Intel has kicked off the new year with its latest laptop silicon, the Core Ultra 2, codenamed Arrow Lake. This series, targeting various laptop segments, comes in three configurations:
- Core Ultra 200HX and 200H for gamers and content creators seeking high performance.
- Core Ultra 200U for ultraportables with an emphasis on power efficiency.
While these processors boast neural processing units (NPUs) for AI workloads, Intel’s new lineup struggles to match the high standards set by Microsoft’s coveted “Copilot+ PC” badge.
Falling Short of the TOPS Standard
To qualify for Microsoft’s “Copilot+ PC” label, a laptop must feature NPUs capable of delivering 40+ trillion operations per second (TOPS). Unfortunately, Intel’s Arrow Lake chips max out at just 13 TOPS for the 200HX and 200H series and 11 TOPS for the 200U processors.
For context, competitors like Qualcomm and AMD are already boasting desktop NPUs with 45 and 50 TOPS, respectively. This makes it clear that Intel’s new chips won’t qualify for the “fastest, most intelligent Windows PCs” badge anytime soon.
Intel tried to counter this limitation by promoting its chips’ “platform TOPS”—a combined performance measure of the CPU, GPU, and NPU working together. In this scenario, Intel claims a peak of 99 TOPS, but the standalone NPU performance remains significantly behind its rivals.
Despite this, Intel insists it’s leading the charge in AI-driven PC innovation. Michelle Johnston Holthaus, Intel’s interim co-CEO and CEO of Intel Products, emphasized the company’s focus on bolstering its AI PC portfolio. She stated, “Intel is only going to continue strengthening its AI PC product offerings in 2025 and beyond.”
Arrow Lake HX Series: Power Meets AI Integration
The HX lineup in the Arrow Lake series is designed to cater to high-performance use cases, such as gaming and professional content creation. Here’s a breakdown of the Arrow Lake HX chips and their specifications:
| Model | Cores | L3 Cache | TDP | Base Clock | Max Boost | Availability |
| Ultra 9 285HX | 24 | 36MB | 55W-160W | 2.8GHz | 5.5GHz | 1H 2025 |
| Ultra 9 275HX | 24 | 36MB | 55W-160W | 2.7GHz | 5.4GHz | 1H 2025 |
| Ultra 7 265HX | 20 | 30MB | 55W-160W | 2.6GHz | 5.3GHz | 1H 2025 |
| Ultra 7 255HX | 20 | 30MB | 55W-160W | 2.4GHz | 5.2GHz | 1H 2025 |
| Ultra 5 245HX | 14 | 24MB | 55W-160W | 3.1GHz | 5.1GHz | 1H 2025 |
| Ultra 5 235HX | 14 | 24MB | 55W-160W | 2.9GHz | 5.1GHz | 1H 2025 |
Notably, Intel’s Lunar Lake processors, released in September 2024, are the only chips in its portfolio to meet Microsoft’s Copilot+ NPU standards with a performance of 48 TOPS.
Arrow Lake H Series: Midrange Options
The H series targets a broader audience, from gamers to everyday laptop users, offering a balance between performance and efficiency. However, like the HX lineup, these processors fall short of Microsoft’s TOPS requirements.
Here’s a detailed look at the H series:
| Model | Cores | L3 Cache | TDP | Base Clock | Max Boost | Availability |
| Ultra 9 258H | 16 | 24MB | 45W-115W | 2.9GHz | 5.4GHz | February 2025 |
| Ultra 7 265H | 16 | 24MB | 28W-60W | 2.2GHz | 5.3GHz | February 2025 |
| Ultra 7 255H | 16 | 24MB | 28W-60W | 2.0GHz | 5.1GHz | February 2025 |
| Ultra 5 235H | 14 | 18MB | 28W-60W | 2.4GHz | 5.0GHz | February 2025 |
| Ultra 5 225H | 14 | 18MB | 28W-60W | 1.7GHz | 4.9GHz | February 2025 |
Arrow Lake U Series: Ultraportable Focus
The U series is optimized for lightweight laptops and ultraportables, prioritizing battery efficiency while delivering capable performance.
| Model | Cores | L3 Cache | TDP | Base Clock | Max Boost | Availability |
| Ultra 7 265U | 12 | 12MB | 15W-57W | 2.1GHz | 5.3GHz | February 2025 |
| Ultra 7 255U | 12 | 12MB | 15W-57W | 2.0GHz | 5.2GHz | February 2025 |
| Ultra 5 235U | 12 | 12MB | 15W-57W | 2.0GHz | 4.9GHz | February 2025 |
| Ultra 5 225U | 12 | 12MB | 15W-57W | 1.5GHz | 4.8GHz | February 2025 |
Expanding to Desktops and the Edge
Alongside its laptop announcements, Intel also unveiled new desktop processors and products designed for edge computing.
Desktop Processors
The Core Ultra 200S series has been expanded to include several 65-watt and 35-watt models. These chips are designed for mainstream desktop users and are expected to find applications in the education, retail, healthcare, and manufacturing sectors.
Low-End Chips
Intel also introduced Twin Lake processors for low-end laptops and embedded devices. These chips offer modest improvements in clock speeds compared to Alder Lake, positioning them as budget-friendly options.
18A Process Technology
Intel highlighted its advancements in chipmaking with the upcoming 18A process, roughly equivalent to a 1.8-nanometer process. While Intel has produced sample products using this technology, volume production isn’t expected until the second half of 2025.
Intel’s AI PC Ambitions: Marketing vs. Reality?
Despite its claims of leadership in AI-driven PCs, Intel’s Arrow Lake chips face stiff competition. Qualcomm and AMD are delivering far superior NPU performance, and even Intel’s Lunar Lake lineup barely meets the Copilot+ standards.
Intel’s marketing emphasizes its vast hardware and software ecosystem, but the underwhelming performance of its latest NPUs could make it challenging to maintain credibility in the AI PC market.
For now, Intel remains optimistic about its trajectory, promising a stronger AI portfolio in the years ahead.
Key Takeaway: While Intel’s Arrow Lake chips may offer incremental improvements and innovative features, their inability to meet Microsoft’s Copilot+ standards underscores the fierce competition in the AI PC space. Buyers should weigh their options carefully, especially with alternatives from AMD and Qualcomm delivering superior AI performance.