The Pixel 9a’s 8GB RAM is a bad joke for an AI phone promising to last seven years

Google has one of the best update policies of any smartphone brand. Even the new and very affordable Pixel 9a will receive the same seven years of OS and security updates as its farm more pricey Pixel 9 Pro models. At $499, that’s an absolute bargain if you plan to keep the handset for the foreseeable future.

However, Google somewhat undermines this premise with a mediocre memory configuration. I can overlook the limited 128GB of storage to save costs, even if most of us will quickly fill it up with new photos, videos, and increasingly hefty apps before the decade’s end. There’s a 256GB version that you absolutely should consider if you want to make the most of the phone for this long.

No, my bigger problem with the Pixel 9a is its measly 8GB of RAM. Measly might sound like an exaggeration; however, Google bills the Pixel 9a as a phone specifically engineered to bring AI features like Gemini Live, on-device Recorder translations, and its best-in-class camera features to the masses. With that, we’re anticipating seven years of Pixel Drops packed full of new AI features over the Pixel 9a’s lifespan. 8GB RAM isn’t going to cut it for running next-gen AI tools, especially if you want the privacy benefits of running them offline. In fact, it’s barely even adequate today.

You only have to look at Google’s flagship models to see why. The Pixel 9 Pro reserves 2.64GB of its RAM for AI, keeping its model in memory at all times for added responsiveness. Thankfully, the Pro phones have a healthy 16GB RAM, so don’t miss that dedicated pool all too much. The smaller Pixel 9, with just 12GB RAM, doesn’t preload an AI model, however, and by Google’s own admission, it can be less responsive when loading AI tools on demand.

Unfortunately, the Pixel 9a’s 8GB of RAM is even more limited. The Pixel 9a doesn’t feature the same version of Gemini Nano as its more RAM-heavy counterparts. Instead, Google ships the phone with an “extra extra small” version called Gemini Nano 1.0 XXS. This is the same model found on the Pixel 8, rather than the XS model in the rest of the Pixel 9 series, and the smaller size jettisons everything except text support, meaning no Screenshots, Call Notes, and possibly other features. Even so, the XXS model is still too large to sit in the Pixel 9a’s memory at all times, meaning it’s less responsive to boot up as well.

So, the AI experience is already cut down, and yet Google’s sub-3GB Nano model is very small compared to larger text-based AI models you can already run on your smartphone. For example, the llama3 8b model found on the ASUS Zenfone 12 Ultra weighs in at 4.7GB, and the compact 8 billion parameter version of Deekseek R1 is 4.9GB. Models of that size would leave the Pixel 9a with very little memory for anything else, and these aren’t even the most cutting-edge models you can already run today. Upgrading to far more capable 14b models, essentially the next size up, clock in at around 9-10GB, already outsizing the Pixel 9a’s total memory pool. Ouch. If Google ever decides that it needs a bigger Large Language Model to keep up with the competition, the Pixel 9a won’t be able to run it.

This trend doesn’t just apply to text-based models. AI is already widely used in image editing and generation, and trends are pointing toward using more models to aid with various aspects of photo and video capture as well. Better models will inevitably have to become bigger to make huge improvements, barring some significant leaps forward in model compression. Who knows how far the AI journey will go, but one thing is for sure: following along will require plenty of RAM as well as a solid promise of future updates.

Of course, Google may not make the jump to such big models any time soon and will most likely continue to rely on cloud computing for the vast majority of its more advanced workloads, especially on lower-end handsets like the 9a. In that case, Google’s budget phone could keep up, at least in theory. However, Video Boost shows that Google already restricts advanced cloud-based AI features to Pro-tier customers. We will surely see more Pro-only features when the Pixel 10 and future generations roll around, and the 9a might still miss out even if they’re backported to other models. Not to mention that online features can come and go or be suddenly paywalled at Google’s will, and the 9a isn’t decked out enough to safeguard against this with robust offline processing.

I was already skeptical that the flagship Pixel 9 would keep up with the Pros due to its limited RAM and lack of advanced cloud computing tools. The budget Pixel 9a is destined to ride all the way back in third class as far as Google’s future AI plans are concerned. I would gladly have traded that new IP68 rating or macro-focusing camera for 12GB of RAM, thereby making the Pixel 9a a properly future-proofed budget AI phone.

There are plenty of other great reasons to buy the Pixel 9a, of course. Google’s photography brilliance on a budget, its flagship-grade Tensor G4 processor, and those funky color options, to name just a few. Plus, plenty of customers will quite rightly not care about this AI nonsense anyway. However, I’d be very wary of Google billing the Pixel 9a as a long-term AI purchase. While it will run many (but certainly not all) of the Pixel AI tools available today, there’s no guarantee it’ll keep up in the coming years. In fact, it almost certainly won’t.