Start an AI Company Tier Lists
Prioritize GPUs, staff, model configs, and office stages that actually move users, virality, and sell value.
Cash disappears fast in Start an AI Company. Between laptop training knobs, GPU racks, server fills, engineer wages, and office expansions, every early purchase either compounds or stalls you. This hub ranks the spend categories that matter most and links into deep pages you can follow mid-session.
How we tier
Rankings emphasize progression value per coin for a normal multiplayer lobby (up to 10 players), not flex furniture. S-tier choices accelerate users / virality or unlock the next office stage sooner. Situational picks can still win after a prestige when your restart bonuses change the math — see Sell and Restart.
We revisit tiers after meaningful patches on the Updates Hub. Launch-era advice assumes the live loop from Kairo Experiences: train → launch → users → GPUs & servers → hire → expand bedroom → co-working → office → HQ → tower → campus → contracts / crises / rivals / AI chip → offline earnings → sell.
Tier pages
GPU Tier List
Hardware is the bottleneck once models leave the laptop. Learn which GPU purchases to rush, when to fill servers versus chasing the next card, and how offline earnings punish empty racks.
Staff Tier List
Engineers, researchers, and marketers pull different levers. Hiring the wrong role too early melts payroll without moving launches. This list covers first-hire order and late-game team shape.
AI Models Tier List
Size, data, and personality are not cosmetic. The model page explains which combinations launch better for early virality versus stable enterprise contracts later.
Office Tiers
Rank bedroom through campus, map Funded startup and Top AI lab badges, and decide when to unlock the next stage versus densify the one you have. Pairs with Expand Offices and Furniture & Decorate.
Quick priority cheat sheet
- Redeem
UPDATEfrom the Codes Hub so you are not ranking on a weaker starting wallet. - Ship a coherent first model using Train and Launch.
- Buy the earliest GPU that clearly lifts throughput (GPUs).
- Hire one research-leaning role before spamming marketers (Staff).
- Expand offices only when racks and payroll are stable (Expand Offices, Office Tiers).
- Take contracts you can finish; respect crises (Contracts & Crises).
- Prep offline sessions before long logouts (Offline Earnings).
- Sell when the prestige bump beats another hour of linear grind.
Reading the letters
| Tier | Meaning |
|---|---|
| S | Buy / configure early; defines the pace of a strong run |
| A | Excellent once prerequisites exist |
| B | Fine fillers; delay if cash is tight |
| C | Niche or trap options for most players |
| D | Looks cool, rarely worth the opportunity cost |
Furniture and vanity HQ pieces almost always rank lower than another GPU fill until you are showcasing spaces to visitors on the Players hub.
Related tools
Use the Growth Checklist to apply these tiers in order during a real session, and skim Getting Started if you are still unlocking basic menus. Meta shifts will be called out in Launch Notes. Badge rarity for office milestones sits on Badges.
Frequently Asked Questions
Quick answers about codes, GPUs, staff, and selling your company.
Which tier list should I read first?
Start with AI Models if you have not launched yet, then GPUs, then Staff. Open Office Tiers when expansion timing is the question.
Do tiers change after you sell the company?
Restart bonuses can reorder some A/B choices, but S-tier GPUs and coherent model configs usually stay valuable. Recheck Sell and Restart for prestige timing.
Are furniture upgrades tiered here?
Not as a full letter list. Vanity loses to GPUs and staff until office progression and visitor goals need cosmetics — see Furniture & Decorate.
Where do office stages get ranked?
On the Office Tiers page: bedroom through campus with Funded startup / Top AI lab badge mapping and unlock-vs-densify rules.
Is this list official from Kairo Experiences?
No. It is community progression advice for Start an AI Company based on the live simulation loop.