
AI isn’t just another “bubble”; this is the most important human invention since the printing press. And while chatbots are cool, it’s clear that the future rests on AIs that can make decisions & perform actions for us (aka “agents”).
Agents will revolutionize productivity – they are infinitely scalable intelligence. Within a few years we will have billions of agents performing trillions of actions a day online. Soon, agents will become the majority of online activity.
For AI agents to accomplish tasks for us, they need to be able to access tools & data. This will include software, paywalled content, proprietary datasets, fine-tuned models, specialized AI agents, etc. Agents will need a way to exchange value with other entities to gain access to these resources. In other words, agents need a way to pay for things.
Agents don’t have a social security number. They can’t open a bank account or a credit card. As individual agents complete thousands of transactions per day, they will choose a way to pay for things that’s permissionless. Agents will choose a financial system where they can send payments whenever they want and aren’t limited by credit card processors, regulations, or other human systems.
Crypto at its core is permissionless finance. Blockchain as a technology is the solution for how to solve trust with decentralized digital money. We already have developed the basic building blocks for the AI economy. AI agents are going to use crypto. The next billion crypto users will be AI agents.
Agents don’t have the same UX problems with crypto that consumers have: they don’t care about the annoyance of creating a wallet, managing a private key, bridging between chains, or managing many different types of currency and applications with their varying user interfaces.
Web3 evangelists have spent years talking about a potential Cambrian explosion of consumer crypto adoption. We’ve been told the average person will learn how to use crypto for daily purchases, international payments, and gaming/entertainment. While inspiring, that vision may never happen. Crypto is too complex for average people to grok, and coordinating everyone to switch from current fiat systems may not be possible. Even still, crypto will be bigger than almost anyone imagines. But it will be AI users driving that shift, not humans.
AI agents & AI agent infrastructure are the hottest segments in web2 VC investing. And yet financial infrastructure for AI agents is completely nascent. Currently, AI agents don’t work for most complex tasks – they are unable to make a coherent plan of actions. However, once GPT-5 or GPT-6 comes out (stronger models, longer context windows), then AI agents will be able to plan tasks like a smart human can. Soon, the bottleneck to AI agents flourishing will no longer be planning actions, but rather executing actions (e.g. paying for a piece of software and using it to complete a task). Payments become a primary bottleneck.
Even though blockchains have native transaction capabilities, they need to be optimized and regularized for agents to make maximal use of their capabilities. And in the near term, most vendors don’t accept crypto so agents need a way to transact via credit cards.
For a developer to enable an AI agent to transact, they need an easy way to
Provision funds to AI agents via crypto wallets & virtual credit cards (fiat)
Set daily spending limits & user interface to approve larger transactions
Select appropriate vendors & prevent fraud
Communicate with vendors & automatically troubleshoot payment issues Etc.
We’re building an API that makes it effortless for every agent developer to enable their agents to take action & transact (regardless of what architecture they’re using, whether they’re in web3 or web2, etc).

Michael Saylor and others have suggested that AIs are going to use Bitcoin to transact. Bitcoin is a great store of value, but it's not a great currency. It’s deflationary & similar to human-run economies, it’s critical to have an inflationary currency for a growing economy. Deflationary currencies mean money is worth more if you (proverbially) stuff it under your mattress. In a deflationary environment, actors spend less on software, CapEx, R&D, etc, leading to less GDP and lower growth. Instead, the AI economy needs a currency where more tokens are being minted frequently (like the Fed printing dollars). In addition to being deflationary, Bitcoin is too slow: it can only process 7 transactions per second. Compare that to 24,000 for Visa.
AI agents will be highly rational and choose whatever digital currency is the combination of fastest, cheapest, and most convenient to use. AI agents will choose a digital currency that’s optimized for their payment architecture and designed to be a medium of exchange (i.e. low Tx cost, fast, inflationary).
Over the coming years, billions of agents will be completing trillions of crypto transactions on our rails. It follows that we, as the payment infrastructure for AI agents, will be best positioned to provide agents with a dedicated token. That’s why Agent Protocol is building $AGENT - the medium of exchange for AI agents.

Our payment infrastructure will be compatible with all major currencies, fiat or crypto. Our tooling will be free for developers, and we will simply charge a transaction fee to end users sending funds over our network. But if transacting in our native token $AGENT, transaction fees will be heavily discounted or free, to be decided by governance. This incentivizes $AGENT token to be the default for all agent-to-agent payments, which will become the majority of total payments over the next decade.
The key insight of the agent token is to use speculative demand to our advantage the same way Bitcoin and Ethereum did. We believe that $AGENT is the only way to access AI agent beta, similar to how NVidia has been a liquidity sink for people hoping to get upside in the future of training models. By incentivizing early users and agents on the network we can build an incredible moat and network that will be unstoppable and grow in tandem with the abilities and abundance of Al agents.
We will be opening up opportunities soon to get access to $AGENT allocation and sharing next steps on the project. Reach out contact@agentprotocol.xyz for any inquiries.

AI isn’t just another “bubble”; this is the most important human invention since the printing press. And while chatbots are cool, it’s clear that the future rests on AIs that can make decisions & perform actions for us (aka “agents”).
Agents will revolutionize productivity – they are infinitely scalable intelligence. Within a few years we will have billions of agents performing trillions of actions a day online. Soon, agents will become the majority of online activity.
For AI agents to accomplish tasks for us, they need to be able to access tools & data. This will include software, paywalled content, proprietary datasets, fine-tuned models, specialized AI agents, etc. Agents will need a way to exchange value with other entities to gain access to these resources. In other words, agents need a way to pay for things.
Agents don’t have a social security number. They can’t open a bank account or a credit card. As individual agents complete thousands of transactions per day, they will choose a way to pay for things that’s permissionless. Agents will choose a financial system where they can send payments whenever they want and aren’t limited by credit card processors, regulations, or other human systems.
Crypto at its core is permissionless finance. Blockchain as a technology is the solution for how to solve trust with decentralized digital money. We already have developed the basic building blocks for the AI economy. AI agents are going to use crypto. The next billion crypto users will be AI agents.
Agents don’t have the same UX problems with crypto that consumers have: they don’t care about the annoyance of creating a wallet, managing a private key, bridging between chains, or managing many different types of currency and applications with their varying user interfaces.
Web3 evangelists have spent years talking about a potential Cambrian explosion of consumer crypto adoption. We’ve been told the average person will learn how to use crypto for daily purchases, international payments, and gaming/entertainment. While inspiring, that vision may never happen. Crypto is too complex for average people to grok, and coordinating everyone to switch from current fiat systems may not be possible. Even still, crypto will be bigger than almost anyone imagines. But it will be AI users driving that shift, not humans.
AI agents & AI agent infrastructure are the hottest segments in web2 VC investing. And yet financial infrastructure for AI agents is completely nascent. Currently, AI agents don’t work for most complex tasks – they are unable to make a coherent plan of actions. However, once GPT-5 or GPT-6 comes out (stronger models, longer context windows), then AI agents will be able to plan tasks like a smart human can. Soon, the bottleneck to AI agents flourishing will no longer be planning actions, but rather executing actions (e.g. paying for a piece of software and using it to complete a task). Payments become a primary bottleneck.
Even though blockchains have native transaction capabilities, they need to be optimized and regularized for agents to make maximal use of their capabilities. And in the near term, most vendors don’t accept crypto so agents need a way to transact via credit cards.
For a developer to enable an AI agent to transact, they need an easy way to
Provision funds to AI agents via crypto wallets & virtual credit cards (fiat)
Set daily spending limits & user interface to approve larger transactions
Select appropriate vendors & prevent fraud
Communicate with vendors & automatically troubleshoot payment issues Etc.
We’re building an API that makes it effortless for every agent developer to enable their agents to take action & transact (regardless of what architecture they’re using, whether they’re in web3 or web2, etc).

Michael Saylor and others have suggested that AIs are going to use Bitcoin to transact. Bitcoin is a great store of value, but it's not a great currency. It’s deflationary & similar to human-run economies, it’s critical to have an inflationary currency for a growing economy. Deflationary currencies mean money is worth more if you (proverbially) stuff it under your mattress. In a deflationary environment, actors spend less on software, CapEx, R&D, etc, leading to less GDP and lower growth. Instead, the AI economy needs a currency where more tokens are being minted frequently (like the Fed printing dollars). In addition to being deflationary, Bitcoin is too slow: it can only process 7 transactions per second. Compare that to 24,000 for Visa.
AI agents will be highly rational and choose whatever digital currency is the combination of fastest, cheapest, and most convenient to use. AI agents will choose a digital currency that’s optimized for their payment architecture and designed to be a medium of exchange (i.e. low Tx cost, fast, inflationary).
Over the coming years, billions of agents will be completing trillions of crypto transactions on our rails. It follows that we, as the payment infrastructure for AI agents, will be best positioned to provide agents with a dedicated token. That’s why Agent Protocol is building $AGENT - the medium of exchange for AI agents.

Our payment infrastructure will be compatible with all major currencies, fiat or crypto. Our tooling will be free for developers, and we will simply charge a transaction fee to end users sending funds over our network. But if transacting in our native token $AGENT, transaction fees will be heavily discounted or free, to be decided by governance. This incentivizes $AGENT token to be the default for all agent-to-agent payments, which will become the majority of total payments over the next decade.
The key insight of the agent token is to use speculative demand to our advantage the same way Bitcoin and Ethereum did. We believe that $AGENT is the only way to access AI agent beta, similar to how NVidia has been a liquidity sink for people hoping to get upside in the future of training models. By incentivizing early users and agents on the network we can build an incredible moat and network that will be unstoppable and grow in tandem with the abilities and abundance of Al agents.
We will be opening up opportunities soon to get access to $AGENT allocation and sharing next steps on the project. Reach out contact@agentprotocol.xyz for any inquiries.

What Does It Mean To Know?
In philosophy, "knowledge" is famously defined as "justified, true belief" (the JTB theory). It includes three properties:Belief: A subjective attitude that a proposition is trueJustification: The belief must be held for some reason (e.g. evidence)Correctness: The belief must be trueComparatively, one of the greatest advancements in the cryptocurrency space has been in "zero-knowledge" proofs. Zero-knowledge proofs enable actors to prove that they know the solution to a problem, without revea...

ETH as Money (A New Medium of Exchange for the Digital Age)
Ethereum has faced an avalanche of FUD on crypto Twitter over the past weeks. As BTC hovers near ATHs, ETH is about half what it was at the peak of last cycle. This is despite approval for an ETF and other regulatory tailwinds that should be fueling growth. Common bear arguments include value capture being cannibalized by L2s and slow execution of Ethereum improvement proposals. People are wondering: where does ETH derive its value? And a primary question is whether ETH is money. Money is def...

Consumer Crypto is Stuck (& why AI users are the path to widespread adoption)
Background: Consumer Crypto Is StuckSome say most of the growth that we’ve seen in crypto has already happened. Crypto projects are being used by the same users as last cycle, most just scheming for an airdrop or to make money trading. Public interest around crypto as a utility has waned, and even reversed over the last few years. Peter Thiel has this view, and has sold most of his crypto assets. As is, the path to universal crypto adoption is going to be difficult. That’s because there are f...

What Does It Mean To Know?
In philosophy, "knowledge" is famously defined as "justified, true belief" (the JTB theory). It includes three properties:Belief: A subjective attitude that a proposition is trueJustification: The belief must be held for some reason (e.g. evidence)Correctness: The belief must be trueComparatively, one of the greatest advancements in the cryptocurrency space has been in "zero-knowledge" proofs. Zero-knowledge proofs enable actors to prove that they know the solution to a problem, without revea...

ETH as Money (A New Medium of Exchange for the Digital Age)
Ethereum has faced an avalanche of FUD on crypto Twitter over the past weeks. As BTC hovers near ATHs, ETH is about half what it was at the peak of last cycle. This is despite approval for an ETF and other regulatory tailwinds that should be fueling growth. Common bear arguments include value capture being cannibalized by L2s and slow execution of Ethereum improvement proposals. People are wondering: where does ETH derive its value? And a primary question is whether ETH is money. Money is def...

Consumer Crypto is Stuck (& why AI users are the path to widespread adoption)
Background: Consumer Crypto Is StuckSome say most of the growth that we’ve seen in crypto has already happened. Crypto projects are being used by the same users as last cycle, most just scheming for an airdrop or to make money trading. Public interest around crypto as a utility has waned, and even reversed over the last few years. Peter Thiel has this view, and has sold most of his crypto assets. As is, the path to universal crypto adoption is going to be difficult. That’s because there are f...

Subscribe to Samuel Spitz

Subscribe to Samuel Spitz
Share Dialog
Share Dialog
<100 subscribers
<100 subscribers
No activity yet