25th September 2025 > > Autonomous A.I. agents.

tl;dr

Two developments, one impacting the survival of autonomous A.I. agents, and one which brings a real-life use-case for autonomous A.I. agents a step closer, might be worth pondering in more depth.

Market Snap

Market Wrap

Spot ETF flows have been mixed this week, leaving crypto markets largely directionless.

Curious Cryptos’ Commentary – It’s the future, innit

Or one possible variation.

Truly autonomous A.I. agents will need to sell their services in exchange for cryptos in the form of USD stablecoins, to buy the processing power they need to function. Darwin’s survival of the fittest will be played out in real time in front of our very own eyes, which should prove highly entertaining.

Coinbase and Cloudflare have announced the means of providing the payment rails for A.I. agents:

https://www.coinbase.com/en-gb/blog/coinbase-and-cloudflare-will-launch-x402-foundation

I will let the press release explain:

“The x402 protocol is revolutionizing how payments work on the internet. By leveraging the HTTP "402 Payment Required" status code, x402 embeds payments directly into web interactions, enabling AI agents, APIs, and apps to transact value as seamlessly as they exchange data. This innovation eliminates the friction of traditional payment systems, paving the way for a new era of automated, scalable commerce.”

The use case for x402 protocol is as wide as one could wish for, for it has two key advantages that must be frightening the life out of TradFi. First up:

“No fees: x402 as a protocol has 0 fees for either the customer or the merchant.”

And secondly:

“Instant settlement: Accept payments at the speed of the blockchain. Money in your wallet in 2 seconds, not T+2.”

Source: https://www.x402.org/

Curious Cryptos’ Commentary – It’s the future, innit, part 2

This recent research paper caught my attention:

https://arxiv.org/pdf/2509.15956

Again, let me quote directly for it is as good an explanation as I could muster:

“Blockchain consensus, rooted in the principle “don’t trust, verify”, limits access to real-world data, which may be ambiguous or inaccessible to some participants. Oracles address this limitation by supplying data to blockchains, but existing solutions may reduce autonomy, transparency, or reintroduce the need for trust. We propose Swarm Oracle: a decentralized network of autonomous robots—that is, a robot swarm—that use onboard sensors and peer-to-peer communication to collectively verify real-world data and provide it to smart contracts on public blockchains.”

The proposal for a robot swarm is one use case for the A.I. agents discussed above. The paper goes on to describe how individual robots would be rewarded for accurate data, and would be penalised for misbehaviour, deliberate or otherwise. This carrot and stick approach would be funded by cryptos – in effect, the reward/penalty structure would be built to encourage the right outcome from the swarm. You may be sceptical that it might work, but if so, I encourage you to delve deeper into the research, for there are some convincing arguments in there.

…

What does this mean for my conviction play for LINK, and PYTH, two centralised oracles that are currently positioned to be at the heart of the tokenisation revolution?

At first glance, this is not good news.

I will dig some more, so I can better understand if the concept of a robot swarm blows my investment hypothesis out of the water. I hope not, but if indeed that is my conclusion, I will have no hesitation selling out of both positions, one flattish/small up (PYTH) and one with some excessively juicy gains since the CC Treasury started buying LINK in 2019.

One cannot afford to be sentimental when it comes to investments.