AI stocks are still rising, but the real opportunity may have already started shifting to a new track
Recently, AI-related assets have continued to draw market attention. But here’s a question worth thinking about:
If investment in AI infrastructure keeps growing, besides GPUs and leading AI companies, what other links can continue to benefit from this round of industry dividends?
I’m more focused on the following five areas👇
⚡ 1. Power
One of the biggest basic necessities for AI data centers is electricity.
The larger the model and the stronger the computing power, the higher the data center’s power consumption.
In the future, AI expansion won’t just be a story about chips—it will also become a story about power grids, transformers, distribution equipment, and generation capacity.
❄️ 2. Liquid cooling / thermal management
AI servers are seeing higher and higher power density, and traditional air cooling faces challenges.
As the number of GPUs increases, cooling isn’t optional—it becomes part of the infrastructure.
Who can solve the problem of “computing power is getting stronger, but the servers can’t overheat,” and who can capture the benefits of industrial upgrading.
🌐 3. High-speed networking
Many people only focus on GPUs, but they can easily overlook an issue:
If tens of thousands—even hundreds of thousands—of GPUs are placed together, how do we enable high-speed communication among them?
The importance of infrastructure like high-speed switches, network chips, optical modules, and more will keep increasing.
🏭 4. Semiconductor equipment
As demand for AI chips grows, at its core that means fabs need to expand capacity.
So besides chip companies, you can also look at the equipment and processes needed to manufacture chips.
🏢 5. Data centers
In the end, all AI must land in the real physical world.
Servers need server rooms, power, cooling, networks, and land.
So data centers themselves may also become very important infrastructure assets in the AI era.
More and more, I feel that:
The logic behind AI investing is shifting from “who can build the strongest model” to “who can provide the infrastructure for AI.”
It’s like the gold rush era.
What’s really worth researching isn’t necessarily only the people who dig for gold.
Those who sell shovels, provide water, supply electricity, or repair railways may benefit just as much.
So going forward, if I continue to study AI, I won’t only focus on GPUs.
I will pay special attention to this chain of logic:
compute → network → power → cooling → data centers → semiconductor equipment
This industry chain may be more worth tracking long-term than simply chasing a single AI hot spot.
#AI #ai股持续上涨还有哪些投资机会
Recently, AI-related assets have continued to draw market attention. But here’s a question worth thinking about:
If investment in AI infrastructure keeps growing, besides GPUs and leading AI companies, what other links can continue to benefit from this round of industry dividends?
I’m more focused on the following five areas👇
⚡ 1. Power
One of the biggest basic necessities for AI data centers is electricity.
The larger the model and the stronger the computing power, the higher the data center’s power consumption.
In the future, AI expansion won’t just be a story about chips—it will also become a story about power grids, transformers, distribution equipment, and generation capacity.
❄️ 2. Liquid cooling / thermal management
AI servers are seeing higher and higher power density, and traditional air cooling faces challenges.
As the number of GPUs increases, cooling isn’t optional—it becomes part of the infrastructure.
Who can solve the problem of “computing power is getting stronger, but the servers can’t overheat,” and who can capture the benefits of industrial upgrading.
🌐 3. High-speed networking
Many people only focus on GPUs, but they can easily overlook an issue:
If tens of thousands—even hundreds of thousands—of GPUs are placed together, how do we enable high-speed communication among them?
The importance of infrastructure like high-speed switches, network chips, optical modules, and more will keep increasing.
🏭 4. Semiconductor equipment
As demand for AI chips grows, at its core that means fabs need to expand capacity.
So besides chip companies, you can also look at the equipment and processes needed to manufacture chips.
🏢 5. Data centers
In the end, all AI must land in the real physical world.
Servers need server rooms, power, cooling, networks, and land.
So data centers themselves may also become very important infrastructure assets in the AI era.
More and more, I feel that:
The logic behind AI investing is shifting from “who can build the strongest model” to “who can provide the infrastructure for AI.”
It’s like the gold rush era.
What’s really worth researching isn’t necessarily only the people who dig for gold.
Those who sell shovels, provide water, supply electricity, or repair railways may benefit just as much.
So going forward, if I continue to study AI, I won’t only focus on GPUs.
I will pay special attention to this chain of logic:
compute → network → power → cooling → data centers → semiconductor equipment
This industry chain may be more worth tracking long-term than simply chasing a single AI hot spot.
#AI #ai股持续上涨还有哪些投资机会