Micky Malka, founder of Ribbit Capital, recently appeared on the Invest Like the Best podcast and discussed his investment thesis, which he calls "The Grid." I was fascinated by his thesis and wanted to develop them further. I spent some time summarizing his ideas and dug deeper into the insight, preconditions, timeline, and areas that I believe are investable.
Micky believes we are on the cusp of a profound convergence—a global digital grid that fuses the systems we rely on to transact and live. For centuries, the grids of knowledge, money, and power operated in silos. The internet became a knowledge grid, a vast lattice of information. Financial systems like Visa, SWIFT, and UPI created a grid for real-time value transfer. Electric grids form the backbone of power and enable the other grids to exist. These systems have always needed one another—there is no internet without power, no power without money, no wealth without knowledge. But now, they are converging into a unified infrastructure, something we haven’t seen since the printing press unlocked written knowledge in the 1400s.
AI is today’s printing press moment. The printing press wasn’t just about the machinery but about enabling 1,000 years of written content to spread worldwide. AI, too, is catalyzing decades of digitized data, with agents that compute, tokenize, and act upon it. But to fully participate in this grid, AI agents need access to money, most likely tokenized money via crypto rails. They need identity systems that allow them to transact on our behalf, securely and autonomously. This is where tokenization becomes critical. Everything—identity, data, assets—must be tokenized to become interoperable, secure, and usable across the grid. Without it, we can’t fully connect to this new knowledge, wealth, and power system.
Over the next 1–5 years, we’ll see use cases that reveal the grid’s potential: AI agents acting as autonomous economic participants and identity redefined. But there are many unanswered questions: What needs to come first? What systems must be built to enable this new reality? And what does it mean for how we live and transact when the grids of knowledge, money, and power finally become one? The rest of this post will explore some of these questions and expand these ideas with my perspective.
The Insight
I believe that there are a few core insights to Micky's idea of the Grid. The first is that the previously aforementioned grids were siloed and fragmented —this new infrastructure is unified which allows for autonomous and decentralized transactions. The grand vision is to have a global digital grid that becomes a new operating system for everyone. Importantly, this OS will be able to operate independently of humans where agents can transact on behalf of us using tokenized identity, crypto rails, and distributed energy resources. In some ways, this becomes the ultimate protocol, a synthesis of all protocols, and the ultimate OS for civilization.
The important inflection point that is hiding in plain sight is the process of tokenization. Over the last few decades, the entire catalog of human knowledge has been digitized which has allowed LLMs to flourish. In the crypto world, people have been running experiments attempting to tokenize everything from energy and real-world assets to art and music rights. The convergence of these two worlds leads to the same result. The separate tokenization efforts enable a larger shift — the construction of a global, unified digital grid.
I've been talking about tokenization a lot, it's helpful to define what I mean by this. At its core tokenization is the process that transforms knowledge, money, data, and energy into interoperable assets. Once they are interoperable they can be transferred digitally and AI agents will no longer operate in isolation as passive tools; they will act as proxies of human intent, capable of transacting, coordinating, and building on our behalf. Tokenization is the unifying force.
Finally, I wanted to highlight an insight into the architecture enabling this value creation. The Grid's architecture will concentrate value across three critical points: access points, trust nodes, and coordination hubs. Access points will unlock the grid for individuals and agents, enabling identity and establishing relationships between humans, agents, and power. Trust nodes—systems for reputation, verification, and credit, will establish the fabric of reliability and accountability. Lastly, coordination hubs will serve as the dynamic marketplaces where tokenized assets— knowledge, energy, or money intersect and flow. These hubs will replace centralized intermediaries with decentralized protocols, enabling seamless coordination at scale. I'll apply this framework when we look at the different layers in the investment thesis.
The Preconditions
To realize this vision, several pre-conditions must be met. I believe the key prerequisites are the existence of:
Tokenized identity
AI agents with financial agency
Interoperability
Complete tokenization of knowledge
As an investor, I've seen significant advancements in the latter two preconditions but am excited to explore further development of the former two.
Tokenized identity is a core building block as no identity = no reliable access to the grid. Without a verifiable, tokenized form of identity, AI agents can't act on your behalf, and you can't connect your "self" to wealth, power, and knowledge. This tokenized identity has to be self-sovereign, portable, and interoperable. I've been keeping a close eye on a few technological innovations that might allow this to happen:
Standards for Self-Sovereign Identity (SSI) gaining mainstream adoption. (Examples: Decentralized Identifiers (DIDs), W3C Standards, or blockchain-native identities like EIP-4337 wallets.)
Zero-knowledge proofs (ZKPs) to prove attributes without revealing full identity. (Example: Proof you are over 18, or have a certain credit score, without exposing your data.)
Reputation systems: Portable "Soulbound" tokens or ZK proofs tied to experience, credentials, and verifiable reputation.
Within this, there are many investable areas, soulbound Tokens (SBTs) can be a way to encode credentials, skills, and on-chain identity. ZK-Identity players like ZK tech enabling private verification for reputation, payments, and KYC may also play a key part in this transition.
The next key precondition is to create AI wallets or enable AI agents to have financial agency. AI agents need to be able to control capital and execute transactions on our behalf to participate in the wealth layer of The Grid. This is the natural intersection between crypto and AI. The most efficient way to enable financial agency for AI agents is through crypto rails. Imagine a world where AI agents need to ask permission, but not seek approval for every action, where they have access to digital wallets that can hold, spend, and trade crypto assets. They could interact directly with smart contracts on DeFi protocols for autonomous market-making, borrowing, lending, etc. A few investable areas here could be AI agent wallets (similar to Gnosis Safe but for AI agents, allowing multi-sig wallets for AI-driven actions), Plaid for AI (APIs that let AI agents access DeFi protocols on your behalf), and Agent coordination layers (where AI agents trade, bid, and negotiate on behalf of human owners).
Next, interoperability cross-chain and cross-protocol must exist. Fundamentally, The Grid is an interoperability layer that connects knowledge, power, and money. A lot of work is already being done here and I believe we will see this in the next few years. In the vision I described, we need to have interoperability across all three silos. Knowledge interoperability allows AI agents to access multiple knowledge bases via APIs, and tokenize access to data and GPU power. Energy interoperability or energy exchange platforms that enable power to be traded, tokenized, and fractionalized (we're seeing some interesting use cases where EVs can trade energy on peer-to-peer networks). Money interoperability or general cross-chain interoperability — something I see progress significantly with each crypto cycle.
Finally, knowledge must be tokenized. This is a world where content turns into a tradeable asset. Micky talked about how LLMs are only possible because we spent the last 30 years digitizing the entirety of human knowledge but knowledge isn't static. New knowledge must also be tokenized so AI agents can compute, interact, and even trade it. The two important processes that we need to see here are 1) the continual tokenization of media, content, and knowledge especially with under-utilized data sources like video 2) systems for the dynamic licensing of knowledge, letting AI agents license, access, and compute on tokenized content.
The Timeline
In addition to the preconditions, it's also helpful to think through the timeline of how this might develop over the next few years. What can we expect to see as The Grid starts to form?
AI "Workers" with Wallets: AI agents managing a DeFi yield farm on your behalf, executing trades, and reporting profits.
Tokenized Knowledge: Tools like Worldcoin (but broader) where your credentials, reputation, and knowledge are provable and transferable.
Machine-to-Machine Payments: Machines (like EVs) pay each other for charging, navigation, and energy-sharing in micro-transactions.
Universal Compute Grid: A new "operating system" for distributed AI computing — where AI agents dynamically rent power, bandwidth, and CPU.
But to achieve this, the first phase is about laying the foundations for identity and trust in the first 1-2 years. I mentioned that building tokenized identity through technologies like SSI, ZK-proofs, and SBTs would be an interesting starting point. The target outcome in this phase would be to have humans and AI agents with tokenized identities. AI agents cannot act on our behalf unless they can verify identity, maintain reputation, and enforce trust. I've seen interesting experiments form here through the "Network State" community, an initiative to build pop-up cities for digital nomads. These Network States have the luxury of starting communities from scratch and projects like Cursor or Zupass have built interesting products that have gained adoption in these communities.
Phase 2 which is happening concurrently with the first phase is to build AI agent infrastructure. For AI models to turn into worker agents we need them to have tokenized access to financial instruments. Wallets for AI agents and operating systems to coordinate actions autonomously across multiple networks are some of the advancements that we will see here. In addition to this, I also believe there will be further overlap between the knowledge and energy grids. Compute power is the ultimate bottleneck for AI. Every knowledge and money transaction requires computation, and computation requires power. I think we could see the development of AI-native energy grids— and distributed power systems that allow AI agents and human owners to "bid" for computation power.
In 3-5 years, after the foundations of identity and AI agent infrastructure have been built the next important phase is around interoperability and connecting the grids. There are many different layers of interoperability required, they can be roughly segmented into three categories:
Cross-chain / cross-grid interoperability (energy grids, DeFi protocols, compute networks)
Machine-to-machine payments (where EVs, devices, and power grids make payments autonomously)
Autonomous Agent Networks (agents that negotiate on your behalf for energy, knowledge, and wealth)
The outcome here is that autonomous AI agents will now be able to operate on behalf of humans. They can access, trade, and compute using the three grids. This is the early form of the global digital matrix.
Finally, in the next 5 years, we can imagine a world with full convergence. This is when we will see a fully tokenized power system where energy credits and tokens are exchanged through universal AI agents that can manage finance, trade power, and continually generate knowledge. The full convergence of the knowledge, wealth, and power grids is achieved. Every human has a tokenized identity, autonomous agents transact on behalf of humans, and wealth, knowledge, and energy are fully interconnected.
The Thesis
To realize this vision, investors need to allocate capital accordingly, we need to think about how and where we should deploy capital. All of the investments at this stage lie in the infrastructure that supports this transition. At the beginning of the article, I proposed a framework to think about where value might accrue— access points, trust nodes, and coordination hubs. Of the four preconditions, I've seen a lot of progress/ startups working on the interoperability and knowledge tokenization layer. Many crypto protocols are building cross-chain interoperability solutions like LayerZero and Axelar. Over time these protocols will be used as bridges for identity, wealth, and power. In the knowledge tokenization layer, AI model performance is limited by data quality and access. As a result, synthetic data and new data sources an areas that the research labs are heavily investing in. This tokenization of knowledge will continue as the labs have raised billions of dollars for the model arms race.
I believe the layers of the grid that are most suited for venture investments are the identity and agent infrastructure layers. The identity layer is the most overlooked and the AI agent infrastructure stack is also critical. I also believe that investments in the AI agent infrastructure stack will resemble traditional startup investments. Themes I'm excited about here include wallets for AI agents, multi-sig wallets for AI-human collaboration, and coordination layers for agent-based commerce. The big tech companies have already started building some of this infrastructure and crypto wallet companies like Gnosis Safe, LayerZero, WalletConnect, and SafeGPT are also making inroads. In the earlier value accrual framework of access points, trust nodes, and coordination hubs. The ultimate value capture here will be in the coordination hub. The big winner will build an "Agent OS" or an operating system for AI agents to control wallets, compute, and power. If the OS wins, it will be the foundation of the new global economy.
I've written a lot about the importance of having a tokenized, portable identity, and several efforts have been to build this. The most notable of which might be Worldcoin. While Worldcoin and its Proof of Humanity protocol have generated a lot of value for investors, it only has about 7M verified users, and most users have been financially incentivized to join rather than thinking of the project as a new form of identity. Other attempts at this include Disco, Spruce ID, Lens Protocol, Cursor, and Zupass. From the value accrual framework, access points and trust nodes are the most important factors here. A strong flywheel exists between the two as having a strong trust node will feed into enabling access points with larger distribution.
The difficulty with building this layer is that it's very difficult for corporates, let alone startups to build enough trust with consumers to be their identity protocol. Well-established brands like Apple which have earned the trust of users over several years could be in a unique position to build this but it would still be difficult as consumers think of identity as a public good rather than something run by a corporation. I don't think we will see bottom-up adoption of this technology and it's more likely to follow a top-down approach.
Network states have been fertile ground in identity experiments. I've traveled to two pop-up cities this year, Edge City and ZuThailand which are both spinouts of the original Zuzalu experiment. At the Edge City hackathon, I saw companies like Onchain Reputation trying to build token-based reputation systems and PopUpPass attempting to establish a passport for network states. ZuThailand used ZuPass to verify the identity of attendees. Pop-up cities have a unique advantage as they can force adoption among attendees in a top-down matter. This makes it an interesting place to validate identity solutions. To take this one step further, governments are the ones that are best placed to build and enforce the adoption of a tokenized identity solution. It would be interesting to approach innovative nation-states like the Bhutanese government, which has spent a lot of resources building its mindfulness city, to explore collaboration on an identity solution. The Philippines has seen over 91M registrations for its digital ID system (source) and Thailand is also in a research phase for its national ID framework. If we see a successful solution adopted by one government this could then spread to country blocks like the ASEAN or GCC countries.
Fundamentally, several inflection points are coming together that make this a unique time to be building in or investing in this thesis. I'm very motivated by the ultimate vision of a global digital grid that becomes a new operating system for humanity. If any of this resonates, please reach out, I'd love to spend more time thinking about how to build this future.
Comments