On this episode of The Weekly Wrap, Steve Eisman covers a massive earnings week headlined by Microsoft posting 30% EPS growth while Meta saw EPS fall 13%. Steve also covers results from Visa, Quanta, Bloom Energy. Starbucks may have turned while Apple disappointments. It’s been a brutal week for FICO. He also explains why he will sell Charter next week and he takes a mailbag question about the history of the K-shaped economy.
The AI Debate Has Changed
Last year, just about everyone was positive on AI.
Every announcement of higher AI capital spending was greeted with massive stock price increases. Companies couldn’t spend enough money. Investors wanted more. The assumption was simple: more capex meant more growth.
Today, the questions are much harder. AI is incredibly capital intensive. AI may not have meaningful moats. Chinese companies have developed very good models that are much cheaper than the American closed-source models. That creates the possibility of a price war, and suddenly investors have gone from believing AI can do no wrong to wondering whether the economics work at all.
The truth is somewhere in the middle.
Not Every AI Company Is the Same
One of the most important distinctions investors can make today is separating the companies building AI infrastructure from those developing AI models.
The hyperscalers—Microsoft, Amazon, Google, and Oracle—are investing extraordinary amounts of capital to build the data centers that power AI applications. Those investments are expensive, but they also create significant barriers to entry. Very few companies have the financial resources to compete at that scale.
The economics of large language models may prove to be different.
As open-source models continue to improve, enterprises have more choices than they did even a year ago. That increased competition could make it more challenging for model providers to maintain pricing power over time.
The technology is advancing rapidly, and it’s still too early to know exactly how this competitive landscape will develop. But it’s an area investors should continue to watch closely.
The AI race is becoming extraordinarily expensive.
Look at Meta. Revenue continues to grow, but expenses are growing even faster. Research and development spending has exploded. Free cash flow has been squeezed. The company continues to increase its capital spending plans because it feels like it has no choice.
Microsoft told a different story.
Azure continues to accelerate, helping justify the billions the company is investing in AI infrastructure. Amazon delivered a similar message through AWS.
These are all spending enormous amounts of money.
The difference is that some companies are already generating returns from those investments while others are still trying to prove that they eventually will. Investors are beginning to recognize that distinction.
Comments
Nothing yet. Say the first thing.
Sign in to join the conversation.