At the start of the year, tech executives were sure that a new generation of phones and PCs powered by artificial intelligence would bring new customers to those categories that were having a hard time. They might have had too much hope. The most recent data shows that the shift in how people use AI is still moving slowly.
This week, Alphabet GOOGL 1.03%’s Google showed off its newest Pixel smartphones, which are said to have AI features like the ability to turn text into images and enhance photos in new ways. They don’t seem interesting enough to make people want to buy a new $1,000 gadget.
The way Apple AAPL 0.59% did things at its developer conference in June wasn’t all that different. It showed off small features like text summarisation and making your own emojis that will be part of Apple Intelligence later this year. AI that is just basic, no matter what you call it, probably won’t start the long-awaited iPod boom.
It’s possible that things will get worse for computers than for phones. Even though AI PCs, which have special chip circuits that let them run demanding AI applications, caused a lot of excitement at first, short-term optimism for PC demand is fading.
Sales of PCs have already had a rough year. According to research firm IDC, shipments of PCs around the world only went up 3% from the same time last year to the June quarter. This was after going up 2% from the March quarter to the June quarter.
It looks like momentum might slow down for the rest of 2024. Intel INTC 0.87% gave guidance earlier this month that was much lower than what Wall Street expected. The company said this was because the economy was not looking as good as they had thought.
Then there’s the lack of data on laptops, even though they make up the vast majority of PC sales.
KeyBanc Capital Markets analyst John Vinh lowered his third-quarter laptop PC unit forecast to a 4% drop from the previous year on Monday. He did this after looking at the most recent shipment numbers from Taiwan’s biggest notebook makers.
More than 90% of the laptop PCs made in the world are made by original design manufacturers (ODMs) in Taiwan. Vinh said that the number of laptops shipped in July was less than what he thought it would be—13.6 million instead of 14.3 million. That’s a pretty big miss.
What’s maybe most disappointing is that new PC chips aren’t driving innovation, even though hardware companies are making a lot of noise about AI. AMD 0.81%’s new Ryzen 9000 processors came out this month. They are based on the company’s newest chip architecture, which is called Zen 5.
Unlike the Zen 4 from two years ago, reviews of the new chips are mixed, and most of them say they don’t make much of a difference in performance. Eurogamer, a popular magazine in the gaming industry, wrote that “the price-to-performance ratio here is much poorer than we’ve come to expect from Ryzen,” which made them question the upgrade itself.
Many PC bulls are likely to be let down by AMD’s new chips if they were hoping for a big upgrade cycle.
Do not think that my doubts are a general view of AI, though. AI on phones and PCs isn’t really taking off yet, but companies and cloud-computing providers are spending more and more on AI infrastructure. And these businesses are getting more and more out of it.
The demand for AI-focused graphics processing units in data centres is still higher than the supply. Lisa Su, CEO of AMD, said last month that she now thinks those AI GPUs will bring in more than $4.5 billion in 2024. This is up from her earlier prediction of $2 billion at the end of last year.
Su also said that the supply chain for AI GPUs would “remain tight” until 2025, even though AMD will get more processing power from its partners later this year.
Nvidia NVDA 1.40%, the market leader in data centre GPUs, will release its second-quarter results on August 28. This will tell us a lot more about how much people want them.
AI is being used by businesses to help with many things, from customer service and marketing to better IT support within the company. AI may be the most important change because it helps businesses learn more about their customers and themselves.
The CEO of Walmart WMT 0.65%, Doug McMillon, said on Thursday that the company was using generative AI to make more than 850 million pieces of data in its product catalogue better. McMillon said that in a world without AI, 100 times as many people would have been needed to do that job. Ai models are also being used by the company to make its website search and app chatbots better.
During Walmart’s earnings call on Thursday, he told investors, “We’re finding real ways to use generative AI to improve the customer, member, and associate experience.” “This technology has a lot of different uses and affects almost every part of our business. We’ll keep testing and deploying AI and generative AI applications around the world.”
Walmart’s use of AI shows how quickly the new technology has spread through the economy as a whole. This isn’t just a movement in Silicon Valley.
I have been saying for a year that companies like Nvidia that are building AI data centres will do very well. I’ve told businesses that deal with PCs and smartphones to be more careful.
That difference is clearer now than ever before.
