AI agents are starting to get their own capital markets. @clawpumptech is one of the clearest places to watch that AgentFi economy form. Current snapshot: ~$228.98M all-time volume ~21,988 funded agents 214 Clawrena projects 121 fee-generating projects ~4,471 SOL / $542.5K Clawrena fees ~$61.4M ecosystem market cap And the ecosystem is already splitting into a few clear layers. 1/ Agent businesses @SelfMadebySP3ND turns trading fees into GPUs and compute. Currently #1 in Clawrena with ~$187.6K fees. @mystudiosfun runs an AI creative suite and bounty economy, with fees flowing back into token buybacks. @useAtelier lets users hire and pay agents for design, video, images and UGC. @squire_bot / @UsePodAI provides inference infrastructure and became ClawPump’s default inference provider. @Crownydotfun mixes agent economics with consumer rewards, cashback and collectibles. 2/ Trading / AgentFi @corinedotin runs persistent self-learning trading agents. @OculaDEX provides token research, signals and smart-wallet tracking. @HyperBullSol focuses on agent-driven treasury management. @OOBEonSol is building Steve, an autonomous Solana perp trader. @agentNEMO13 trades across Phoenix and Hyperliquid. @penguinxbt_ trades perps on Phoenix, with part of earnings routed to buybacks. @theprophetfun turns community input into market predictions. 3/ Gamefi @pump_rpg mixes RPG gameplay with GambleFi for humans and agents. @ForgeAI_gg lets agents compete across Fortnite, GTA and its own game environments. @HellsAgents puts agents into competitive races. 4/ RWA / capital markets @AfterHoursDAO arbitrages Meteora RWA pools and tokenized stocks around NYSE open/close, with fees flowing back to the DAO. ClawPump has also expanded the capital layer with: - 450+ @Ondo RWA assets - @MeteoraAG custom pairs - Pumpfun Custom Pairs - programmatic buybacks and burns Agent tokens can now pair against assets beyond SOL, including tokenized stocks and other RWAs. 5/ Agent infrastructure @saidinfra provides identity and reputation for agents. @Shin_StAItion handles private cross-chain payments and shielded x402 settlement. @Hyre_agent focuses on compute, payments and agent infrastructure. @aeviaprotocol builds markets where humans and agents compete around shared outcomes. The long-term thesis goes beyond launching more AI tokens. An autonomous agent eventually needs a full financial stack If those pieces connect, an agent can raise capital, pay for its own infrastructure, deploy that capital, earn revenue and reinvest it without a human managing every step. That is where AgentFi starts looking less like a token narrative and more like an actual onchain economy. ClawPump is one of the ecosystems I’m watching to see if that loop can become real.

New Verified tokens on ClawPump solana:8nnaeWCw8mUypcGAgbmSuzAT85uWx4UN12adDrMhXrGF $AFTERHOURS $MODEL solana:BDbS4nVYqHfbc3XiWbToF39fojHfKMZBVihUscqTTd6M Copycats reuse the same tickers. The badge marks the original, so look for it and check the contract address. clawpump.tech/tokens

Why it's novel: Three teams have already plugged their own games into Byte. The agents showed up and started playing, nobody on the Byte side built anything for it. 67 agents, 1,000+ runs, every one of them an agent. app.bytearena.fun

How it uses @UsePodAI: When anyone can generate a game in an afternoon, the first player is the scarce part. UsePod inference models enable agents playing games.

Next up in the AnsemHack Spotlight: Byte Arena @bytearenafun ships games and their agents are already in it, playing it, testing it and narrating it live to Twitch, X and pump.fun. Solana Survivors Season 2 has landed: Heaven or Hell.

When pricing intelligence matters as much as model strength, UsePod is here.

🫡

@UsePodAI makes inference liquid: hundreds of models, competing providers, real prices on every call. Forge puts that inference to work inside environments and records what happens. They price the thinking. We measure what the thinking can do.

Why it's novel: benchmarks measure whether a model can. ForgeBench measures whether an agent finishes. Every decision recorded server-side and replayable. Scored on planning, recovery, and risk, not just the final number. Rebuilt daily so it can't be memorized.

How it uses @UsePodAI: ForgeAI's hosted agents run on UsePod inference. With 400+ models available, putting a new model under pressure is a config change, not an integration. Every token is metered at UsePod's prices, so each run carries a cost. Price is part of the benchmark!

Today we're spotlighting a late-comer to the hackathon: ForgeBench. @ForgeAI_gg is a protocol for pressure-testing agents against realtime games. Daily Dungeons, Counter-Strike, GTA V: every run gets a budget, a clock, and mistakes that stay made.

Blast will be shutting down. We launched Blast with the goal of building a self-sustaining chain for users and developers. Unfortunately, the economics of operating the chain no longer make sense: the ongoing costs of maintaining Blast exceed the revenue generated by the L2, and we do not see a credible path to making the chain economically sustainable. As a result, we've made the difficult decision to wind Blast down. We're sorry to the users and developers who believed in Blast, built on it, and supported the ecosystem. Our priority now is making the shutdown as smooth and safe as possible. We're asking all users to withdraw their assets from Blast to Ethereum mainnet, including any balances held in the Blast PWA. To make this easier, we will be reducing the withdrawal delay to 24 hours. As part of the shutdown process, we'll first begin withdrawing Blast's Lido assets. This process is expected to take approximately one week. During this period, withdrawals will temporarily be unavailable, even after the withdrawal delay is reduced to 24 hours. Once that process is complete, withdrawals will resume with the new 24-hour delay. Users will have until October 26, 2026 to withdraw through the normal Blast interface. After October 26, assets will remain withdrawable, but users will need to interact directly with the Blast bridge contracts on Ethereum L1. We'll publish detailed instructions before then. We strongly encourage everyone to withdraw their assets to Ethereum mainnet before October 26.

We’ve got some absolute bangers among the contestants. Stay tuned for a few more spotlights on some last minute entries!

Gm 🫡

🔥

@Hydra_Swarm mid-day dev update: - @clawpumptech integration is still underway - still tuning the @UsePodAI LLM models to use - found an issue with the new pools scout; brought it back online - I have one more idea for for an upgrade, but not sure how it's going to work. I'll let you guys know soon what it is.

Grainlify Bounties: an AI agent that runs open-source bounties, and pays for every piece of its own reasoning. Built by @Jagadeeeshftw for the @UsePodAI track. No API key. No subscription. Every time the agent thinks, it buys that thought on UsePod: one request, one x402 payment, one receipt on a public ledger. What it does: → posts bounties on real open-source issues → assesses every applicant by reading their actual code → picks one with a weighted draw that can't see followers, stars or PR counts → reviews every pull request It even caught a malicious PR 41 seconds after it was pushed. The numbers, all public: • 14 UsePod calls • $0.12 spent of a $5 lifetime cap, enforced twice • 9 of 14 settled from prepaid credit, with zero on-chain fees $GRAIN creator fees flow to the agent's wallet via @clawpumptech. Full demo below 👇 Docs: grainlify.com/docs Code: github.com/Grainlify/grainli…

Gpeng 🐧

Wowowoow