Intel Panther Lake vs AMD Gorgon Point: Which Laptop CPU Should You Buy?

Laptop CPUs became properly interesting again in 2026. Not "a slightly faster spreadsheet" interesting. Properly interesting. Intel's Core Ultra Series 3, known as Panther Lake, brought a new manufacturing process and a more complicated mix of CPU, graphics and low-power cores. AMD's Ryzen AI 400 family, codenamed Gorgon Point, took a different tack: familiar Zen 5 foundations, higher clocks, revised power management and an NPU that stretched to 60 TOPS in the flagship.

For most people, the comparison comes down to a handful of chips. On Intel's side, the Core Ultra X9 388H is the all-singing flagship, while the Core Ultra 7 358H is a more representative high-performance option. AMD counters with the Ryzen AI 9 HX 475 at the top end and the Ryzen AI 7 460 for laptops that still need serious muscle without going completely bananas on specification.

The annoying but important answer is that there is no single winner. Intel had the stronger single-threaded story and a very ambitious efficiency design. AMD had more CPU threads at the top, the higher NPU ceiling and a formidable RDNA 3.5 integrated GPU configuration. The right choice depends much more on what you do between charges than on the badge beside the trackpad.

Meet the Chips: Panther Lake and Gorgon Point at a Glance

Intel's Panther Lake family sat beneath the Core Ultra Series 3 name, with Core Ultra 5, Core Ultra 7 and Core Ultra 9 parts joined by the more graphics-focused Core Ultra X7 and Core Ultra X9 range. It is a deliberately broad platform. At one end were configurations with eight CPU cores and four GPU cores; at the other, the Core Ultra X9 388H offered up to 16 CPU cores, 12 Xe3 graphics cores and a 50 TOPS NPU.

AMD's Gorgon Point family arrived as Ryzen AI 400. The star of the range was the Ryzen AI 9 HX 475: 12 cores, 24 threads, a 5.20GHz maximum boost clock, a 16-CU RDNA 3.5 integrated GPU running up to 3.10GHz, and a 60 TOPS XDNA 2 NPU. That is a lot of silicon to put inside a laptop chassis, even before the cooling system starts having opinions about it.

Neither family should be treated as a single CPU. That is the first buying trap. A laptop built around a lower-tier Core Ultra 5 is not a slightly less exciting Core Ultra X9 388H; it may have a very different core and graphics configuration. The same is true of Ryzen AI 400. The Ryzen AI 5 430, for example, has four CPU cores and four graphics compute units, while still carrying a 50 TOPS NPU. AI capability, CPU performance and gaming potential are no longer tied together as neatly as they once were.

The useful comparison is not simply Intel versus AMD. It is the specific Core Ultra or Ryzen AI configuration inside the laptop you are considering, especially where CPU-core and integrated-graphics counts change.

Intel Panther Lake Architecture: What 18A Actually Changed

Shop Intel Panther Lake Architecture on Amazon UK

Panther Lake's headline was Intel 18A, a process node that introduced gate-all-around transistors and backside power delivery. Those phrases are not exactly the stuff of a relaxed Sunday afternoon, so here is the human version. Gate-all-around technology gives the transistor gate closer control of the channel carrying electrical current. Backside power delivery moves power connections behind the transistor layer rather than making signal and power routing fight for the same space above it.

The intended result is more efficient switching and better use of the available silicon area. In a laptop processor, that matters because every watt is negotiated between performance, fan noise and battery life. You do not buy a machine because it contains a clever transistor. You buy it because that transistor can help the machine feel snappier without making the fans sound like it is trying to leave the runway.

Intel did not present Panther Lake as one giant, pure-18A slab of silicon, however. Its multi-tile layout is important. The CPU compute tile used Intel 18A, while the powerful 12-Xe3 graphics tile used TSMC's N3E process. The Platform Control Tile used TSMC N6 and integrated I/O features including PCIe 5.0, PCIe 4.0, Thunderbolt 4, USB, Wi-Fi 7 and Bluetooth 6.0. It is a pragmatic design rather than a purity test.

At the top end, Panther Lake offered a 4P + 8E + 4LPE CPU arrangement. The P cores are intended for the heavy lifting. The standard E cores take on efficient parallel work. The LPE, or Low Power Efficiency, cores are there for the quiet background jobs: low-demand activity, light responsiveness and tasks that do not need the big cores waking up and drinking the battery for no good reason.

There is a practical benefit to this division of labour. A laptop spends far more time handling modest, irritatingly ordinary work than it does flattening a video render: keeping the system awake, moving through browser tabs, receiving messages, playing media and managing background processes. A design that can handle those moments efficiently has a better chance of feeling composed across a whole day, not just impressive during a five-minute benchmark.

AMD Gorgon Point Architecture: A Sharper Zen 5 Refresh

Shop AMD Gorgon Point Architecture on Amazon UK

Gorgon Point was built on 4nm silicon and evolved the Strix Point approach rather than throwing it in the bin and starting again. That is not a criticism. A successful laptop platform should get better at power management and clocks without making developers, manufacturers and buyers relearn everything from scratch every year.

The architecture combined full Zen 5 cores with compact Zen 5c cores. In the flagship arrangement, four full-sized Zen 5 cores shared a 16MB L3 cache, while eight Zen 5c cores shared an 8MB L3 cache. The end result was a 12-core, 24-thread chip in the Ryzen AI 9 HX 475 and HX 470. It gave AMD a useful blend: fast higher-performance cores where responsiveness matters and a larger group of compact cores for parallel work.

Move down the stack and AMD scaled the configuration sensibly. The Ryzen AI 9 465 used 10 cores and 20 threads, made up of four Zen 5 and six Zen 5c cores. The Ryzen AI 7 460 used eight cores and 16 threads, split between four Zen 5 and four Zen 5c cores. The Ryzen AI 7 445 had six cores and 12 threads, while the Ryzen AI 5 430 came in with four cores and eight threads.

The important bit is that the AI engine did not disappear once you left the flagship. The Ryzen AI 9 465, Ryzen AI 7 460, Ryzen AI 7 445 and Ryzen AI 5 430 all retained a 50 TOPS NPU. AMD therefore made a clear distinction between general compute and local AI acceleration. Someone buying a less extravagant laptop could still get substantial NPU hardware, even if they did not get the flagship's 16-CU graphics setup or 12 CPU cores.

In other words, Gorgon Point was not only a chip for people who wanted the biggest number on the box. It was a scalable laptop platform. That makes it attractive to manufacturers because they can build thin-and-light, mainstream and performance systems around a broadly shared architecture. For buyers, it makes the choice simpler: first decide how much CPU and graphics power you need, then make sure you are not accidentally paying flagship money for a configuration that does not match your work.

Gorgon Point's Ryzen AI 400 stack kept XDNA 2 acceleration well beyond the flagship, whilst CPU-core and integrated-graphics resources scaled more aggressively between tiers.

CPU Performance: Who Wins the Core War?

CPU performance is where the headline gets nuanced very quickly. Panther Lake was clearly ahead of both Strix Point and Gorgon Point in single-threaded performance in the comparison results, with an edge that could reach 10% in Cinebench. It was not a demolition job, and there were individual single-threaded tests where Strix Halo pulled ahead, but Intel's lead was described as steady rather than accidental.

That matters in the bits of laptop life that feel immediate: opening an application, browsing heavier sites, working through documents, responding to inputs and running software that still cares more about one very fast thread than it does about a heroic thread count. These are not glamorous jobs. They are the jobs that decide whether your laptop feels brisk or mildly cross with you.

AMD's answer was workload breadth and sustained multi-thread potential. The Ryzen AI 9 HX 475 combined 12 cores and 24 threads with a 5.20GHz maximum boost clock. Its 15W-to-54W configurable power range also gave laptop designers room to build very different systems around the same basic chip. A thin laptop tuned for restraint and a larger machine given more cooling headroom are not going to behave identically, even when the processor name is the same.

Intel's Core Ultra X9 388H had up to 16 CPU cores in its top configuration. But raw core count alone does not settle the question because the core types are different, power limits differ by machine, and sustained performance depends on how long the laptop can maintain its boost behaviour. This is why a single "which is faster?" result is often less useful than a buyer would like. I know. It would be lovely if there were one neat answer and we could all get on with our lives.

For a writer, student, developer working in lightly threaded tools, or anyone who values a laptop that reacts sharply in day-to-day use, Panther Lake has a very persuasive argument. For workloads that can keep many threads busy, the high-end Ryzen AI 9 HX 475 remains a serious proposition, especially in a laptop designed to use its higher power envelope. Do not choose based on core count alone. Choose based on the jobs you actually run, and whether you carry the laptop around enough to care how loudly it makes its case.

Integrated Graphics: Intel Arc Xe3 vs AMD RDNA 3.5

Integrated graphics stopped being a polite backup a while ago. In a well-specified modern laptop, they handle photo work, multi-display setups, creative acceleration, casual gaming and local AI tasks without requiring a separate graphics chip. That is especially useful in thin machines, where a discrete GPU can add heat, thickness and a noticeable appetite for battery power.

Intel's strongest Panther Lake graphics setup paired the Core Ultra X9 388H with an Arc B390 integrated GPU carrying 12 Xe3 graphics cores. Intel quoted up to 122 GPU TOPS for AI workloads from that graphics engine alone. Across CPU, GPU and NPU, Panther Lake reached a platform figure of up to 180 TOPS, with up to 120 TOPS attributed to the GPU and 50 TOPS to the NPU.

AMD's top integrated graphics configuration came with the Ryzen AI 9 HX 475 and Ryzen AI 9 HX 470: 16 RDNA 3.5 compute units, boosting up to 3.10GHz. The Ryzen AI 9 465 reduced that to 12 CUs, the Ryzen AI 7 460 to eight, and the Ryzen AI 7 445, Ryzen AI 7 435 and Ryzen AI 5 430 to four CUs. Again, the exact suffix matters. "Ryzen AI" is not a graphics performance promise by itself.

AMD's 16-CU configuration is the one to seek out if gaming and graphics work are high on your list. The Radeon integrated graphics approach had already built a strong reputation through the Hawk Point and Strix Point generations, and Gorgon Point continued with RDNA 3.5 at the top end. Intel's Arc Xe3 design, meanwhile, brought substantial AI throughput through the GPU and made Panther Lake a much more rounded platform than an "office laptop CPU" label would suggest.

The graphics battle is more than gaming: Intel's Arc Xe3 hardware contributes heavily to Panther Lake's total AI figure, while AMD's 16-CU RDNA 3.5 configuration is the standout Gorgon Point iGPU tier.

The NPU Arms Race: 50 TOPS vs 60 TOPS

The NPU is the bit of the chip designed for AI workloads that do not need to hammer the CPU or graphics hardware. It is becoming a more useful part of the laptop specification, although I would still resist buying a computer solely because one number is 10 TOPS higher than another. A local AI workload has to be supported by the software, fit the hardware and be something you actually want to run.

Intel delivered up to 50 TOPS from the Panther Lake NPU in its top models. The Core Ultra X9 388H combined that NPU with up to 122 GPU TOPS and a total platform AI figure of up to 180 TOPS. That wide platform figure is useful for workloads able to spread AI processing across different engines, but it should not be confused with the NPU rating. If software specifically uses the NPU, 50 TOPS is the relevant number.

AMD's Ryzen AI 9 HX 475 raised the dedicated NPU ceiling to 60 TOPS. The Ryzen AI 9 HX 470 landed at 55 TOPS. Below those parts, the Ryzen AI 9 465, Ryzen AI 7 460, Ryzen AI 7 445 and Ryzen AI 5 430 all delivered 50 TOPS. This is a neat strategy because it keeps a strong baseline across much of the range, while letting the flagship claim a concrete hardware advantage.

For most buyers, the sensible conclusion is straightforward. Both platforms had ample local AI hardware for the software features aimed at this generation of laptops. AMD's 60 TOPS figure gives the Ryzen AI 9 HX 475 extra headroom at the high end. Intel's wider 180 TOPS platform figure makes its flagship compelling where software can make use of CPU, GPU and NPU resources together. They are different advantages, not identical measures.

Battery Life: Intel's 27-Hour Claim vs AMD's 24-Hour Claim

Battery claims always need a cup of tea and a raised eyebrow. Intel stated that Panther Lake laptops could reach up to 27 hours of Netflix streaming. It also quoted up to nine hours of Microsoft Teams video conferencing in a 3×3 setup with studio effects enabled. Those are impressive claims, and they reflect the obvious purpose of Panther Lake's low-power architecture: let the LPE cores handle lighter tasks without waking bigger resources unnecessarily.

AMD quoted up to 24 hours of video playback and up to 20 hours of web browsing for Gorgon Point laptops. Those are also very strong figures, and they underline how seriously both companies now treat efficiency. A laptop that lasts through a day of work and an evening of entertainment without drama is more useful than a benchmark monster that begins begging for a socket just after lunch.

But neither claim should be read as a promise for your own routine. Screen brightness, display technology, battery capacity, wireless activity, browser tabs, background syncing, the installed memory, cooling settings and whether the machine is using its integrated graphics at full tilt all change the outcome. Video playback is comparatively friendly to modern laptop hardware. A mixed day with video calls, documents, cloud services, browser work and a few hundred open tabs because apparently we enjoy suffering is a harder test.

Intel's 27-hour streaming claim is the larger headline figure, and the LPE-core approach gives Panther Lake a particularly convincing theory of operation for low-demand work. AMD's claims were still excellent, though, and Gorgon Point's updated power management was central to its design. In practice, the individual laptop matters at least as much as the processor family. A gorgeous high-resolution display and a modest battery can undo a lot of silicon cleverness.

Battery life remains a laptop-level decision. Panther Lake's LPE cores and Gorgon Point's revised power management matter, but the display, battery and chassis configuration still decide the result you live with.

Memory, Connectivity and Expansion: The Quiet Buying Factors

Processor comparisons can become a bit obsessed with the loud numbers: clocks, cores, TOPS and graphics units. Memory support and connectivity are less glamorous, but they are often what determine whether a laptop remains useful for several years. A strong processor paired with too little memory is still a frustrating computer. It just frustrates you very quickly.

Panther Lake supported up to 96GB of LPDDR5 memory or up to 128GB of DDR5 DIMMs, with LPCAMM support and transfer rates up to 9600 MT/s. That is a notably flexible memory proposition. It gives manufacturers routes to compact, integrated-memory designs as well as systems with DDR5 DIMMs, and it creates a very high ceiling for people working with large projects, demanding development environments or a frankly adventurous number of simultaneous applications.

AMD's top four Gorgon Point processors supported memory speeds up to 8533 MT/s, while the rest of the range maxed out at 8000 MT/s. Those are still rapid memory specifications, and the important point is that Ryzen AI 400 covered a wide class of machines. The top-end HX parts and the mainstream chips did not have identical memory limits, which is another reason to check the exact processor rather than assuming every Ryzen AI 400 laptop is configured the same way.

Intel's Platform Control Tile integrated up to 12 PCIe 5.0 lanes and up to eight PCIe 4.0 lanes, alongside two Thunderbolt 4 ports, two USB 3.2 ports, eight USB 2.0 connections, Wi-Fi 7 and Bluetooth 6.0. This is a useful strength for buyers who spend their days docked, moving large files, connecting storage or using a broader ecosystem of accessories.

For office work and web use, either platform's memory support is more than capable. For serious creative work, software development, local AI experimentation or anything that can gobble RAM for breakfast, Intel's 128GB DDR5 ceiling is a meaningful advantage. For lighter machines, AMD's high memory-speed support remains perfectly compelling. As ever, the exact notebook specification matters: support from the processor is not the same thing as a manufacturer choosing to fit the maximum amount.

Choosing the Right Tier: Flagship, Performance and Mainstream

The flagship comparison is fun, but most sensible purchases happen one or two steps below the maximum-spec part. That is where the Ryzen AI 7 460 becomes especially interesting. It offered eight cores and 16 threads, a maximum boost frequency of 5.10GHz, eight integrated graphics compute units and a 50 TOPS NPU. It is a balanced configuration rather than a stripped-down AI token.

The Ryzen AI 9 465 takes a step up for people who need more CPU and graphics resources: 10 cores and 20 threads, a 5.00GHz maximum boost clock, 12 graphics compute units and a 50 TOPS NPU. At the very top, the Ryzen AI 9 HX 475 has the full 12-core/24-thread CPU, 16-CU iGPU and 60 TOPS NPU. The HX 470 keeps the same CPU and iGPU configuration but brings the NPU down to 55 TOPS.

At the other end, the Ryzen AI 5 430 has four CPU cores, eight threads, four integrated graphics compute units and a 50 TOPS NPU. That makes a useful point: it is possible to buy into local AI capability without buying into top-tier general CPU or gaming performance. If your needs are documents, browsing, calls, media and occasional AI-assisted features, that can be absolutely fine. If you want to edit video, run demanding creative work or play more ambitious games on integrated graphics, it is plainly not the same class of chip as the HX 475.

Intel's range worked similarly. Core Ultra 300V was an eight-CPU-core configuration with four GPU cores. Core Ultra 300H offered 16 CPU cores but dropped the integrated GPU for gaming-oriented systems. Core Ultra X 300H combined 16 CPU cores with a 12-core GPU, and the Core Ultra X9 388H sat as the headline model with up to 16 CPU cores, 12 Xe3 GPU cores and a 50 TOPS NPU.

This is why I would not recommend buying purely by generation. Buy by tier. A well-balanced mainstream chip paired with sufficient memory, a sensible display and a decent battery is often a smarter laptop than an expensive flagship CPU compromised elsewhere. The processor is the engine, not the whole car.

The sensible middle of each range matters more than the flagship halo. Ryzen AI 7 460 and higher-tier Core Ultra options are where many buyers will find the best balance of CPU, graphics and AI hardware.

Best For: Which Platform Fits Your Type of Work?

This is the part where we stop admiring architecture diagrams and make an actual decision. There are clear strengths on both sides, but the best choice changes depending on whether you want a long-lived work machine, an integrated-graphics powerhouse, a local AI test bed or simply a laptop that gets out of its own way.

There are two particularly easy mistakes to avoid. First, do not assume the flagship chip is automatically the best fit. If you do not need a 16-CU iGPU or 12 CPU cores, you may be better served by a cooler, quieter, more balanced design. Second, do not assume an NPU figure tells you everything about future-proofing. Memory capacity, graphics resources, connectivity and the software you use are all part of the longer-term picture.

The Practical Verdict Before You Buy

Panther Lake and Gorgon Point are both excellent examples of where laptop CPUs had reached by mid-2026. They are no longer just processors; they are mixed compute platforms, juggling CPU cores, integrated graphics, dedicated AI acceleration, media work, memory bandwidth and connectivity. The buyer's job is not to pick the chip with the most exciting slide. It is to find the laptop implementation that gets the balance right.

Intel's case is strongest if you value a consistent single-threaded advantage, a sophisticated low-power core arrangement, a very high memory ceiling and a leading claimed video-streaming figure. The Core Ultra X9 388H also gives Intel a powerful integrated Arc Xe3 configuration and an up-to-180-TOPS platform number when the CPU, GPU and NPU are considered together.

AMD's case is strongest if you want the highest dedicated NPU figure, the most substantial Gorgon Point integrated graphics configuration or a top-end CPU with 12 cores and 24 threads. The Ryzen AI 9 HX 475's 60 TOPS NPU and 16-CU RDNA 3.5 graphics made it the obvious maximum-spec AMD option, while the Ryzen AI 7 460 looked like the sensible all-rounder for buyers who wanted a lot of capability without chasing the biggest possible configuration.

There is no universal winner between Panther Lake and Gorgon Point. Intel's strengths centre on single-threaded responsiveness, memory headroom and efficiency design; AMD's centre on flagship NPU throughput and top-end RDNA 3.5 graphics resources.

Intel Panther Lake vs AMD Gorgon Point FAQ