Phone companies do not usually explain themselves. When a flagship arrives with last year’s processor, the launch deck simply talks about something else, and the spec sheet is left to be discovered by whoever reads that far down.
Poco went the other way this week. The F9 Ultra launched on September 1 with the same Snapdragon 8 Elite Gen 5 that powered last year’s F8 Ultra, and when asked about it directly at the launch press conference, the company gave a straight answer: memory prices.
That answer is more interesting than the phone, because it is the first time a major Android brand has publicly connected the memory shortage to a specific product decision. Everyone in the industry has been dealing with it for a year. Poco is the one that said it out loud.
📱 The short version
- The Poco F9 Ultra reuses the Snapdragon 8 Elite Gen 5 from its predecessor, breaking a long run of Poco flagships launching on the newest chip available.
- Poco named the reason on the record: rising memory costs, plus a view that the existing chip is already fast enough.
- It still costs more. $799 against $729 for the F8 Ultra, and £899 against £749 in the UK.
- Memory prices have gone vertical. Mobile DRAM and NAND contract prices have roughly doubled inside a year, driven by AI datacenter demand eating wafer capacity.
- The upgrades went elsewhere: an 8,050mAh battery, a 200MP main camera, and a 185Hz display with a dedicated graphics chip beside the Snapdragon.
- This is the template for 2026. Expect more phones that hold the chip and raise the price, and fewer that do the reverse.
What Poco Actually Said
The question came up at the F9 Series press conference, and two Poco executives answered it. Senior Product Manager Kang Lou took the cost question head on.
“The overall situation is that rising memory costs are putting a lot of pressure on all of us in the smartphone industry. Poco is not escaping these costs. From our end, we believe performance is not always about chasing the latest chipset. Of course, that’s important, but operationally speaking, having the latest chipset will cost us a lot more on the latest phone.”
Kang Lou, Senior Product Manager, Poco
Global PR lead Olivia Jiang followed with the strategic framing, saying that Poco does not make launch decisions based on the next generation of hardware, and that product timing is not something the company treats as fixed.
Read those two answers together and the shape of the decision is clear. Poco could have waited for Qualcomm’s next flagship silicon and charged accordingly, or shipped now on a chip that is already extremely fast and spend the budget elsewhere. It chose the second, and it decided the reasoning was defensible enough to say in public.
The Memory Problem Is Worse Than Most People Realize
The background to that quote is a component crisis that has been building through 2026 and shows no sign of resolving.
The cause is not a factory fire or a trade dispute. It is demand. High bandwidth memory for AI accelerators is enormously profitable, and it consumes wafer capacity that would otherwise produce the conventional DRAM and NAND that goes into phones and laptops. Manufacturers have followed the money, and everything downstream of them is now paying for it.
The numbers are genuinely difficult to absorb. Mobile LPDDR5X, the memory in every phone discussed here, was forecast to rise 78 to 83 percent quarter on quarter in the second quarter of 2026, on top of a DRAM contract market that had already climbed 90 to 95 percent in the first. NAND flash pricing for Android manufacturers moved in the same range.
Those are not retail price increases, and nobody is passing all of it through. But memory is one of the largest single line items in a phone’s bill of materials, and analysts expect overall component costs to rise another 5 to 7 percent as a result, with global smartphone output falling around 10 percent year on year. Some of the pressure comes out of margins. Some of it comes out of the spec sheet. The rest arrives at the till.
We looked at how this plays out across the wider market when phone shipments hit their worst year since 2013 and manufacturers started quietly trimming RAM, and Poco’s decision is the same force acting on a single product, just visible for once.
The Chip Poco Kept Is Not a Downgrade
Here is the part that stops this being a straightforward case of a company cutting corners.
The Snapdragon 8 Elite Gen 5 is a 3nm part with two prime cores running up to 4.6GHz and six performance cores at 3.62GHz. It is the same chip inside the Xiaomi 17 Ultra and the OnePlus 15. It was, until very recently, the fastest thing you could put in an Android phone, and in ordinary use it remains indistinguishable from whatever replaces it.
Mobile chipsets crossed into diminishing returns some time ago. The gap between last year’s flagship silicon and this year’s is now largely invisible outside benchmarks and sustained gaming sessions, and the second of those depends more on cooling than on the processor. Poco leaning on its cooling system in that quote is not deflection. For the workloads the F9 Ultra is built around, it is the more relevant variable.
The wider context is that even the chip roadmap itself is fragmenting under cost pressure. Qualcomm’s next generation is expected to split into more than one flagship tier, a change we unpacked when Qualcomm decided to divide its flagship chip in two. Manufacturers are being asked to pick a price point rather than simply take the best chip available, and that changes what “flagship” means on a spec sheet.
Where the Money Went Instead
If the processor stayed put, the upgrade budget went somewhere. It went to the three components buyers actually notice.
| Component | Poco F8 Ultra | Poco F9 Ultra |
|---|---|---|
| Chipset | Snapdragon 8 Elite Gen 5 | Snapdragon 8 Elite Gen 5, unchanged |
| Battery | 6,500mAh | 8,050mAh silicon carbon |
| Main camera | Lower resolution sensor | 200MP, f/1.68, stabilized |
| Display | Standard high refresh AMOLED | 6.9 inch, up to 185Hz |
| Extra silicon | None | VisionBoost D8 display chip |
| Launch price | $729 / £749 | $799 / £899 |
The battery jump is the most consequential change, and it is the same silicon carbon chemistry driving capacity gains across the Android market this year. Going from 6,500mAh to 8,050mAh inside a phone that has not grown noticeably is a bigger real world improvement than any processor upgrade would have delivered.
The 200MP main camera with optical stabilization and an f/1.68 aperture is joined by a 50MP periscope telephoto at 5x optical and a 50MP ultrawide. The display runs at 2608 by 1200 and peaks at 4,500 nits across a quarter of the panel. Memory is 12GB or 16GB of LPDDR5X with UFS 4.1 storage up to 1TB, which is exactly the component that has become expensive.
The 185Hz Question
The VisionBoost D8 deserves a note, because it is the one piece of genuinely new silicon in the phone and it is easy to misread.
It is a dedicated display processor sitting alongside the Snapdragon, and its job is frame generation and upscaling. Games that render internally at a lower frame rate get interpolated up toward the panel’s 185Hz ceiling, and more than twenty titles currently support it, including Call of Duty Mobile and PUBG Mobile.
Worth understanding before you weigh it
Generated frames are not rendered frames. Interpolation makes motion look smoother. It does not make the game respond faster, because the extra frames contain no new information about your input.
For a racing game or a platformer that is a real visual improvement. For a competitive shooter, where the thing you care about is latency rather than smoothness, the benefit is much smaller than the number suggests. Treat 185Hz as a display claim, not a performance one.
None of that makes it a bad feature. A dedicated chip doing frame generation is genuinely cheaper in power terms than asking the main GPU to render those frames, which matters for both heat and battery. It is just not the same thing as a faster processor, and the marketing will not make that distinction for you.
What This Means for Your Next Phone
Poco is a value brand, which is why this surfaced here first. Its buyers are unusually price sensitive and its margins are thin, so it feels component costs before a company selling $1,200 flagships does.
But nothing about the underlying pressure is specific to Poco. Every Android manufacturer is buying the same memory in the same market, and the responses available to them are the same three: absorb the cost, raise the price, or hold something back on the spec sheet. Most will do some of each, and most will not explain which.
What to expect over the next year
Chip reuse becomes normal. A new phone on last year’s flagship silicon is no longer a warning sign. Check what the chip actually is rather than whether the generation number moved.
RAM and storage tiers get quietly worse. Base configurations shrinking, or a storage tier disappearing, is the cheapest way to hold a price. Compare against the previous model rather than against the marketing.
Last year’s flagship becomes the value buy. When the new model reuses the chip and raises the price, the outgoing one at a discount is arithmetic, not a compromise.
The Bottom Line
The Poco F9 Ultra is a good phone that costs $70 more than its predecessor and runs the same processor. Whether that is fair depends entirely on how much you value a substantially larger battery, a far better main camera and a faster display, because those are what you are paying the difference for.
The more useful takeaway is the admission underneath it. A significant Android brand stood up at its own launch event and said that the cost of memory now shapes which processor goes into its phones. That has been true across the industry for a year. It is the first time anyone has been willing to describe it plainly, and it will not be the last phone this year where the same calculation was made quietly.
When the next flagship arrives with a familiar chip inside, you will at least know what happened.
