Futures
Access hundreds of perpetual contracts
CFD
Gold
One platform for global traditional assets
Options
Hot
Trade European-style vanilla options
Unified Account
Maximize your capital efficiency
Demo Trading
Introduction to Futures Trading
Learn the basics of futures trading
Futures Events
Join events to earn rewards
Demo Trading
Use virtual funds to practice risk-free trading
CFD
Stock CFD Derivatives
US Stocks
Access real US stocks and ETFs
HK Stocks
Trade quality Hong Kong-listed stocks
Korean Stocks
SK Hynix
Real Korean stocks and top assets
Stock Futures
High leverage, 24/7 trading
Tokenized Stocks
Backed by real stock assets
IPO Access
Unlock full access to global stock IPOs
GUSD
3.8%
Mint GUSD for Treasury RWA yields
Stocks Activities
Trade Popular Stocks and Unlock Generous Airdrops
Launch
CandyDrop
Collect candies to earn airdrops
Launchpool
Quick staking, earn potential new tokens
HODLer Airdrop
Hold GT and get massive airdrops for free
Pre-IPOs
Unlock full access to global stock IPOs
Alpha Points
Trade on-chain assets and earn airdrops
Futures Points
Earn futures points and claim airdrop rewards
Promotions
AI
Gate AI
Your all-in-one conversational AI partner
Gate AI Bot
Use Gate AI directly in your social App
GateClaw
Gate Blue Lobster, ready to go
Gate for AI Agent
AI infrastructure, Gate MCP, Skills, and CLI
Gate Skills Hub
10K+ Skills
From office tasks to trading, the all-in-one skill hub makes AI even more useful.
The greatest unsolved mystery: how do companies that sell AI models actually make money?
I read an insightful article from @michael_lwy that found the answer from the Hong Kong MTR.
MTR (MTR Corporation) is one of the very few subway systems in the world that doesn’t rely on government subsidies and can sustain itself. It’s even a publicly listed company, distributing dividends every year. However, when it comes to making money, it’s not from ticket prices.
Public data shows that at its best time, the railway operating profit was HK$2 billion, but in the same period the three years of capital expenditures were HK$87.9B. That means ticket revenue only covers 8% of construction costs.
So relying on selling tickets is never enough to build the next railway line—this was never designed as MTR’s profit model. Then where does the money come from? The answer is from the buildings above the stations. You might not know that MTR owns 13 shopping malls and 47 station-overlaid property projects. The railway makes surrounding land appreciate in value, and MTR develops those assets itself; the property profits then feed back into railway construction.
Now replace “railway” with “AI model,” and the structure is actually exactly the same. API pricing drops by 10x each year, open-source models chase down closed-source ones, and any AI lab with pricing above marginal cost will have its customers poached.
Each model may generate slim profits at the operational level, but profits can never support the next round of training.
This is the same logic as subway ticket fares: it can cover operating costs, but it can’t cover construction costs.
So asking “when can an AI company make money from APIs,” is the same question as “when can a subway recoup construction costs from ticket prices.” The answer is the same too: you can’t, and you shouldn’t; you have to build your own buildings.
For example, Google puts Gemini into Gmail, Docs, and Search, giving the model away for free—what it earns is probably the money from you being even more dependent on Google’s whole suite of products.
Conversely, companies that only sell bare model APIs and don’t have their own “buildings” are just like subways that only sell tickets—they’ll have to survive on subsidies sooner or later.