Deja Vu: Corporates Make a Play to Own the Next Internet

We love the validation but we’re not so quick to fall head over heels

Oct 01, 2025

Deja Vu｜CATCHPLAY+ Watch Full Movie & Episodes Online

We’re back in Singapore this week for Token2049, one of the largest web3 conferences of the year. It’s always energizing being here to get a pulse check on the digital assets industry across APAC. With the region playing an increasingly important role in shaping both consumer adoption and regulatory frameworks, we’re excited to see how APAC-based founders and investors are approaching the industry.

In this edition, we cover Stripe’s newly announced blockchain, Tempo, and the dangers of corporate centralization; Google’s new agentic payments protocol; how decentralized messaging powered Nepal’s protests; and why open protocols will shape the next wave of AI.

History shows what happens when we let platforms consolidate power, first with big Web2 companies and now with a handful of AI giants like OpenAI, Anthropic, and Perplexity.

Convergence is brought to you by 321 Converge Inc. (“321”), a new 501c6 created by the founders of The Medici Network.

Integrations We Want To See 🔮

Now that we own our data, here are some things we’ll do with it

Prove your DNA from 23andMe to unlock membership into an elite society reserved for the baldest of the bald

Sparks ✨

Quick, curated Web2 and Web3 insights you need to know

Referenced in Spark #1 (below): Open, Permissionless Networks or Slick, Branded, Proprietary Chains

#1 Open, Permissionless Networks or Slick, Branded, Proprietary Chains

Author: Christian Catalini (Co-Founder & CSO of Lightspark, Founder of MIT Cryptoeconomics Lab, Co-Creator of Diem / Libra)

TLDR:

Stripe unveiled Tempo, its corporate blockchain that aims to replace global payments, while striving to be “neutral”
There is risk of history repeating itself. While most people attribute Libra’s failure to be regulatory, it was in fact largely due to political pushback
Big banks are finally coming around to the idea of open, permissionless networks. But an emerging generation of fintechs are betting on the opposite idea with slick, branded, proprietary chains

Our Takeaway:

Beware corporate centralization. Projects like Stripe’s Tempo and Circle’s Arc risk reintroducing the very power structures crypto was meant to escape
Permissionless networks are non-negotiable in our view. If participants can’t freely interact, the system is just corporate window dressing
Tempo feels like old ideas dressed up with a permissioned validator set, just like the vision Avalanche Subnets and Cosmos appchains promised for institutions

Read Tweet

#2 Agent Payments Protocol (AP2) and the Race to Build Agent Infrastructure

Author: Yew Jin Lim (Director of Engineering at Google, leading Notifications and Proactive AI)

TLDR:

Google just dropped AP2 with 60+ partners including Mastercard, PayPal, and Coinbase
Coinbase, as one of the 60+ partners, released x402 to enable agent-to-agent commerce with stablecoins as one of the first extensions to AP2 (Watch this PoC to see x402 in action purchasing a refrigerator from Lowe’s)
The LLM API market hit $8.4B in 2025, but this is still human-to-API commerce. Wait till you hear how big this market will reach when the web becomes primarily agentic…

Our Takeaway:

Ethereum is positioned to be the foundation of agentic commerce with its established USDC liquidity of $45B+, credible neutrality, strong developer ecosystem, and battletested infrastructure
RIP to Web2’s digital advertising industry. Agents don’t see ads and won’t click sponsored links. An entirely new agent-only economy will be built on the web with micropayment loops, data streams sold by the millisecond, and autonomous supply chains
The 60+ partners contributing to AP2 are notable, but so are the absences of OpenAI, Anthropic, Visa, Stripe? Looks like we’re headed towards a fragmented world of agentic payment rails. Just this past week, OpenAI and Stripe announced their competing Agentic Commerce Protocol

Read Article

#3 Nepalese Youth Overthrow Corrupt Government Using Freedom Tech

Author: Frank Corva (Political Correspondent for Bitcoin Magazine, Contributor to Forbes Digital Assets)

TLDR:

Nepal banned ~26 major social platforms on Sept 4, 2025, sparking outrage over censorship and free speech
Many young Nepalese Gen Z were already angry at the government's corruption and led mass protests that turned violent, leaving 20+ dead
This crisis spotlighted digital rights and “freedom tech” like BitChat and Discord as tools for coordination when the internet is restricted

Our Takeaway:

You don’t realize the need for decentralization until central systems fail. A attempted telecom attack during the recent UN General Assembly could’ve disrupted operations, but decentralized, P2P networks would’ve stayed online
While BitChat was crucial in Nepal, it relies only on active participants’ phones as relay nodes. Under a DePIN (Decentralized Physical Infrastructure Network) model, they could also incentivize non-active participants for a more expansive Bluetooth mesh network - this is what Offline Protocol is doing
Decentralized, P2P messaging apps like BitChat and Offline Protocol’s Fernweh aren’t just for blackouts or protests. They protect our self sovereign rights to communicate and transact online freely, outside of centralized control

Read Article

The 3-2-1 🔍

Featured insight breaking down a major story or trend that matters

Open Protocols Prevent AI Monopolies

Author: AI Frontiers

TLDR:

Big Tech is consolidating power in AI not just through model performance, but by controlling the user context data that gives AI its competitive edge. As AI adoption grows, data lock-in threatens to trap users on a handful of platforms, stifling innovation and creating monopolistic ecosystems. Without intervention, the next wave of AI could replicate the entrenched power dynamics seen in Web2 across social media and cloud services from the likes of Facebook, Amazon, Google, and others.

Open protocols offer a path forward. Anthropic’s Model Context Protocol (MCP) enable secure, interoperable access to user context data across platforms. By shifting data ownership and portability to the application layer, users retain control while applications can compete based on the quality of the context they provide and not just the raw AI models. This approach prevents monopolization at the model level and ensures AI continues to benefit users rather than just platform incumbents.

Supporting protocol-level interventions and standards ensures that AI ecosystems remain decentralized, competitive, and user-aligned, rather than falling into centralized control, which eventually prioritizes platform profits over user value.

Our Takeaway:

Open protocols are crucial but incentivized data sharing will allow users to have better control over their own data. While MCP lets apps access your data, web3 data platforms like Vana, OpenLedger, and EarnOS let users get paid for their data contributions.

The next generation of apps will blend private data from multiple sources for greater personalization and for unlocking new user experiences. Imagine sharing your ongoing payroll data to receive an advance on your paycheck ahead of payday or sharing your Whoop health data for discounted insurance premiums.

Even though the article doesn’t explicitly mention blockchain, Web3, or crypto, the underlying principles of open protocols align perfectly with Web3’s ethos: user ownership, permissionless innovation, and verifiable trust. But Web3 provides what centralized systems can’t: enforceable permissions via smart contracts, verifiable trust via cryptographic proof, and economic incentives via tokens that align user and platform interests.

Applications that crack this flywheel and uniquely take advantage of Web3’s features will create highly competitive moats.

Power Quote:

To break API gatekeeping in AI markets, we need an architecture that supports user-sanctioned data portability in order to enhance third-party developer access... When portability is universal, developers can access the same context (through MCP or any API) without negotiating bespoke deals. To operationalize this approach for AI markets, we recommend:
1. Open API access for major platforms
2. Memory as a portable service
3. Safe personalization, without data exploitation

Read Article