--- title: "AI Is Making Software Cheaper and Your Next Phone More Expensive" date: 2026-09-02 category: Chips & Compute site: NeuroAI canonical: https://neuroai.site/a/na-ai-makes-software-cheaper-hardware-costl language: en --- # AI Is Making Software Cheaper and Your Next Phone More Expensive > From 1 September, Huawei, Xiaomi and Honor raised prices across current models, some by RMB 200–1,000. The cause is memory: procurement prices reportedly went from around RMB 500 to RMB 2,200. AI data centres took the capacity. ## Key takeaways - From **1 September 2026**, Huawei, Xiaomi and Honor all raised prices on models currently on sale, with several devices up **RMB 200 to RMB 1,000**. Meizu said current prices would hold at most one more month. - The direct cause is memory. Per supply-chain reporting, memory chip procurement prices rose from roughly **RMB 500 to RMB 2,200** — a more than fourfold increase. Qualcomm chips were also reported to be getting a double-digit adjustment from September. - The mechanism: a high-end AI server carries tens of times the memory of an ordinary phone. Fabricators allocate capacity to the higher-margin AI orders first, and phone memory queues behind them. - Memory expansion to market typically takes **12–18 months**, so this pricing cycle most likely runs into next year. ## The illusion we have been living under There is a widespread impression from the last two years that as AI advances, everything gets cheaper. Copywriting costs nothing, translation costs nothing, even video editing has fallen to a tenth of what it did. Meanwhile, from 1 September, the phone in your pocket got more expensive. ## Does AI deserve the blame? It does, though not in the way people assume. The proximate cause is memory chips. The reason memory is expensive is that AI servers are absorbing production capacity at speed. A single high-end AI server needs tens of times the memory of an ordinary phone. Suppliers naturally allocate wafer capacity to the AI orders that pay better, and consumer memory ends up at the back of the queue. This is not phone-maker greed. It is global capacity re-queueing. ## The counterintuitive part Here is the change most people have not registered. For twenty years, Moore's law meant compute kept getting cheaper. In the AI era the demand curve has changed shape: training one large model consumes hundreds of thousands of accelerators, and every one of those accelerators needs large amounts of high-bandwidth memory. The result: **intelligence at the software layer is deflating, while cost at the hardware layer is inflating.** You will feel that split more and more sharply. AI writes your proposal for free. The phone, laptop and car you run that AI on get more expensive every year. ## How to handle the increase The short answer: do not panic-buy, and do not wait for a quick reversal. Memory takes 12 to 18 months from capacity expansion to market. That means this cycle most likely extends into next year. Some practical notes: - **If you genuinely need a new phone**, do not wait — waiting probably costs more. Look at last year's flagships; they are currently the best value. - **If it can go either way**, replace the battery and keep the current device another year. A battery swap costs far less than the premium on a new phone right now. - **If you are hunting a bargain**, avoid launch pricing. Memory cost pressure hits new models first, so old-stock clearances create a window. ## The knock-on effect people miss It is not only phones. Solid-state drives, memory modules, graphics cards, smart cars — anything needing large-capacity storage is affected by the same capacity squeeze. Builders are already comparing notes: an SSD that cost RMB 500 last year sells for close to RMB 1,000 this year for the same model. If AI is the origin of this inflation, then its cost is ultimately shared out across every consumer — whether or not you use AI. ## The longer read Behind the price rise is a structural signal: **AI demand has begun passing cost through to the consumer side.** Until now, AI-driven price increases were a data centre matter, invisible to ordinary people. Now they are on your phone bill. Over the next few years, every product that needs a lot of memory — phones, computers, smart cars, games consoles — will feel the same push. Once you understand that, you stop asking why things got more expensive in the age of AI. AI deflates software and inflates hardware. The subscription fee you saved comes back out of the price of your next phone. ## Honest limitations The memory price figures come from supply-chain reporting rather than vendor disclosure, and procurement prices vary by specification and contract. Retail price adjustments differ by model and region. The 12–18 month capacity timeline is an industry rule of thumb, not a forecast for any specific product. *Sources: supply-chain and retail pricing reports from 1–2 September 2026; vendor statements. Information only.* --- Published by NeuroAI (https://neuroai.site/) — https://neuroai.site/a/na-ai-makes-software-cheaper-hardware-costl Free to quote with attribution and a link to the original.