12 Months of Trendvesting: The Real Unit Economics of a Production AI Side Project

A developer's 12-month cost audit of Trendvesting, an AI signal-intelligence platform for equities and options, reveals that production AI side projects cost far more than API pricing pages suggest, with tokens never being the expensive part. The project, spanning 1,672 commits since February 14, 2024, across a multi-mode Go backend, Next.js app, React Native client, and Python FastAPI consensus service, shows that infrastructure costs dominate before a crowd can be afforded.

Member-only story 12 Months of Trendvesting: The Real Unit Economics of a Production AI Side Project Everyone shipping AI side projects publishes the demo. Almost nobody publishes the ledger. Here’s mine. This is the complete cost history of Trendvesting, an AI signal-intelligence platform for equities and options that has been running in production since a first commit dated 2024-02-14 and now spans 1,672 commits across a multi-mode Go backend, a Next.js app, a React Native client, and a Python FastAPI consensus service. The original bet was automated signal intelligence: that a swarm of bots plus a community of humans could crowd-source the detection of market moves faster than any one trader watching charts. The vision was the collective layer. The reality, twelve months in, is that before you can afford a crowd, you have to afford a cluster — and this post is about that second part. The headline finding: a “small” AI project in production does not cost what the API pricing page implies — and the tokens were never the expensive part. What follows is where the money actually went, the three taxes I paid to learn that, and the real cost per signal once you stop lying to yourself about it.