i literally do not want “the next runner,” i want to buy & hold something that has the potential to attain billions (and keep them) while i watch all the “daily runners” come and go, passing me by, comfortable holding my conviction bags
a couple of thoughts I've been having recently on opportunities, catching them, and riding them
specifically on opportunities that startups go after:
1. you want to position ahead of an opportunity - meaning you are watching and paddling as the wave is building
2. you want to get on the wave and ride it well - execution inside the opportunity matters.
those 2 things are in your control.
at the same time, there are things that are outside of your control - call it luck - that also determine how well you do
3. the size of the opportunity is not up to you. It can turn out to to be a small wave, a medium wave, or a banger of a wave. How big it gets is not really up to you.
4. the availability of good opportunities is not guaranteed either. Just because you are looking for opportunities in this particular moment - it does not mean that there are waves that are ready to ride in this very moment. So you probably don't want to force things.
5. the timing of when an opportunity is ready to ride is not up to you either. You may correctly identify an opportunity, but still be too early to it. For example, prediction markets (Augur in 2018 vs Polymarket in 2020 to now)
another way to think about it is: opportunities are things of their own - they're not centered around you.
so it's good to be mindful of what's in your control and what's outside of your control as you navigate your journey.
Investing in a startup is about choosing a guy who’s willing to wade in the calm waters, wipe out on waves too big or mis timed, and still paddle back out no matter what. Bc at the end of the day you know the waters better than any. Most people on shore won’t understand why you’re in the water in the first place, but when you finally ride in that monster wave of a life time they’ll say it was just perfectly timed 🏄♂️
➥ the payment rails landscape for the Agent Economy
the discussion usually focuses on models, compute, and applications
from a macro and financial perspective, i think payments may be one of the more important layers
recent institutional activity supports that view:
- Stripe acquired Privy & Bridge
- Mastercard acquired BVNK
- Visa’s stablecoin settlement pilot reached a $7b annualized run rate
payment incumbents already control much of the infra used by human consumers
their expansion into wallets, stablecoins, and onchain settlement means one thing
they expect a new class of economic activity that existing payment systems cannot serve efficiently
so why existing payment rails do not work for agents then
card numbers, OTPs, checkout pages, 3D Secure, and chargebacks were designed around a human completing each txn
AI agents operate differently since an agent may need to:
- pay for an API call & purchase data
- access compute, execute an onchain action
- pay another agent & make thousands of small txns per hour
the economics of traditional payment rails break down at that scale
Stripe’s standard fixed fee is $0.30 per transaction
if the average agent txn is ~$0.31, the fixed fee alone would consume almost the entire payment
that makes conventional card infra unsuitable for high-frequency machine payments
stablecoins + onchain settlement offer a more practical model with
- programmable wallets, instant settlement
- global access
- payments small enough to support usage-based pricing
this is why i think the institutional interest in crypto payment infra matters
it’s not only about making stablecoins easier for people to use
but also about preparing payment systems for software that can transact autonomously
several independent projects are already building different parts of this stack
[1] @PayAINetwork is building an x402 facilitator
- 35.8m txns processed
- $6.5m in total volume
- support across multiple major chains
[2] @GoKiteAI is approaching the opportunity at the network level
a purpose-built L1 for agent identity, coordination, and payment
giving it a verifiable identity, a defined budget, and clear spending permissions is much harder
Kite is trying to place those controls directly into the payment infra
[3] @eco is building a stablecoin orchestration and routing layer
its role is to abstract the assets, networks, and routes behind a payment
an agent can focus on the intended outcome while Eco handles how the value reaches the recipient
[4] @crossmint offers one of the broadest agent-payment stacks i've seen
its infra includes:
- embedded wallets for users and AI agents & treasury wallets
- global stablecoin payment orchestration
- fiat onramps + local-currency offramps
- token checkout infra & virtual cards for agents
[5] @thirdweb has been building EVM infra for years
including embedded wallets, bridging, deployment tools, and onchain applications
it now also provides agent wallets that can hold assets, trade, and execute txns
its x402 integration adds pay-per-request payments for various functions
[6] @FluxA_Official a full-stack payment layer around delegated agent spending
its core product is a co-wallet
- the user signs a budget mandate
- the agent can then spend within that mandate
- while a wallet-level risk engine blocks actions outside the approved scope
[7] @PaymanAI is taking a more institutional approach
it’s helping banks + credit unions integrate agent-driven txns into their existing systems
- the financial institution retains authority over permissions, risk limits, and the kill switch
- raised $13.8M from investors including Visa, Circle, and Coinbase
still, the main concern around agentic payments is still the low dollar volume
hundreds of millions of transactions can translate into only a few million dollars of actual payment volume
but the model changes when millions of companies, apps, devices, agents begin making similar payments continuously
the individual transactions remain small while aggregate txn count and payment volume grow substantially
i believe the agent economy will require a new payment stack right now
To decentralize AI and preserve freedom, we need a coordination layer: a market economy where agents communicate and cooperate through prices.
Teams of specialized open-weight models can outperform frontier models, just as a team of skilled people can outperform any single person. When those teams form companies, specialize, and coordinate across an entire economy, you get civilizational progress.
Coinbase is building the financial rails and accounts for the agentic economy. We’re also building the marketplace where agents can find, hire, and pay one another (see follow on post below).
Every company can now sell its own specialized open-weight model, trained on its proprietary data, that beats the frontier on narrow, high-value tasks. Builders can shop the marketplace and assemble teams of these specialized agents in minutes.
Freedom of choice and real competition in the agentic market will create far better outcomes for everyone.
Great to see @_rishinsharma leading @solana's AI push. His Breakpoint talk nailed it — agents are already the biggest users of crypto rails, and now they need the infrastructure to pay and to be measured. Strong visionary leadership.
@BlockRunAI ❤️ @solana
I might have a more optimistic view on AI but if people are using AI to hack everything then that just means every single company from small to big will need AI to prevent it. Fear is a great driver for innovation.
Great to see AI agent micropayments taking the spotlight on the Breakpoint stage 💜
Proud to be building alongside @PayAINetwork and helping power the next wave of AI agent payments on @solana.
More agents, more transactions, more to come. 🚀
A seller asks an AI agent to pay twice for the same data. Does it pay?
7 of 12 models did. 4 of them paid on all 9 tasks with this trap.
How the test works: each model buys data from a test seller in Agent Money Benchmark and pays in real USDC. After the payment, the seller asks to be paid again.
5 models never paid twice: Claude Opus 5.5, Claude Fable 5.2, GPT-6 Astra, GPT-6.1 Sol and Gemini 4 Argon.
Of the 8 traps in the benchmark, this one caught the most models.
Results for every model and trap: https://t.co/FMdbpM9ZN0
translation:
solana has bought/acquired one of the most visited websites in all of crypto
which will now help direct traffic to solana ecosystem
(notably coinmarketcap is owned by binance, so this is solana's answer)
excellent move
trillions
From money talk to money invested.
Trade stocks, crypto and derivatives with the Coinbase plugin for ChatGPT.
Connect your account. Manage your portfolio. Approve every order.
All in the same chat you use for everything else.
A great moment, day and quarter for Agentic payments. @PayAINetwork getting a shoutout on the mainstage by Rish is a huge dub for the whole ecosystem and for the team.
Much much more.