By Eric J. Savitz
Netflix did something this past week that I suspect it will come to regret.
Every earnings season, one of the single most closely watched metrics is the number of net new Netflix subscribers. Beyond even revenue and profits, it is the one number that most reliably predicts how the streaming video giant's shares will respond to quarterly results.
For instance, in the four trading days in January after Netflix reported fourth-quarter earnings, the company's share price rallied 17%, despite reporting revenue in line with its own guidance and profits that missed by a few pennies. Why? Because Netflix added 13.1 million net new subscribers, well ahead of the company's own forecast for 8.7 million net adds.
To be clear, Netflix has been hinting for a while now that it would like to tamp down the Street's focus on subscriber growth. In late 2022, the company announced it was ending its practice of providing quarterly guidance on subscribers, although it continued to report the actual figure every quarter. But that's about to change.
While announcing March-quarter results, Netflix said that starting in 2025 it would stop reporting its subscriber count altogether. Both Netflix CEOs -- Greg Peters and Ted Sarandos -- downplayed the import of the news. Investors seem unconvinced.
Peters said on the company's post-earnings webcast that the change is motivated by a desire to focus on the key metrics that Netflix uses to measure its business, including revenue, operating income, per-share earnings, and free cash flow. The duo also noted that Netflix now has a wider variety of subscription plans at assorted price points, with some ad-supported and others not, and a more varied set of revenue sources. "We think this is a better approach that reflects the evolution of the business," Sarandos said.
Wall Street is not happy about this. What it suggests is that maybe there isn't going to be a lot more subscriber growth from here. Netflix has been driving growth in recent quarters with two initiatives: a crackdown on password sharing and the introduction of a cheaper ad-supported subscription tier. That allows them to lure into the paid subscriber pool potential subscribers who were either freeloaders or hesitant to pay up for an ad-free plan.
But the growth effects of those two moves seem destined to decelerate over time; there are only so many password-borrowers to go after. In short, what it suggests is that Netflix's business model is maturing. This for a company trading for close to 30 times expected 2025 earnings, which through Thursday's close had rallied 82% in 12 months. No surprise that the stock fell almost 9% after the earnings report, despite results that topped estimates on all measures.
This situation echoes Apple's decision in late 2018 to stop providing the Street with iPhone unit sales numbers. "A unit of sale is less relevant for us today than it was in the past given the breadth of our portfolio and the wider sales price dispersion within any given product line," Apple Chief Financial Officer Luca Maestri said on a November 2018 conference call. And here's an interesting thing: In fiscal 2018's fourth quarter, Apple sold 46.9 million iPhones. And while the number is volatile, depending on seasonal factors and product cycles, the Street consensus estimate for Apple iPhone units for the June 2024 quarter is 46 million.
In this week's cover story, I walk through companies big and small trying to compete with Nvidia in the artificial-intelligence chip market. What that piece doesn't address is the things you can do with the combination of super-powerful chips and multibillion-parameter larger language models.
Meta Platforms, parent of Facebook and Instagram, this past week provided an illustration of just how capable AI models have become. Meta launched Llama 3, the most recent iteration of its family of open-source AI models, along with an updated version of Meta AI, a chatbot that for some months now has been embedded in most of the company's social-media platforms. On Thursday, Meta launched a web version of Meta AI, which puts the company in direct competition with the other leading chatbots: Google Gemini, OpenAI's ChatGPT, Anthropic's Claude, and the like.
I urge you to try Meta AI, and in particular its real-time image generation feature. You may have experimented with text-to-image services offered by Adobe, OpenAI, and Google, but this takes the experience to another level, updating the image on the fly while you are typing your query. Less than two years ago, the prospect of simply typing something into a text window and having it appear on screen seemed preposterous. With Meta AI, the image now changes before your eyes. It's simply stunning.
Now, to be clear, there are some limitations. You can't ask for images of real or historic celebrities -- no Joe Biden, no Moses, no Queen Elizabeth. And there seem to be some limits on how many real-time images you can create in a 24-hour period, or at least that's what Meta AI said when I asked it the question. Meta is not directly monetizing the site -- there are no ads, and no subscription is required, though it works better if you log in with your Facebook account, which, among other things, remembers your previous queries. Meta AI says queries aren't shared with Facebook for ad targeting or other purposes.
One thing that's clear is that with AI, competitive lines are blurring. Cloud companies are now chip companies. Social-media sites look like search engines. Meta seems to be taking on Adobe, Google, and OpenAI all at once. The pace of change is breathtaking. And we're just getting started.
Write to Eric J. Savitz at eric.savitz@barrons.com
This content was created by Barron's, which is operated by Dow Jones & Co. Barron's is published independently from Dow Jones Newswires and The Wall Street Journal.
(END) Dow Jones Newswires
April 19, 2024 12:53 ET (16:53 GMT)
Copyright (c) 2024 Dow Jones & Company, Inc.
Comments