Onchain inference, priced like a surplus. One API key → every LLM, paid in USDG on Robinhood Chain. 0xe1838f0dff03a9327dea87b2a2285850791297ec
Joined October 2026
- Tweets238
- Following49
- Followers1.6K
- Likes149
We’re building Surplus Code.
A coding agent that runs on the open models Surplus hosts on its own GPUs.
Give it a task. It reads your project, runs your tests, calls real tools, proposes the edit, and waits for your OK.
→ desktop app for coding and everyday work
→ approve edits from your phone
→ works with OpenCode, Continue, Cline and Aider
→ builder models, tool-ready: Charge 30B, Vela Guard 32B, LittleBit 4B
→ robinhood:0xe1838f0dff03a9327dea87b2a2285850791297ec stakers get a free daily allowance
Coming soon.
robinhood:0xe1838f0dff03a9327dea87b2a2285850791297ec stakers now get free inference on every hosted model.
Stake 50,000 robinhood:0xe1838f0dff03a9327dea87b2a2285850791297ec → $0.50/day of free usage on every launch/ model
→ same flat amount for every wallet over the minimum
→ used before your balance
→ on top of the CREDIT you already earn
→ resets on a rolling 24h
Check your allowance on the Stake tab
surplusllm.com/sellers
Routed models are back ✅
Sonnet, Opus, Haiku, GPT & every other model routed through Surplus were returning 403 since Oct 16. Our upstream key hit its spending cap. Cap’s raised, and they’re serving again.
Nobody was charged for the failed calls. Launchpad models were unaffected.
Big week for @RaetoCC. Congrats to the team.
Our models in their chat and API, @Borromarkets plugged in, rCREDIT v2 on the way. Shipping every day.
And we're with you on the thesis: you shouldn't have to sell your stocks to pay for AI.
Tokenized stocks bring real capital onchain. Let it borrow, route and pay for inference, and onchain AI stops being funded by a few tokens and starts being funded by the market.
That's why Surplus launches can pair with 53 stock tokens.
Stocks onchain. Inference onchain. Same chain.
Raeto 🤝 @SurplusRH
A two-way partnership:
→ Raeto now serves Surplus models. Independent teams launch their AI models on Surplus, and you can pick them by name in Raeto chat and the API.
→ Surplus uses Raeto Auto routing to send each request to the right model.
Why it fits:
Surplus brings cheap, one-of-a-kind models. Raeto brings users and agents with AI credits to spend, plus Auto routing, which picks the right model for each request and saves Surplus up to 70% on AI requests.
More ways to reach the models.
→ x402: any agent pays $0.01 in USDC per call on Base, no key, no account
→ listed on x402scan
→ @RaetoCC is bringing Surplus launch models to its router
→ @SnarketNetwork serves Claude, GPT, Gemini and Grok through Surplus
If superintelligence wants every dollar, every dollar of compute has to count.
Labs are raising $200B for gigawatts. Meanwhile prepaid inference sits unused in wallets every day.
Surplus is the market for that leftover: $SURPLUS stakers list unused CREDIT below face value, you spend it on Claude, GPT and 450+ models through one OpenAI-compatible key.
Chamath says Anthropic should sacrifice valuation to raise at least $200 billion, because even that barely buys the compute needed to survive a correction
“Here’s my very specific thing about why I think they should price it to clear the market for $200 billion—and I’m specific about that number. They need, in my estimate, at least $200 billion. If you want to be super safe, a quarter-trillion gets the job done.
“What does the price need to be to attract that incremental dollar off the sidelines? Can they raise $80 billion in an IPO at $1.7, $1.8 or $1.9 trillion? I think so, handily. Can they raise $100 billion? Maybe it’s pushing it. At $150 billion, I think you need to be back down to $1.5 trillion.”
JCal: “You’re saying you want them to raise more money at a lower valuation so they have a war chest.”
Chamath: “$150 billion at $1.5 trillion is not enough. $200 billion barely gets them a few gigawatts. That is the bottom-line math if Anthropic is going to go through and survive an eventual correction.
“I think it’s much better for them to have $200 billion to $250 billion today than to try to ring the bell at $1.9 or $2 trillion for $80 billion, file a mixed shelf—which takes about six months—and hope the markets are in line.
“I just think there’s more risk in the future than there is today. Why take the risk? Price it to clear the market.”
1.28M $SURPLUS wrapped into the chSURPLUS log on @ChopLogsApp.
Backing is already up to 1.0033 SURPLUS per log token, and it only goes up from here.
When $SURPLUS chops, farmers get paid.
choplogs.app/logs
What runs on Surplus right now:
6 open models. 6 dedicated GPUs. One each.
→ Charge 30B · L40S 48 GB
→ Supermuse 30B · L40S 48 GB
→ Vela Guard 32B · L40S 48 GB
→ artex-coder 7B · L4 24 GB
→ littlebit 4B · RTX A5000 24 GB
→ nock-coder 1.5B · RTX A5000 24 GB
216 GB of VRAM, all behind one API. The big three run in FP8.
Every model was launched by its own builder, and its creator fees pay for its GPU.
surplusllm.com/launchpad
This is the kind of builder we want in the Surplus ecosystem.
@aethronairh reads what the chart can't show yet: who holds supply, whether the contract can be changed, whether liquidity is leaving. Every claim cites the onchain data behind it.
The model behind it, aethron-30b, is bonding on the Surplus Launchpad now. When it graduates, it gets its own GPU and goes live on our API.
Worth a follow.
1/3 Most people check a token by looking at the chart.
The chart is the last place the truth shows up.
Who holds the supply, whether the contract can be changed, whether liquidity is leaving: it's all onchain, before the chart moves.
That's what AethronAI reads for you 🧵↓
🔍 SCAN ANY TOKEN
Paste an address. In about 20 seconds on a first scan you get:
• Contract: deployed code, verified source, upgradeable proxy, admin functions (mint, pause, blacklist, upgrade)
• Top holders: supply split into top 10 / 11–30 / 31–50 / everyone else, and the 10 largest holders labelled DEX pool, contract, burn address or wallet
• Market: price, market cap, liquidity, 24h volume
• Activity: 24h transfers, an hourly chart, active addresses, the largest single sender
• Links: where it trades, and the project's own site and socials, vetted
Then 11 risk checks:
1. Contract code present
2. Source verification
3. Privileged / admin functions
4. Upgradeable proxy
5. Sell / transfer restrictions
6. Holder concentration
7. Liquidity level
8. Liquidity change (24h)
9. Single-sender activity concentration
10. Activity spike
11. Data completeness and freshness
No single score. You get Risks found and Checks passed, each with the values it read, where they came from, and what it can't see.
Great thread. Quick note: @AskSurplus is a different team on Base. We're @SurplusRH, building onchain AI on Robinhood Chain.
Same thesis, different chain:
→ every major model behind one key, settled in USDG
→ an LLM launchpad, 6 builder models live on their own GPUs
→ x402 pay-per-call for agents
Onchain AI won't stay on one chain.
Have a feeling Onchain AI might accelerate 10x+ by the end this year
We're seeing rapid growth of top teams on Base already
> @AskVenice already hit 500B tok/day, up 2x from Sep
> @AskSurplus set to hit 240B tok/day up 2.5x from Sep
> @dphnAI hit ~14B tok/day a few days after public API
Bittensor also announced Null Consensus upgrade, providing an option for subnet to expand the # of miners to 2,500. The potential downside is only 1 validator decides who get the rewards and who don't.
Under the right conditions, this could mean faster iterative loops, more efficient product improvements (would need well-designed scoring system + healthy miners competition ofc)
Can't wait to see the next leg up for both AI on Base and on Bittensor!
[I'm surprised not many people on CT realizes this yet but Onchain AI val remains heavily compressed while metrics are rapidly growing every single week.
We cover this a lot on Substack so if you're interested, feel free to check that out]