Why Smartphones Are Getting More Expensive to Build

Flagship launches often foregrounded the processor: bigger cores, faster benchmarks, a new chip name printed across the marketing slide. In 2026, the component actually straining a flagship's budget isn't the chip doing the thinking. It's the chip doing the remembering.

That's the short answer to why smartphones are getting more expensive to build: a sharp spike in memory prices has made DRAM, the phone's working memory, costlier than the processor itself. The research behind this piece documents higher costs for phone makers. It does not document a confirmed retail-price increase on any specific device sold to a consumer, and that gap matters. Nothing here shows what a 2026 flagship costs on shelves versus its 2025 predecessor at the same tier, which is the difference between why building a phone costs more and why buying one might.

The numbers are worth sitting with. Smartphone memory prices jumped 300% year over year in the second quarter of 2026, a rise large enough to push memory costs ahead of processor costs in every smartphone segment, from budget to flagship, according to Counterpoint Research. Separately, Counterpoint's teardown analysis confirms that DRAM has overtaken the system-on-chip, or SoC, the phone's main processor, as the single most expensive component inside a modern device. The research identifies memory as the main documented driver of this shift; it does not provide a complete account of what happened to cameras, batteries, displays, or any other component over the same stretch.

The effects show up further downstream, too. Global smartphone SoC shipments fell 15% year over year in the first half of 2026, a decline Counterpoint attributes to rising memory costs alongside cautious inventory management by phone makers and longer replacement cycles among consumers, not memory prices alone (Counterpoint Research).

What follows separates what's confirmed from what isn't. The cost of building a phone has genuinely shifted. Whether that shows up as a higher price, a slimmer margin a manufacturer absorbs quietly, or a spec trimmed somewhere shoppers rarely check is the open question this piece works through, along with what buyers can actually verify the next time a flagship launches.

Why smartphones are getting more expensive: memory costs outrun the processor

A quick glossary helps. The SoC handles computation and, increasingly, on-device AI. DRAM is short-term working memory, separate from storage, that everything running on the phone depends on. For this analysis, BOM, or bill of materials, means the estimated cost of a phone's components, before R&D, marketing, or margin get layered on top.

The scale of the increase looks even sharper at a shorter time horizon. Beyond the 300% annual figure, Counterpoint's Memory Price Tracker shows prices rose more than 80% quarter over quarter in Q2 2026 alone (Counterpoint Research). Those two figures measure different windows, one spanning a full year and one a single quarter, so both can be accurate at once. Together they describe a sharp increase over a short period, not a gradual drift.

Here's a hypothetical worth walking through, not a reported figure: why doesn't a 300% jump in one part translate into anything close to a 300% jump in the finished phone? Memory is one line on a much longer bill of materials, and a phone's retail price includes plenty that has nothing to do with parts at all: distribution, marketing, retailer margin, the manufacturer's own profit. If memory made up a modest slice of a phone's component costs before the spike, tripling that slice still moves the total BOM by far less than 300%, and the BOM itself is only part of what the phone costs to sell. R&D, shipping, and retailer cuts do not necessarily move in direct proportion just because one chip got pricier. That gives a manufacturer several places to spread the cost before a shopper ever sees a different number on a price tag.

Counterpoint describes the underlying shift as a "sustained structural impact" on smartphone BOMs. Its teardown-based models show the cost structure of 2026 phones has already moved measurably compared with same-tier 2025 models (Counterpoint Research). Counterpoint's own forecast suggests the pressure may extend well beyond one quarter, expecting the underlying memory shortage to persist until the second half of 2027, a projection worth treating as exactly that: a forecast, not a settled outcome (Counterpoint Research).

Think of a phone's BOM the way a restaurant thinks about food cost. If one staple ingredient triples in price, the ledger changes even though the menu hasn't. Why memory got scarce in the first place isn't addressed in the research reviewed here, and that's a real gap worth flagging rather than papering over with a guess.

A market splitting between budget phones and flagships

Memory's cost surge hit, in Counterpoint's own words, "all smartphone segments." Every tier is exposed to higher memory costs; this isn't a story confined to cheap phones or reserved for expensive ones (Counterpoint Research).

The forecast damage isn't landing evenly, though. Counterpoint projects overall SoC shipments will fall 14% year over year across all of 2026, with entry-level shipments dropping more than 30%, a far steeper decline than the market average (Counterpoint Research). As a general point of business economics rather than a figure Counterpoint quantifies directly, lower-priced phones typically run on thinner margins to begin with, which may leave them less room to absorb a memory-cost spike than premium models.

At the other end of the market, GenAI-capable smartphone SoC shipments, chips built to handle on-device AI features, grew 24% year over year across the smartphone SoC market, which Counterpoint ties to consumer demand for AI features and an ongoing premiumization trend (Counterpoint Research). Chip suppliers are feeling this unevenly, too: Qualcomm and MediaTek shipments each fell more than 25% year over year, while Apple, Samsung, Google, and UNISOC gained share, for reasons Counterpoint notes differ by company without specifying which (Counterpoint Research).

Growth in AI-capable chip shipments suggests continued demand for AI-capable phones, though that's an inference, not proof the entire flagship tier is healthy or insulated. Flagships are exposed to the same memory-price shock as everything else on the shelf, even though their memory configurations may differ from model to model. They may have more room to absorb the hit before it shows up anywhere a shopper would notice, though the research doesn't confirm flagship margins are thicker than other tiers, only that flagship margins themselves are expected to shrink from previous generations, which the next section covers.

Higher prices, thinner margins, or quieter cuts

Counterpoint states directly that flagship gross margins are unlikely to match previous flagship generations in the near term, concrete evidence of margin pressure rather than a prediction about retail prices or hardware specs (Counterpoint Research). That's the clearest evidence in the supplied research of pressure on the flagship ledger.

The cost side is already documented. Based on Counterpoint's teardown modeling, the 2026 BOM cost structure has moved measurably compared with same-tier 2025 models, a cost-side finding rather than a confirmed change to retail prices or component quality (Counterpoint Research). The pressure isn't done building, either. Counterpoint expects next-generation 2nm SoCs to add further upward pressure to smartphone BOM costs, layering a second cost increase on top of the memory problem rather than replacing it (Counterpoint Research).

Faced with a pricier BOM, a manufacturer has several realistic options, and this list isn't exhaustive: absorb the cost and accept a smaller margin, raise the retail price at an equivalent spec tier, or hold the price steady and quietly adjust something less-marketed, such as base RAM or storage allocation, battery capacity, or camera hardware. Companies can also mix approaches by delaying launches, renegotiating supply contracts, shifting their product mix toward higher-margin models, or pricing differently by region. These represent the basic ways a business responds when component costs rise faster than the price it charges, not a complete menu of every move available.

The available data doesn't identify which lever any specific 2026 flagship has pulled. Counterpoint's figures describe the market as a whole, not any individual phone, so there's no way from this research alone to say a named model cut its RAM, storage, or camera hardware in response to the memory squeeze. Confirming that requires comparing actual specs and teardowns model by model.

A flagship's price tag staying flat from one generation to the next doesn't automatically mean nothing changed behind it. The change may simply have happened somewhere other than the number at checkout.

What to watch when shopping for a flagship

Buyers can still check several concrete signals rather than taking a launch event's word for it.

• Compare the base RAM and storage tier against the equivalent 2025 model at the same price point. A quiet downgrade could show up as "same price, less storage included," which is easier to overlook than a sticker increase and worth checking specifically because it doesn't announce itself.
• Don't treat a new AI-branded chip as automatic proof of better value. GenAI-capable smartphone SoC shipments grew 24% year over year across the smartphone SoC market (Counterpoint Research), so the label alone says little about whether memory allocation or camera components held steady.
• Look for independent teardowns, official spec sheets, and side-by-side pricing comparisons as new flagships ship. Teardowns are strong evidence for identifying what memory a phone actually contains, though they can't confirm what a manufacturer paid to source it; treat them as one signal alongside published specs and pricing history, not a complete read on the economics.

What would actually prove a phone got more expensive

A single teardown or spec sheet won't settle this on its own. What would, checked together, are things a reader could realistically verify: a launch price higher than the maker's previous generation at an equivalent tier, a base RAM or storage allocation that shrank rather than held steady, an independent BOM estimate showing that model's memory configuration became materially more expensive, and evidence that memory prices were rising during the window when that phone's components were sourced. Counterpoint's data supplies the last piece for 2026 broadly. The other three depend on the specific model.

That evidence doesn't require waiting for phones that haven't launched yet. Several 2026 flagships have already shipped as of this writing, so model-level comparisons and teardowns may already exist for some devices; the gap is that the research reviewed here operates at the market level, not the model level, so it can't be used to name a winner or a loser among specific phones.

It's also worth separating two things that get conflated in coverage of this story: the BOM and the total cost of building and selling a phone. The BOM is a snapshot of what the parts inside cost to buy. It leaves out R&D amortized across a product line, software development, warranty reserves, shipping, tariffs, and the retailer's own cut, all of which move independently of what DRAM costs that quarter. A memory spike can shift the BOM sharply while barely denting the far larger total, which is exactly why a 300% jump in one component doesn't require anything close to a 300% jump in what a phone costs on a shelf.

Counterpoint has established the cost shock: DRAM now outweighs the SoC as the priciest single part in a smartphone, across every segment. What it hasn't established, because the data doesn't reach that far yet, is which manufacturers passed that shock along, which absorbed it, and which quietly moved the weight onto a spec sheet. That answer will come model by model, from launch prices, spec comparisons, and teardowns, not from a market-wide report.