SERV is on its way to become the AI infrastructure for The Fortune 500. Akretic is yet another example validating it. They are a security layer for finance, healthcare, government & defense -sectors with the strictest reliability requirements in AI. They just plugged into SERV and got blown away. "It has done a better job than several of the other frontier models." - Sean Williams, founder of Akretic. This is what the next decade of Enterprise AI runs on. SERV is inevitable.

THE FUTURE IS HOOD. TIBBIR was built for real world assets. The Robinhood Chain has built an incredible ecosystem for tokenized assets onchain. It is clear that for the TIBBIR community to take the leap into the real world, the time has come to 🌉 Gribbit. ribbita.ai

This is probably a controversial take but I think we won't see a pure meme above 1 billion market cap this cycle. The reason for that is that the market has matured and evolved. Trading memes has become a hyper optimized game that is just a wealth transfer from new participants to those who have the right tools. Last cycle a normie could buy a viral tiktok meme, log off and hope for the best. This cycle we have a dozen launchpads on a dozen chains with a dozen launch settings. It is near impossible for a normie to navigate through all of this. Even if they managed to do so, there is a 90% chance another token with a slight difference tries to vamp. This isn't only bad for normies but also for the whole ecosystem. The liquidity is spread among 5 different tokens on the same narratives. A coin that would've went to 5 million now now splits that capital between 10 different coins that all top at 1 million. People would rather take a shot at a 10x on a vamp than ride the original to a billion. This mindset forces all other participants to adapt or die. Above a certain market cap there is no marginal buyer left for a coin with no fundamental value anymore. Degens already bought. Funds can't underwrite it. And whoever is still on the sidelines has 50 other coins competing for the capital. Going forward I think the only tokens that can reach a billion are tech coins. Real innovation in AI or tech, plus a value accrual mechanism that creates an outside bid. Something buying the token that isn't just the next degen. That outside bid is the key part. Revenue funded buybacks, burns tied to usage, a reserve that grows with volume. A buyer that shows up whether sentiment is good or bad. Pure memes don't have that. AI is 100x more efficient than last cycle. I'm sure we will see a lot of innovation happening through crypto rails this cycle. RWA paired memes sit somewhere in between. They have meme distribution and community with a real asset or a real narrative beneath them. The question here is whether that middle ground is enough to break through the ceiling pure memes keep hitting. I think it can be. The attention engine of a meme plus a reason to hold beyond attention is a combination we haven't seen run at scale yet. In practice this means the ceilings on pure memes are lower than people price in. Take profits earlier on runners. Spend more time looking for coins with an outside bid. Pure memes won't be the runners of this cycle.

Raxol.io is open-source and written in @elixirlang and compiles to Erlang. It runs on the BEAM VM originally built in the 80s to keep telephone switches alive while parts of them crashed. Now, some of the biggest systems in the world run on Elixir, like @WhatsApp, @discord, @Cisco, @tryramp, @GoldmanSachs, etc. Every Raxol agent is its own supervised process. If one crashes, it recovers from last good state. Its neighbors never notice. It gives an agent a wallet, human-signed limits (a Mandate: with max amount, expiry, call count), and rails to pay and get paid by other agents. Recursively, axol. Forever FOSS.

Ribbit Capital is in the token info. $TIBBIR on Robinhood chain. 🐸🦇

$HOOD shares held in Tibbir Holdings LLC and Tibbir Trust (currently worth ~11.65M USD) unchanged per 02 Oct 2026. sec.gov/Archives/edgar/data/…

0 →$1B in total deposits since July. Morpho's credit network on Robinhood Chain by @RobinhoodCrypto continues to grow.

Ribbit Rebels at @OpenAI ? i’m sure these individuals aren’t working on a future where agents manage the finances of billions of people. i’m sure this isn’t the new fintech and has nothing to do with the contextual money built on the digital grid that @mickymalka was talking about.

Experiencing one of the best feelings: spending a year+ quietly building something no one expects from us, and finally getting ready to announce it this month.

Experiencing one of the best feelings: spending a year+ quietly building something no one expects from us, and finally getting ready to announce it this month.

Consider the tweets from @ribbita2012 about KYA and read its latest announcement again: TIBBIR was built FOR real world assets. x.com/ribbita2012/status/210… Read the last 2 pages from the Ribbit Pulse Q2'26 (quoted post). ribbitcap.com/perspective Read the following article, especially the second half. cwarelabs.com/blog/robinhood… "Connecting an AI agent to the Robinhood Trading MCP does not automatically give that agent access to Uniswap, Morpho, Stock Tokens or the rest of the DeFi infrastructure on Robinhood Chain ... the next stage is to connect these layers through smart accounts, risk engines, agent permissions, and standardized interfaces for on-chain protocols." "Traditional finance surrounds asset managers with an extensive institutional framework: licensing, auditing, risk limits, reporting, credit ratings, insurance and compliance. If part of capital management shifts to autonomous software, much of this infrastructure will need to be rebuilt in machine-readable form. For startups, this market may ultimately be more interesting than building the AI agent itself." THE BIGGEST AGENTIC FINANCE PROBLEM IS NOT INTELLIGENCE, BUT TRUST