#亚马逊投资OpenAI
To be honest, seeing Amazon confirm in its earnings report that it will pour the full 50 billion into OpenAI, my first reaction was: what a smooth move—playing the shell game with left hand, right hand $AMZN $AMZNB
If I really have to choose between betting and a bubble, I’d lean toward viewing this as a bundled cloud sale dressed up as a strategic bet
The key variable I care about most is whether the $100 billion cloud order OpenAI has promised can actually translate into real cash flow on the ground
First, look at the $100 billion cloud orders. Eight years to reach $100 billion sounds shocking, and it directly locks AWS into huge long-term revenue. But no matter how big the compute demand is, it still depends on OpenAI’s own ability to commercialize and monetize its applications
Moreover, OpenAI is smart—it keeps tugging between Microsoft, Oracle, and Amazon. In essence, it’s exploiting the big players’ anxiety over falling behind to arbitrage compute
Next, consider preferred shares and liquidity risk. On the surface, liquidity is locked up—you have to wait for a listing to convert into common stock. But I think the market is overestimating this risk. Most of the money Amazon invests will flow back to its books through OpenAI’s purchase of AWS services and its Trainium in-house chips. The capital gains are just icing on the cake. The real benefit is locking in AWS market share and real deployment scenarios for the in-house chips—those are the tangible takeaways
As for where things may go next, I have two predictions:
One is that this financing loop—where giants pay and unicorns buy cloud services—will soon face strict scrutiny from capital markets. Going forward, everyone will pay more attention to the true net profit after stripping out related-party transactions
The other is that the shift of compute chips away from Nvidia will accelerate. OpenAI has started consuming a large amount of Amazon’s Trainium chips, which shows that model companies are desperately hunting for alternatives to cut costs. This will also force the cloud giant’s in-house chips to quickly go mainstream
Not investment advice—DYOR
To be honest, seeing Amazon confirm in its earnings report that it will pour the full 50 billion into OpenAI, my first reaction was: what a smooth move—playing the shell game with left hand, right hand $AMZN $AMZNB
If I really have to choose between betting and a bubble, I’d lean toward viewing this as a bundled cloud sale dressed up as a strategic bet
The key variable I care about most is whether the $100 billion cloud order OpenAI has promised can actually translate into real cash flow on the ground
First, look at the $100 billion cloud orders. Eight years to reach $100 billion sounds shocking, and it directly locks AWS into huge long-term revenue. But no matter how big the compute demand is, it still depends on OpenAI’s own ability to commercialize and monetize its applications
Moreover, OpenAI is smart—it keeps tugging between Microsoft, Oracle, and Amazon. In essence, it’s exploiting the big players’ anxiety over falling behind to arbitrage compute
Next, consider preferred shares and liquidity risk. On the surface, liquidity is locked up—you have to wait for a listing to convert into common stock. But I think the market is overestimating this risk. Most of the money Amazon invests will flow back to its books through OpenAI’s purchase of AWS services and its Trainium in-house chips. The capital gains are just icing on the cake. The real benefit is locking in AWS market share and real deployment scenarios for the in-house chips—those are the tangible takeaways
As for where things may go next, I have two predictions:
One is that this financing loop—where giants pay and unicorns buy cloud services—will soon face strict scrutiny from capital markets. Going forward, everyone will pay more attention to the true net profit after stripping out related-party transactions
The other is that the shift of compute chips away from Nvidia will accelerate. OpenAI has started consuming a large amount of Amazon’s Trainium chips, which shows that model companies are desperately hunting for alternatives to cut costs. This will also force the cloud giant’s in-house chips to quickly go mainstream
Not investment advice—DYOR