What BGR saw on Amazon
BGR's Christian Wait reported that Amazon's best-selling CPU list was almost completely one-sided. His count was one Intel chip in the top 20, the Core i7-14700K at No. 17. He identified the Ryzen 7 9800X3D and Ryzen 7 5800X3D as the top two sellers. BGR says the rest of the list is Ryzen.
BGR called the 14700K a previous-generation Raptor Lake part, not one of Intel's newest Core Ultra chips. It also said Newegg's desktop bestseller list looks similar. BGR gave no timestamp, screenshot or sales figures for either retailer, so treat the Newegg claim as unverified. This piece does not independently confirm either list.
Rankings move
BGR's snapshot is one data point. Two earlier reports show the Amazon US list shifting within weeks:
- August 24: Tom's Hardware's Zak Killian reported that two Intel chips had entered Amazon's top 10. The Core Ultra 7 270K Plus was No. 6 at $290, and the Core i7-14700K was No. 9 at $312.
- September 14: Club386's Ben Hardwidge reported the i7-14700K at No. 14 on Amazon US. On Amazon UK the 14700KF was at No. 10, in each case as the first Intel chip on its chart.
The storefronts and dates differ, so these are not directly comparable. Together they show that "Intel has one chip in the top 20" describes one day, not a permanent verdict. A ranking also says nothing about how many units sold or how good the chips are.
Why Ryzen sells so well at retail
Club386 gives plausible reasons, which are its opinion rather than measured causes:
- DDR5 memory is very expensive. The 14700K accepts DDR4 or DDR5, while the Arrow Lake Refresh chips need DDR5.
- AMD's sockets have lasted longer. The Arrow Lake socket takes only Arrow Lake chips, and Club386 notes that it cannot run Raptor Lake or Nova Lake CPUs.
- X3D chips with 3D V-Cache are strong in games. Club386 counted five X3D parts above all of Intel's chips on Amazon US.
Tom's Hardware adds that the 14700K can slot into aging LGA 1700 systems and that it works with DDR4. That likely explains why older platforms keep selling.
BGR's own gaming claims are broader. It says X3D chips usually beat Intel's equivalents by 30% or more in games. It also says AMD has committed to the AM5 socket into 2029. I have not verified either claim, and the 30% figure in particular depends heavily on the game, the resolution and the GPU.
Shipment data tells a different story
The best independent context is Mercury Research's Q2 2026 estimate of unit shipments, as reported by Tom's Hardware and others. AMD held 34.9% of desktop CPUs, against Intel's 65.1%. That is up from 33.2% in Q1 2026 and 32.2% in Q2 2025.
Other segments from the same report:
| Segment (Q2 2026) | AMD | Intel |
|---|---|---|
| Desktop | 34.9% | 65.1% |
| Client (desktop plus mobile) | 30.3% | 69.7% |
| Mobile | 28.9% | 71.1% |
| Overall x86 | 30.7% | 69.3% |
Desktop volume fell sharply. Mercury reports desktop shipments down more than 20% year over year, and The Register notes that AMD's decline was smaller than Intel's. Mercury's explanation: higher PC prices and limited GPU supplies are hurting end demand for desktop PCs. AMD's desktop share therefore rose partly because its shipments fell less, not only because they grew.
A Business Stats compilation cited by shattered.io puts AMD's Q4 2025 desktop share at 36.4%, a record, with Q1 2026 dipping below that. I have not confirmed that figure against Mercury's own publication.
Don't repeat the 92% figure
BGR says AMD was estimated to hold 92% of the CPU market in 2025, citing TechPowerUp. I found no support for that as a CPU number. A TechPowerUp story from June 2025 does carry a 92% figure, but it is about graphics cards. It reports NVIDIA's 92% share of the add-in board GPU market in Q1 2025, per Jon Peddie Research, with AMD at 8%. It looks like BGR conflated the two, and the CPU figure should be dropped.
What PassMark measures
BGR also cites PassMark, which says AMD is starting to outpace Intel in overall desktop share. PassMark's own page explains what its chart is. The graph counts benchmark baselines submitted by users, so it reflects CPUs in use rather than CPUs purchased. PassMark adds that it also reflects the installed base from the last 5 to 10 years. It is a participation measure from a self-selecting sample of benchmark users, not a sales statistic.
Three different measures
| Source | What it measures | What it can't tell you |
|---|---|---|
| Amazon bestseller chart | Recent sales ranking at one retailer | Total units, other stores, product quality |
| Mercury Research | Estimated units shipped by quarter, including OEM and prebuilt PCs | Which individual models retail shoppers prefer |
| PassMark | Share among CPUs whose owners ran its benchmark | Current sales or the whole installed base |
The measures agree on direction. AMD is gaining, and DIY retail is where it is strongest. They disagree on magnitude, and "Intel has little relevance" fits only the Amazon list. Prebuilt desktops, laptops and business PCs, where Intel has much more weight, don't appear in a DIY bestseller chart.
Practical takeaways for builders
- Check the date and storefront before citing any bestseller list.
- Match the CPU to your platform. An existing LGA 1700 board with DDR4 makes the 14700K a reasonable upgrade. A new DDR5 build opens up AM5 and Intel's newer socket.
- Count the memory in the total cost. Mercury and Club386 both point to DDR5 prices as a major factor in current buying.
- Think about the upgrade path. BGR notes that Intel's expected Nova Lake desktop chips are reported to need a new LGA 1954 socket, while Club386 says AMD's current sockets can take future chips. Both depend on rumors and roadmaps that may change.
- Weigh your workload. BGR says the Core Ultra 7 270K Plus is strong in productivity at about $300, a point Tom's Hardware's price snapshot partly supports. Check reviews for your own applications.
Amazon's chart is a useful read on what DIY builders are buying this week. It does not prove which brand is best, and it does not show that Intel's market has disappeared.