Nvidia hit a record high on October 2 and is closing in on a $6 trillion valuation, a milestone that underscores how much of the market's AI enthusiasm still runs through chip demand. Dan Niles, founder of Niles Investment Management, is watching that rally from the sidelines. His two preferred plays remain Alphabet and Meta Platforms, even after a stretch that's made at least one of them harder to buy.
Google Cloud's 82% jump changes the math
Alphabet reported second-quarter results on July 22, and the cloud number did the heavy lifting. Google Cloud revenue rose 82% to $24.8 billion. Niles called that the fastest acceleration among Google, Microsoft, and Amazon, a distinction that matters because the three are chasing many of the same enterprise customers.
The search business didn't collapse either. Google Search and other revenue rose 17% to $63.3 billion in the same quarter, a result that pushes back on the idea that AI chatbots are eating the core franchise. Backlog — signed customer commitments that haven't been booked as revenue yet — reached $514 billion in June. That figure gives investors a rough sense of how much work is already contracted.
What Alphabet doesn't have yet is a competing AI agent. OpenAI has launched its Dots agent, and Niles flagged that gap directly. It's the kind of product hole that doesn't show up in a revenue line but could matter if rivals turn agents into the default way people search. For now, Niles said investors should keep checking search data to make sure Google Search isn't slipping.
Meta's Muse app and the missing cloud arm
Niles first flagged Meta on September 7, when the stock was down about 7% year to date. The next day, Meta launched Muse, an AI agent that browses the web and completes tasks for users. By Niles's account, the Muse app now ranks above ChatGPT on the App Store — a consumer signal that's hard to fake and easy to measure.
Meta's structural difference from its peers is that it doesn't rent out AI computing power through a public cloud. Google, Microsoft, and Amazon all do. Niles sees a cloud arm as Meta's next way to monetize its AI spending, which so far has been poured into models and products without a direct rental revenue stream attached.
He likes Meta less than he did in September, after the run-up, but he still sees momentum in the name. JPMorgan raised its Meta price target to $920 from $820 last month, a move that reflects how sell-side analysts are re-rating the stock as Muse gains traction.
Why Nvidia's record matters here
Nvidia's October 2 record high and its approach toward a $6 trillion valuation are both a vote on AI infrastructure spending and a warning about concentration. If chip demand is the market's AI proxy, then Alphabet and Meta are the companies trying to turn that spending into consumer and enterprise products.
Niles's positioning reflects a split view: he's comfortable owning the application layer through Alphabet and Meta, while treating Nvidia's run as a benchmark rather than a buy signal he's acting on.
What to watch next
The next hard data points are Alphabet's coming quarterly report and any update on how Muse is holding its App Store ranking. Search data, as Niles keeps saying, is the tell for whether Google's core is holding. And Meta's cloud intentions remain an open question — no product has been announced, but the logic of monetizing AI compute is sitting right there.
