The Grid’s Beta is LIVE! We can get your AI API costs down by up to 80% by making suppliers compete for your requests. Your first 200M tokens are on us, start building →

v. nice! I’ll be testing this. it feels like the right way to fix out-of-control AI budgets. Congrats on the beta launch 🙌

Excited to have you testing, let us know how it goes!

will do! Snagging those 200M free tokens now. LFG 💥

Just tested, very useful. Hermes agent integration was so fast. and thanks for the 200mn free tokens

Love hearing this! What workflows are you running Hermes on through The Grid?

mainly repo summaries, debugging, quick scripts

Kinda like this tag line... before it was PAYG with The Grid it's PLAYG Pay LESS As You Go

Pay Less As You Go only on The Grid. 👉

Proud to be building with you guys!

The feeling is mutual 🙌

A spot market for inference!

API costs are the silent killer for indie builders. Watching suppliers fight over your request volume is genuinely smart. 200M free tokens is a real reason to test it.

Making model providers fight for traffic is very funny and very useful.👌

80% off API spend changes what a small team can actually afford to ship. That gap between scrappy and scaled just got narrower.

Small teams now have the same firepower without the budget anxiety.

Spending half a day building a router to compare provider pricing only to realize someone already shipped it. The Grid doing the boring infra work so nobody else has to.

not sure the 'suppliers compete' model holds for latency-sensitive apps when i tried similar routing in pennywise the p99 tail times got ugly and my transaction screens started feeling broken. does the grid lock in SLAs per supplier or is it best-effort routing?

This is the inevitable direction for AI infra. Price discovery + supplier competition should’ve happened much sooner. Excited to see where this goes.

Exactly. Cloud computing became a commodity when people stopped caring which server ran their workload. Inference is heading the same way. The winning platform won't be the one with the biggest model—it'll be the one delivering the required quality at the lowest cost and highest reliability.

lowest bidder routing assumes all tokens are fungible. my last project had a 3pm traffic spike where cheap providers failed over to each other like dominoes. saved 60% on the bill, spent it all on the pagerduty.

@PathOfMen_ yeah he get with them to how or wtv

The Grid Beta is now live

first 200M tokens on US 👀

congrats. lools super promising!

Last month I corrected an AI draft that missed three red flags. Fast isn't accurate. The judgment call is the job. What field are you in?

The framing here is sharp. Suppliers competing for requests instead of locking you into one provider. That alone fixes the biggest cost trap in AI dev.

The Grid making AI API suppliers compete to slash costs by 80% is genius

Most cost optimization advice is just "use a cheaper model and pray." Routing each request to whoever bids lowest is the version that actually moves the bill.

The Grid always cooking 🔥👀