As some of the tech titans report their earnings, we’re starting to see a possible pattern relating to the value of AI. Google, Meta, and Microsoft all reported earnings on Wednesday of last week, and of the three, only Google saw a positive Street reaction. Meta fared worst of the three. So, what can we learn from this, and why do we need to learn it?
The “why” is the first point we should address. No public company is going to invest in AI if the result is a dip in their stock price. Apart from the fact that most senior executive compensation is linked in part to stock performance, public companies have a legal duty to shareholders and so doing things that hurt the stock price can result in activist-investor intervention and shareholder suits. All these companies are conspicuous in their commitment to AI, and in all cases management cites AI progress in their public comments on earnings.
When a company can say positive things, backed up by good financial results, there’s an active encouragement for the company to pursue AI investment. Where it can’t do that, we have to assume that it may be forced to pull back a bit, and also that the company’s particular AI strategy may not only be bad for the company, but perhaps questionable overall.
I’ve noted many times that hedge funds create most market trade volume, and that unlike the average investor, they can make money on market declines by “short selling”, borrowing a stock and selling it so that if it drops in price they can buy back the shares at a lower price and then return what they borrowed. This strategy can also be used to drive down stocks, since stocks go up or down based on whether, in the net, they’re being sold or acquired. On Thursday of last week, Google opened up, Microsoft down a bit, and Meta down over 10%. By the end of the day on Friday, after Amazon had a positive report to share and the tech-heavy NASDAQ was up almost 150 points, all three stocks were off a bit, with Meta hit worst and Google the least. This week, short-selling took down the entire market, but obviously targeted AI. There’s a lot of investor uncertainty about AI, obviously, and the three giants who reported earnings last week are at least poster children for that.
All these companies talked about raising their investment in AI, so it’s hard not to see the Street reaction as a measure of what investors think about the potential each of the three has to earn a return on what they’ve promised to spend. Notably, Nvidia hit the $5 trillion market cap level on Thursday, demonstrating that whatever the long-term prospects for ROI might be for the three giant buyers of AI, the seller was going to make out in the near term. So what are the long-term prospects for our big three? Let’s explore.
What’s the problem with Meta, the obvious loser of the group? I think the answer to that is clear, and it’s something I’ve mentioned in past blogs. Consumers think that everything online should be free, and they’re resistant to paying for any sort of new service. For companies like Meta, who like all AI players find future profits from AI are surely tied to someone paying for, that means you have to make a very strong business connection to succeed. They tried that with the metaverse and they seem to have given up, focusing instead on augmented/virtual reality (AR/VR) glasses without pushing a strong business mission for them.
In fact, it may be that the whole AR/VR thing is a diversion. It’s not rocket science to make it possible for glasses to take the place of a computer screen, or to pop up texts or notifications on them in such a way as to make it possible to see the real world around or through these notices. It’s much more challenging to blend these with the real world, and to ensure the augmentations don’t interfere with vision to the point where they risk injury to the wearer. These issues make it hard to see the glasses as a persistently useful tool to consumers, and impossible to see them as an aid in working unless the glasses are very well synchronized with the actual environment. That seems to me to require thinking of AR/VR as being a visualization of a digital twin, and of course you need such a twin to visualize before you start worrying about the “how”. Pushing glasses might advance the second piece too far in advance of the first.
If we focus on AI, that’s particularly true. OK, Meta might be able to use AI as a means of doing better ad targeting, but it’s very hard to see how ad-oriented AI missions aren’t playing into a zero-sum game; adspend isn’t something you can improve just by improving targeting; you steal other share instead. How does AR/VR engage AI in any way, much less a way that generates a willingness to pay?
Microsoft, who saw a bit of a stock decline after their report, has a bit of the same issue. It’s not so much that Microsoft doesn’t have a business-facing AI strategy, or that it doesn’t have business users, but that it doesn’t seem to be generating significant revenue growth. Based on what enterprises tell me, the problem seems to be that the “copilot” concept isn’t seen as a proven way to cut costs, and so doesn’t generate much enthusiasm. While almost 100% of enterprises I chat with say their company uses it, less than a sixth say that they believe it generates any monetizable benefit.
The big problem I see with Microsoft is that it’s really relying on integration with office tools to promote a revenue stream, and that strategy doesn’t seem to resonate with enterprise planners, for the most part. Is Microsoft hoping to capture SMBs, or individuals? There’s nothing wrong with integrating AI with office packages (Google does this too) but it seems better to have specific AI tools available to pull demand through, and also to promote all of the “AI agent” models enterprises find useful.
Then there’s Google (Alphabet, if you want to be strictly correct). This is the AI play that, among the Big Three, earns the most respect. Both Gemini’s Deep Research and NotebookLM tools get the nod by enterprises as being useful, and enterprises also like Google’s initiatives in providing mechanisms to integrate AI agents with applications and support self-hosting. However, they still say that IBM is a better source of insight and support in both these areas.
What seems to make the difference for Google is the value of the AI tools it provides, the fact that these tools are independent and thus not only can be easily fit into multiple missions, but tend to validate AI by being specifically and independently based on AI. Enterprises told me that many users of Microsoft’s copilot-in-office-software users didn’t know they had AI at all, or didn’t associate the features with AI, which hardly builds demand for AI among the users.
Another data point here is that a couple Cisco executives threw a little shade on the AI craze in comments to Wall Street, saying that some AI business models may be eliminated through market experience, and that this could create casualties among AI providers and perhaps an AI market correction. This is essentially how I see the AI market; a lot of hype and a slower-developing but powerful value proposition. The earnings reports of the three biggest AI spenders seems to show they’re all confident, but that investors are a bit less so. Does that mean we face a real AI correction? The Street loves a bubble, and all the excitement around the OpenAI IPO suggests they’re not ready to cast AI hype aside…yet.
