Apple released the iPhone 16 series on Monday, and the biggest selling point is Apple Intelligence. The artificial intelligence system on Apple devices offers some compelling features, such as rewriting emails, generating custom emojis, and a significantly upgraded Siri. But behind it all, artificial intelligence is bringing another big change to the iPhone: more RAM.
Although Apple has never talked about the RAM in its smartphones, mike rumor Every iPhone 16 model now comes with 8GB of RAM, up from 6GB of RAM in last year’s base model. Apple isn’t the only one making such changes. Last month, Google made similar changes to its AI-heavy Pixel 9; both the standard and Pro versions got more RAM, to 12GB At least You can get it this year.
The driving force behind these RAM increases appears to be artificial intelligence. Artificial intelligence is the new must-have feature this year, and its memory requirements are also very large. Smartphone makers are adding memory now because they need to—whether they say it out loud or not.
AI models need to respond quickly when users call them, and the best way to achieve this is to permanently load them into memory. RAM reacts much faster than the device’s long-term storage; it would be annoying if you had to wait for an AI model to load to get a quick email summary. But AI models are also quite large. Even a “small” device like Microsoft’s Phi-3 Mini takes up 1.8GB of space, which means taking up memory for other smartphone functions that previously used that memory.
You can see this directly on Pixel phones. Last year, Google did not enable local AI capabilities on the standard Pixel 8 model due to “hardware limitations.” Spoiler: it’s RAM. Seang Chau, vice president and general manager of Android, said in March The Pixel 8 Pro can better handle the company’s smaller AI model, the Gemini Nano, as the phone has 4GB more RAM than the Pixel 8, at 12GB. 8 Too much memory will be lost when supporting this feature.
“It’s not that easy to just say, okay, we’re going to enable this on Pixel 8 as well,” Chau said. Google eventually allowed the Gemini Nano to run on the Pixel 8, but only to those willing to run the phone in developer mode — people who, Chau said, “understand the potential impact on the user experience.”
These trade-offs are why Google decided to boost RAM across the board on the Pixel 9. Stephanie Scott on edge.
So, will all the extra RAM be used just for AI, or will users see performance improvements across the board? A lot depends on the implementation and the size of these models. Google added 4GB of memory to support native AI features and says you’ll see improvements in both. “Just for our latest Pixel phones,” Scott wrote, “you get better performance and improved AI experiences from the extra RAM.” She added that Pixel 9 phones “will be able to keep up with future artificial intelligence “Intelligence advances.” But if these advances mean larger models, it likely means they will consume more memory.
The same memory improvement trend is also happening in the notebook computer field. Microsoft stipulated earlier this year that only machines with at least 16GB of memory can be considered Copilot Plus PCs, which are laptops that can run native Windows AI features. Rumor has it that Apple is also planning to add more RAM to its next laptops, after years of offering 8GB of RAM by default.
Additional memory will be needed, especially if the laptop manufacturer wants to load larger models natively. “I think most operating systems will always have LLM loaded,” Hugging Face CTO Julien Chaumond told me in an email, “so 6-8GB of RAM is optimal, in addition to other things the operating system is already doing. Also unlock this feature.” The model can then be loaded or unloaded with “a little model on top of it to change certain properties,” such as the style of image generation or the LL.M.’s specific domain knowledge, Chaumond added. (Apple describes a similar approach.)
Apple has not yet made it clear how much RAM is required to run Apple Intelligence. But every Apple device running it (going back to the 2020 M1 MacBook Air) has at least 8GB of RAM. It’s worth noting that last year’s iPhone 15 Pro with 8GB of memory could run Apple Intelligence, while the standard iPhone 15 with 6GB of memory couldn’t.
Apple’s head of artificial intelligence, John Giannandrea, told Daring Fireball’s John Gruber in a June interview that limitations such as “device bandwidth” and the size of the neural engine would make artificial intelligence features too slow to use on the iPhone 15. “RAM is part of the whole,” Federighi said in the same appearance.
The iPhone 16’s 2GB RAM boost ultimately isn’t a lot, but Apple has long been slow to expand baseline RAM on its devices. Even if the company is small, any growth here feels like a usability win.
We still don’t know how useful Apple Intelligence will be, or whether a slight jump in memory will be enough to allow today’s iPhone to run tomorrow’s AI features. One thing seems certain, though: As artificial intelligence becomes more prevalent throughout the industry, we’ll see more improvements in these types of hardware.
