#AIStocksWhatNext
$NVDA $AMD $AVGO $MSFT $AMZN $META
I keep watching the AI trade, and lately I’m less interested in the loud headlines.
Everyone talks about models, GPUs and who is “winning” AI.
I’m watching what sits underneath it.
$NVDA just showed how massive the demand has become, with $96.2B in quarterly revenue and roughly $89B from Data Center.
But the interesting part to me is what comes next.
Training gets attention.
Inference could become the recurring business.
Then there’s $AVGO. Its AI semiconductor revenue reached $16.7B in Q3, up 221% YoY. That tells me the story is getting bigger than GPUs. Hyperscalers are increasingly building custom chips and controlling more of their own infrastructure.
Microsoft and Amazon are spending enormous amounts on data centers and compute too.
And this is where I start asking different questions.
Who is actually using all this capacity?
How efficiently?
Who is paying for it?
And when does the spending turn into durable cash flow?
Even power is becoming part of the story. Data-center projects are competing for grid capacity in places like Texas.
So I’m watching the boring stuff now:
Power. Cooling. Networking. Memory. Utilization.
The AI race may look like a software story from the outside.
Underneath, it increasingly looks like an infrastructure race.
That’s the part I’m watching.
$NVDA $AMD $AVGO $MSFT $AMZN $META
I keep watching the AI trade, and lately I’m less interested in the loud headlines.
Everyone talks about models, GPUs and who is “winning” AI.
I’m watching what sits underneath it.
$NVDA just showed how massive the demand has become, with $96.2B in quarterly revenue and roughly $89B from Data Center.
But the interesting part to me is what comes next.
Training gets attention.
Inference could become the recurring business.
Then there’s $AVGO. Its AI semiconductor revenue reached $16.7B in Q3, up 221% YoY. That tells me the story is getting bigger than GPUs. Hyperscalers are increasingly building custom chips and controlling more of their own infrastructure.
Microsoft and Amazon are spending enormous amounts on data centers and compute too.
And this is where I start asking different questions.
Who is actually using all this capacity?
How efficiently?
Who is paying for it?
And when does the spending turn into durable cash flow?
Even power is becoming part of the story. Data-center projects are competing for grid capacity in places like Texas.
So I’m watching the boring stuff now:
Power. Cooling. Networking. Memory. Utilization.
The AI race may look like a software story from the outside.
Underneath, it increasingly looks like an infrastructure race.
That’s the part I’m watching.
