Skip to content
Zayne Lovecraft
All posts

AI · · 4 min read · Updated

Why I Canceled All My AI Subscriptions for Google AI

I was paying $200 a month for Cursor plus $600 to $1,000 in overages and ChatGPT on top. Here is why I moved my entire coding workflow to Google.

Zayne Lovecraft

Software Engineer III

Google Antigravity official logo and agent platform banner

I mainly use AI for coding. Up until recently, that meant paying $200 a month for Cursor, plus blowing past my usage limit with $600 to $1,000 in overage charges on heavy months. On top of that, I was still paying for ChatGPT and had previously paid for GitHub Copilot.

Last month I canceled all of them and moved everything over to a single Google AI subscription. Here is why I made the switch and why I do not see a need for anything else.

The Cursor overage trap

When you write code all day as a software engineer, a base AI plan rarely covers your actual volume. I was on the $200/month Cursor tier, and I would still go over my limit. Some months the overages hit $600. On my heaviest months, they got close to $1,000.

Add a separate ChatGPT subscription on the side for general questions, and I was paying a massive bill across multiple tools just to get my daily work done.

Closeup of a MacBook Pro keyboard and code editor on a dark desk Photo by Rodrigo Santos on Pexels

One subscription that replaces the whole stack

Moving to Google put everything I actually use under one subscription:

What I Was Paying ForWhat Replaced It Under Google
Cursor ($200/mo + $600 to $1,000 overages)Google Antigravity (integrated IDE and agent harness)
ChatGPT Plus (daily chat and research)Gemini (Gemini 4 Argon, Hydrogen, and 3.8 Flash)
GitHub Copilot (autocomplete)Antigravity IDE inline completions and supercomplete
Disconnected apps and storageGoogle Drive, Docs, Sheets, Chat, and Gmail with AI built in

1. Google Antigravity and the integrated IDE are ahead of the game

Because I am a software engineer, my primary tool is the editor. Google's subscription includes Google Antigravity and its integrated IDE.

Google Antigravity IDE overview banner

After using it daily, the UI and interface are way ahead of the game. It feels cleaner and faster than Cursor and Grok, and the agent harness combined with the integrated IDE simply gets the job done:

  • The agent harness actually finishes work: It reads my local filesystem-first OS rules, runs pnpm lint and pnpm build in a terminal, connects to MCP servers, and checks its own work in a built-in browser.
  • Better model choices right in the picker: I use Gemini 3.8 Flash for fast research and quick lookups, and switch to Gemini 4 Argon or Hydrogen when I am doing deep architecture work across a Next.js 16 and Supabase codebase.
  • Predictable usage and credits: Instead of surprise $600 to $1,000 editor overages, I can check my exact quota windows and manage AI credits directly from the settings panel or CLI.

2. Gemini replaces ChatGPT and upgrades the Google tools I already love

Outside of writing code, I already live in Google's tools every day: Google Drive, Google Sheets, Google Docs, Google Chat, and Gmail.

Having one personal subscription means Gemini replaces ChatGPT for all of my general questions and research, while also putting AI directly inside my emails, docs, sheets, and 2 TB of Drive storage. Everything is connected under the same account instead of spread across different logins.

Close-up of server racks in a data center highlighting modern technology infrastructure Photo by panumas nikhomkhai on Pexels

Why I trust Google when building for the web

Google is the number one website in the world. They practically drive the internet.

When you spend your career building web apps, caring about 100 Lighthouse scores, technical SEO, and Generative Engine Optimization, you realize how much of the web runs on Google's standards and systems. Working across the web feels much more reliable when your AI and agent harness come straight from the company that indexes the internet.

Detailed view of fiber optic patch cables connecting to a blue patch panel in a data center Photo by Brett Sayles on Pexels

They are also the tip of the spear in AI infrastructure:

  • They invented the Transformer architecture that modern AI is built on.
  • They train frontier models at DeepMind and run them on their own custom TPU hardware instead of renting servers from someone else.
  • Their models pull live context straight from Google Search, which matters every time a framework ships a breaking update.

Takeaway

Going from a $200/month Cursor plan with $600 to $1,000 in overages plus ChatGPT down to a single Google subscription was the easiest call I have made all year. Between Google Antigravity's IDE and agent harness, Gemini, and AI built straight into Gmail, Drive, Docs, Sheets, and Chat, I have everything I need in one place.

  • AI Subscriptions
  • Google AI
  • Developer Tools
  • Antigravity
  • Gemini

Keep reading