass KYC. They can't hold a debit card. But they can hold a crypto wallet. And that is precisely why crypto becomes the default financial rail for the agent economy, not just another asset class.

The numbers already tell that story. Between May 2025 and April 2026, AI agents completed over 176 million on-chain transactions, 98% of them micropayments too small for card rails to handle profitably.

Coinbase isn't alone. Robinhood launched a comparable product the month before. BNB Chain deployed ERC-8004, a verifiable on-chain identity standard for AI agents, in February. Lightning Labs released tools for agents to operate natively on Bitcoin's Lightning Network.

CZ put it bluntly in March: AI agents will make one million times more payments than humans, and those payments will run on crypto.

W3DC: The next wave of crypto adoption won't be driven by retail investors discovering Bitcoin. It will be driven by AI agents that need to pay each other, and crypto is the only financial system they can actually use.

According to Stablecoin Insider’s May 2026 report, the market is now around $321bn, up roughly 41% year-on-year.

The top five stablecoins control around 90%+ of total supply, with USDT and USDC still dominating. Hundreds of stablecoins have launched.
Very few have kept real scale, liquidity or trust.

Stablecoins are easy to create, but hard to make useful.

The winners have more than a peg. They have distribution.

PayPal has merchants.
Tron has low fees.
TON has Telegram.
Circle has institutional trust.
Tether has global liquidity.

The report also shows the market remains overwhelmingly dollar-led. Non-USD stablecoins are emerging, but slowly.

So stablecoins may be the “new cash” in some places, but for now they are mostly extending the reach of the dollar.

W3DC: The next phase of stablecoins will not be won by the best launch announcement. It will be won by trust, liquidity, regulation, redemption and distribution.

Hyperliquid is no longer just a DeFi derivatives venue.

It is becoming a test case for what on-chain markets might look like when they move beyond crypto-native assets.

HYPE has been trading near recent highs, helped by a simple but powerful mechanism: protocol fees are used to buy back the token. That makes trading activity directly relevant to token demand.

And trading activity has not been quiet.

Hyperliquid has seen huge interest in pre-IPO-style perpetual futures, including contracts linked to SpaceX. These are not shares. They are speculative derivatives tied to expected future valuations, but the volumes show something important: traders want 24/7 exposure to assets they cannot easily access elsewhere.

That is the bigger signal.

Hyperliquid is expanding from crypto perps into prediction markets, synthetic real-world assets and private-market-style exposure.

HIP-4 brings event and prediction markets into the stack.

SpaceX-linked markets show demand for synthetic access to high-profile private companies.

Phantom and Ledger integrations make the product easier to reach from wallets users already know.

This is why institutional money is paying attention.

Not because Hyperliquid is simply another exchange.

But because it is showing how on-chain infrastructure can create markets around things traditional finance struggles to offer quickly, globally or continuously.

There are obvious risks.

Synthetic assets are not the same as owning the underlying asset. Pricing can detach from reality. Liquidity can disappear. Regulators will care. And token buybacks do not remove market risk.

But Hyperliquid keeps asking a difficult question:

What happens when an on-chain exchange starts to look less like a crypto app and more like a new market structure?

W3DC: Hyperliquid’s real story is not just HYPE price action. It is the move from trading crypto to trading access itself, private companies, prediction markets, synthetic assets and whatever the market wants next.

Until next time