MCG Live · July 11, 2026 · 58 min
P0 Systems is an AI inference gateway that routes users to multiple open-source models and offers cheaper access to AI. Founder Corey describes how they pivoted from an AI agent launchpad three months ago, now processing over 1 billion tokens per hour and hitting $60k monthly recurring revenue while ranking #1 on OpenRouter—the surprising part being that their Chinese connections, cheap GPU sourcing from post-mining-ban equipment, and subsidized energy costs give them structural cost advantages that could 10x revenue once they self-host infrastructure.
P0 Systems is currently processing 1.9 billion tokens per hour with over 30,000 CLI downloads and doing 60k MRR, ranking number one on OpenRouter.
P0 raised 319k in non-dilutive grant funding and is based in Thailand and China, sourcing cheap GPUs from post-mining-ban equipment and benefiting from subsidized energy costs in China.
The founder and team launched P0 in December 2025 as an AI agent launchpad, then pivoted three months ago to focus on the inference market after seeing opportunity in companies like OpenRouter and GitLab.
P0 has generated 21k in verified revenue from paid models on OpenRouter, with growth accelerating from 3-5k MRR previously to 60k MRR recently due to new targeted ads spending 2-3k on Twitter.
Chinese AI labs provide free model inference to P0 to boost their own user metrics and revenue, creating a mutually beneficial top-of-funnel strategy where P0 then upsells paying customers on premium tiers.
P0 accepts payments via local methods including Thai QR codes, Alipay, and WeChat Pay alongside crypto, targeting Southeast Asian users who cannot afford traditional AI service subscriptions.
The founder plans to allocate 40-60% of monthly profits to token buybacks and liquidity, with future goals to offer token-gated equity ownership to align holders with company success.
P0 plans to build its own GPU infrastructure and potentially train proprietary models, which the founder estimates could 10x revenue by eliminating reliance on provider infrastructure.
The founder is currently seeking Series A funding with a pitch deck prepared, targeting 500 billion tokens processed by end of July to strengthen fundraising metrics and valuation.
When evaluating inference projects, investors should look at time spent shipping through market downturns, verifiable token volume on OpenRouter and CLI downloads on NPM, and actual user growth rather than marketing claims.
This summary was written from the recording's transcript by a language model. For the exact words, ask the MCG Live archive: every answer there is quoted word for word and linked to the second it was said. Market Bubble Search is independent and not affiliated with the show. Nothing here is advice.