Daily Update

Latest in Blockchain | December 2, 2025

Kalshi and crypto, Vanguard and crypto, AI and crypto. Crypto Twitter had a busy day.

Aren Rendell

02 Dec 2025 — 4 min read

Welcome to the Meridian Update. Kalshi, Vanguard, and AI hype walk into a bar. That’s not a bar you will catch us in. Let’s dive in.

Peanut butter and jelly, crypto and prediction markets

Kalshi is one of the two big companies in the prediction market space. If you haven’t heard of prediction markets, we’re talking about the thing where people take a financial position based on their views about who will win elections, when government shutdowns will end, what the Federal Reserve will do, and more. We wrote about Kalshi back in October.

Anyway, from Kalshi:

“Tokenized predictions powered by Kalshi are live on @Solana”

When we wrote about Kalshi back in October, here’s what we said:

“If the Meridian Update did predictions, which it does not, it would probably predict more tokenized prediction market products in the coming months. As a reminder, tokenization is a major theme of the on-chain world in 2025, according to us and the CEO of BlackRock.”

And that was spot on. The Meridian Update does not do predictions. And here is a tokenized prediction market product in a month that came. And this fits perfectly into a theme of 2025 that we talk about often: people are taking things from the off-chain world and tokenizing them, or making them ownable on-chain. This time, it’s for Kalshi calls “event contracts, where you can buy Yes or No positions with respect to whether an event will happen or not.”

By the way, our friends at Meridian Research got a shoutout in the announcement video, with their app included in the list of front-ends for tokenized prediction markets powered by Kalshi. Pretty fun to see your friends catch Crypto Twitter attention.

Vanguard accepts the reality: People want exposure to on-chain assets in off-chain ways

Vanguard is one of the largest asset managers in the world. It is well known for pioneering index funds and low-cost investment approaches.

Over the last decade, it has also been well-known to Crypto Twitter as a major crypto skeptic. Jack Bogle, the founder of Vanguard, said people should “avoid Bitcoin like the plague.”

In 2025, this has become a difficult position to hold as a massive asset manager with a broad base of customers. Why has this become a difficult position to hold? Well, because in 2025, people want exposure to on-chain assets in off-chain ways. And, well, because other asset managers like BlackRock are bragging about how big their revenues from a Bitcoin exchange-traded fund (ETF) have become.

So, sure enough, from Vanguard:

“Today, Vanguard allows trading of select third-party cryptocurrency ETFs and mutual funds through a Vanguard brokerage account—but we do not offer our own crypto products.”

Crypto Twitter had fun with this. It was a mix of, “finally people are seeing the technology for its potential” and “see, everyone bends the knee eventually.”

Vanguard was careful to clarify that it does not and will not offer its own crypto ETFs. Who knows if that will last or not.

Our favorite part of the Vanguard’s updated view of blockchain assets and product offerings was, of course, at the very end:

“Trading in cryptocurrency ETFs and mutual funds may involve significant risk and may not be suitable for all investors. If you’re considering investing in cryptocurrencies or crypto-related funds, we encourage you to:

- Understand the product’s structure and underlying exposure.
- Assess risk tolerance and investment horizon.
- Consider how the product fits within the broader asset allocation.
- Review the fund or ETF’s prospectus and disclosures carefully.”

Right indeed, right indeed, and the Meridian Update also encourages everyone to do those things. And yes, as of publishing, they did have three spaces between “investors.” and “If you’re…” Weird typo. Three spaces. Something indeed.

We don’t mention AI often (you’re welcome)

This newsletter is also not written using artificial intelligence (AI). We honestly don’t like the stuff that much. Though we do occasionally use the “Ask Meridian” feature from our friends at Meridian Research to make sure we haven’t missed any stories.

Anyway, this story isn’t about how we are one of the few newsletters that isn’t all AI this, AI that. It also isn’t a way for us to brag about how we actually write this thing. Every day. With thought, care, and genuine joy. Not because we have to. Because we want to, and it’s interesting, and someone should.

No. It’s about this:

“...[AI models] developed exploits [of blockchain smart contracts] collectively worth $4.6 million, establishing a concrete lower bound for the economic harm these capabilities could enable…[AI agents] uncovered two novel zero-day vulnerabilities and produced exploits worth $3,694, with GPT-5 doing so at an API cost of $3,476. This demonstrates as a proof-of-concept that profitable, real-world autonomous exploitation is technically feasible...”

Anthropic is a late-stage startup that builds AI models. It is widely respected in Silicon Valley. While less-well-known than ChatGPT, its models are increasingly used by enterprise customers. Some extremely smart people at Anthropic do something called red teaming. Red teaming is basically hacking to avoid actual hacks. Try to hack the system to figure out how it can be hacked. Then, when you hack it, tell someone to fix the issue that allowed you to hack it.

The Anthropic people were excited because they had established evidence that AI systems could hack things and that there would be financial cost. They aren’t red teaming blockchain technology vulnerabilities. They are red teaming societal vulnerabilities to AI.

We were excited about the news because it’s fun, and it’s some proof that blockchain systems may have a new partner-in-red-teaming. Hacks are nothing new in the blockchain world. System hacks are costly, because blockchain systems deal with genuine stores of value. They happen all the time, frankly. One of the most famous was when someone exploited issues with systems built by Wormhole and took hundreds of millions of dollars.

It turns out AI may be a useful partner in identifying vulnerabilities. Something to keep an eye on.