Claim Validationdata analysis skilltechnology

Is Every New iPhone Really the "Biggest Leap Forward Ever"?

What everyone believes

“Apple frames each new iPhone as a major leap over its predecessor — in reality, most year-over-year changes in performance, price, design, and sales are small, incremental improvements rather than the dramatic overhauls the marketing implies.”

The data's verdictmixed

Every September, Apple says some version of the same sentence: this is the best iPhone we’ve ever made. Eighteen generations in, that phrase has described a 64-bit chip transition, a headphone-jack removal, and everything in between. iPhone 18 is due at Apple’s presumed September 2026 keynote, and it will almost certainly get the same line. The question isn’t whether Apple says it — it says it every year. The question is what the data behind it actually looks like.

The claim vs. the reality

Start with the number Apple puts on stage. Five documented cases, spanning over a decade, show three different outcomes for the same kind of claim.

Apple's keynote-day performance claim vs. the independently measured Geekbench result
Not a complete 18-generation record — Geekbench's own benchmark versions aren't numerically comparable across eras, and two of this analysis's planned primary sources (Geekbench Browser, a full EveryMac benchmark pull) weren't accessible this pass. These five are independently verified, cited spot checks across different points in the product's history.

The iPhone 12 case is the sharpest example. Apple’s own headline that year — “up to 50% faster” — sounded like a claim about the previous iPhone. It wasn’t. It was a comparison to a competing Android chip. The actual generation-over-generation gain, A14 versus A13, independently measured: about 19%. Two years later, the iPhone 16 Pro’s “up to 15% faster” claim was, if anything, modest — Geekbench measured closer to 18%. Same kind of sentence, same stage, two very different relationships to the underlying number.

Price: it went up once, then started coming back down

$0

the iPhone 7's 2016 launch price — the year Apple's pricing stopped being a $199-with-contract number and started being what you actually pay unlocked.

Comparing 2007’s $499 iPhone to 2025’s $799 iPhone as one continuous line is misleading in a specific way: for most of that stretch, the sticker price was never the real price. From 2007 through 2015, iPhones sold for $199 (sometimes less) because the number was paired with a two-year carrier contract that recouped the rest of the cost elsewhere. That regime ended with the iPhone 7 in 2016, when Apple moved to unlocked, contract-free retail pricing — and the “price” jumped to $649 not because the phone suddenly cost 3x more to make, but because it stopped being subsidized.

Base iPhone price, inflation-adjusted to 2025 dollars
Two different pricing regimes, not one trend: 2007-2015 reflects carrier-subsidized contract pricing; 2016 onward is full unlocked retail. The dashed segment is a rough trend-based forecast for the iPhone 18, not an Apple-confirmed price — see Methodology.

The genuinely surprising finding is what’s happened since: in real, inflation-adjusted terms, the base iPhone has gotten cheaper every single year since the iPhone 12’s 2020 peak — from $1,003 down to $799 in five straight years, a trend that runs directly against the “prices only go up” assumption. Trend-extrapolating that same line to iPhone 18’s expected launch puts the rough forecast at roughly $760-$875 — essentially flat, not the price hike that tends to circulate in pre-launch rumor coverage.

Design: real overhauls happen, on a schedule

Major redesign vs. iterative refresh, by year
Coded against documented physical/material changes (chassis, display technology, biometric system, port) — not aesthetic opinion. Highlighted (larger) dots are the genuine chassis-level redesigns, named in the text below; everything else refined the previous year's body. Hover any point for its model name.

Coded strictly by physical/material change — new chassis, new display technology, a new biometric system, a new port — seven of eighteen launches were genuine redesigns: the iPhone 4 (2010), iPhone 5 (2012), iPhone 6 (2014), iPhone X (2017), iPhone 12 (2020), 15 Pro (2023, the first port change since 2012), and iPhone Air (2025). That’s a real overhaul roughly every 2-3 years — the rest were refinements of the previous year’s body. It’s also, not coincidentally, close to Apple’s old “tick-tock” pattern (3G/3GS, 4/4S, 5/5S, 6/6S, X/XS all shared a chassis across two generations) — the cadence people vaguely remember turns out to hold up numerically once you actually code it. Whether iPhone 18 lands as a “tock” year is exactly the kind of thing reported-but-unconfirmed coverage is currently speculating about — more on that below.

Two smaller claims worth checking

0GB

the base iPhone storage tier from the iPhone 5 (2012) through the iPhone 6s (2015) — four straight years, while app and photo/video file sizes kept growing. A certified class action (Sept 2024) covers California buyers of 16GB iPhones and iPads from that era over how much of it iOS itself consumed.

“More storage than ever” wasn’t true for a real four-year stretch — base capacity only doubled to 32GB with the iPhone 7 in 2016. And in 2020, Apple removed the charger and EarPods from the iPhone 12 box, citing carbon and e-waste reduction; reporting estimates the move saved Apple roughly $6.5 billion, with no corresponding price cut passed to buyers. Both are real, one-time, checkable events — not part of a smooth incremental trend in either direction.

Sales: mostly real growth, one real decline, then a plateau

iPhone unit sales / shipments by year
Solid dots: Apple's own disclosed figures (through fiscal 2018). Hollow dots: third-party shipment estimates (IDC), the only source available since Apple stopped disclosing units in fiscal 2019. Years without a confirmed figure are left out rather than estimated — the line simply spans a wider gap there.

The overall shape is real, sustained growth — from 1.4 million units in a partial first year to a 231-million-unit peak in 2015. But “sales keep growing every year” isn’t quite right either: fiscal 2016 posted the first-ever year-over-year decline in iPhone history, down 8.4%. And in the years Apple has stopped disclosing exact figures, independent shipment estimates show a plateau, not sustained growth — 206 to 234 million units through the early 2020s, before a fresh IDC-estimated record of 247.4 million in 2025, driven by iPhone 17 demand.

The other incremental claim: software

Hardware isn’t the only place a keynote promise and its delivery date can drift apart. Apple Intelligence’s overhauled Siri was demoed and promised at WWDC in June 2024, tied directly to that year’s iPhone 16 marketing. Apple delayed it in March 2025 — by its own account, the agentic features only worked correctly about two-thirds of the time — then delayed it again, per reporting in February 2026. It’s now targeting spring 2026, meaning the “biggest Siri overhaul ever” will likely ship closer to iPhone 18’s window than to the iPhone it was actually announced alongside.

Verdict

Mixed — and mixed in both directions, not just “somewhere in between.”

Some of Apple’s specific claims genuinely don’t hold up: the iPhone 12’s “50% faster” line wasn’t even measuring the thing it sounded like it was measuring, and the iPhone 6s’s 70%/90% claims ran meaningfully ahead of what independent benchmarks found. The base storage tier really did stagnate for four years. Software promises really have slipped from their announced ship dates, more than once.

But the flat “it’s incremental hype every year” read doesn’t survive contact with the data either. Real chassis-level redesigns happen on a genuine, countable cadence — seven of eighteen generations, roughly every 2-3 years, not zero. Two of the five performance claims checked out closely against independent measurement, one undersold itself. And the most counter-intuitive finding here cuts against the “everything’s overpriced now” instinct entirely: in real dollars, the base iPhone has gotten cheaper every year since 2020.

What actually varies isn’t whether Apple says “biggest ever” — it says that every year, regardless. What varies is whether the specific, checkable number behind that sentence backs it up. Some years it does. Some years it doesn’t. iPhone 18 will get the same sentence either way.

Preparing for iPhone 18

This section is explicitly separate from the verdict above — everything here is a trend extrapolation or sourced pre-announcement reporting, not evidence the verdict rests on.

Price: if the 2016-2025 real-price trend holds, the base iPhone 18 lands around $760-$875 in today’s dollars — essentially flat versus the iPhone 17’s $799, not the increase that tends to get predicted before every launch.

Design: iPhone 18 Pro, Pro Max, and (per widely-reported, not Apple-confirmed, coverage) Apple’s first foldable iPhone are expected at a presumed September 2026 keynote; the standard iPhone 18 and 18e are reported for spring 2027 instead of alongside the Pro models. If the foldable reporting holds up, 2026 would land as a “major redesign” year in the same cadence the seven prior ones followed — not a break from the pattern, a continuation of it.

Software: Apple Intelligence’s delayed Siri overhaul is currently targeting spring 2026 — meaning by the time the standard iPhone 18 ships, the feature promised alongside the iPhone 16 in 2024 should, finally, actually be there.

None of this is a prediction of what Apple will say on stage. It’s a rough map of what the last eighteen generations of actual data suggest is plausible — worth having on hand the next time a keynote calls something the biggest leap ever.

Sources

Methodology

Four independently-checkable axes: performance, price, physical design, and unit sales, plus two smaller documented callouts (base storage capacity, in-box accessories) and a software-promise-lag timeline, chosen because each could independently support or undercut the belief. Performance uses five curated, independently-cited claimed-vs-measured pairs spanning 2013-2024 rather than one continuous 18-generation index — Geekbench 2 through 6 aren't numerically comparable across versions, pre-2015 iPhones only ran old versions, and this pass's automated tools couldn't pull a complete raw benchmark table from the two candidate primary sources (Geekbench Browser blocks automated access; EveryMac's archive didn't extract completely). Price is charted in two internally consistent regimes — carrier-subsidized contract pricing (2007-2015, cheaper on paper but a different product than what it's compared against) and full unlocked retail pricing (2016-present) — inflation-adjusted to 2025 dollars via BLS CPI-U; the two regimes are shown but never blended into a single trend line, since the 2016 jump reflects a business-model change, not organic price growth. The iPhone 18 price forecast is an ordinary-least-squares linear regression on the 2016-2025 real-price series only (slope ≈ -$12.5/year, residual standard error ≈ $56) — a rough trend extrapolation, not a prediction of what Apple will announce. Unit sales use Apple's own disclosed figures (10-K/earnings-release reporting) through fiscal 2018, when Apple stopped disclosing them, and third-party shipment estimates (IDC) after. Design changes are coded major-redesign vs. iterative based on documented physical/material changes (chassis, materials, display technology, port), not aesthetic judgment. Full source-by-source notes, including which fetches failed and how each gap was handled, are in research/iphone-incremental-upgrades/ in the project repo.

Limitations

The performance axis is five documented spot checks, not a complete generation-by-generation record — it cannot say what every one of the other 13 generations' claim-accuracy looked like, only that overstatement, accuracy, and outright misleading framing all demonstrably occurred at different points across the decade. A few individual unit-sales fiscal years (2008 exact figure, 2014, 2017, 2019, 2022) are not pinned to a single confirmed number and are omitted from the chart rather than estimated — the chart connects the confirmed years it has, which occasionally spans more than one year per line segment. Both the SEC's own filings and the BLS's calculator blocked automated access during this pass; the unit-sales and CPI figures used here are corroborated secondary reporting of that primary data, not the filings themselves. The design-change classification is a documented-physical-change coding, not a reception or sales-impact judgment — a 'major redesign' year isn't claimed to have sold better or been better reviewed than an iterative one. iPhone 18 does not exist yet: the Pro/Pro Max/foldable models are expected at Apple's presumed September 2026 keynote and the standard iPhone 18/18e are reported for spring 2027, per pre-announcement industry reporting, not confirmed Apple specifications. Nothing in the forecast section states or implies an actual iPhone 18 spec, price, or feature — the regression is a trend extrapolation with a stated uncertainty band, and the design/sales/software context is explicitly sourced to reporting, flagged as unconfirmed where it is.

~ $ produced with the data analysis skill — the six-phase playbook from this site's data era.