Intel, AMD Detail Major AI Chip Push at CES 2026

January 6, 2026 at 19:47 UTC

6 min read

AI chip launch by Intel and AMD at CES 2026 with focus on edge and data center markets

Key Points

Intel launches 18A-based Panther Lake and Core Ultra 3 as its AI PC and edge platform
AMD unveils MI455 and MI440X data center GPUs plus Helios rack-scale AI system
Both chipmakers target on-device and edge AI with new Ryzen AI Embedded and Core Ultra designs
AMD projects AI compute demand reaching more than 10 yottaflops within five years

Intel Showcases 18A-Based Panther Lake and Core Ultra Series 3

Intel used CES 2026 in Las Vegas to present what executives described as a pivotal product moment, unveiling its Core Ultra Series 3 processors and new laptops powered by the Panther Lake design. Both are built on Intel’s US-manufactured 18A process, which the company is positioning as the foundation of its next generation of AI PCs and edge systems. Top Core Ultra Series 3 chips such as the X9 and X7 offer up to 16 CPU cores, 12 Xe GPU cores and 50 NPU TOPS, with Intel citing up to 60% better multithread performance, 77% faster gaming and battery life of up to 27 hours versus the prior generation. Intel said more than 200 PC designs, along with certified embedded and industrial systems, are planned around the platform. Pre-orders for Series 3 laptops begin January 6, with additional designs rolling out through early 2026.

Panther Lake Laptops and Intel’s Turnaround, Manufacturing Strategy

At CES, Intel outlined new laptops from major global manufacturers using Panther Lake, with orders starting Tuesday and worldwide availability from January 27. Senior vice president Jim Johnson described 2026 as a strategic inflection point for Intel and the broader semiconductor industry, highlighting AI as a key opportunity. Panther Lake, now folded into the Core Ultra Series 3 lineup, is built on the 18A process and targets AI-related workloads in notebooks and future handheld gaming devices. Intel said improved manufacturing yields after earlier production issues are enabling a larger 2026 deployment. The 18A node, which introduces gate-all-around transistor technology, underpins Intel’s effort to re-establish leading-edge in-house manufacturing while still using Taiwan Semiconductor Manufacturing Co. for some production. Management has linked the success of these products to its broader turnaround narrative after a 2025 focus on cost reductions and attracting outside investment.

Intel’s AI, Edge Ambitions and Market Expectations

Intel is tying its PC roadmap directly to ambitions in advanced manufacturing and industrial AI by combining integrated AI acceleration and Arc graphics into a single system-on-chip that can also serve robotics, smart cities and other edge workloads. The company highlighted partner Advantech’s plans to integrate 18A-based Core Ultra Series 3 chips into industrial motherboards and edge AI systems, showing that the same silicon can target both consumer PCs and higher-value edge deployments. Investment narratives cited in the coverage project Intel revenue of $58.1 billion and earnings of $5.2 billion by 2028, implying 3.1% annual revenue growth and a $25.7 billion earnings increase from a current loss position. One fair value estimate referenced in the articles put Intel at $38.14 per share, about 3% below its current price, while more bullish analyst scenarios assume revenue could reach roughly $62.1 billion and earnings about $8.7 billion by 2028 if AI and foundry catalysts ramp as planned.

AMD Expands Data Center AI Lineup With MI440X, MI455 and Helios

Advanced Micro Devices used its CES 2026 keynote to broaden its AI data center portfolio. CEO Lisa Su introduced the MI440X accelerator, aimed at smaller corporate data centers that prefer on-premise deployments, and emphasized the higher-end MI455X as a step forward in data center AI performance. AMD said MI455-based systems power data center racks sold to customers including OpenAI, whose president Greg Brockman joined Su on stage to stress that chip innovation is key to safely and effectively scaling AI. AMD also previewed Helios, a rack-scale platform built around Instinct MI455X accelerators and EPYC “Venice” CPUs, designed so entire racks of servers can operate as a single machine. Helios systems based on MI455X and Venice are expected to go on sale later this year. Looking further ahead, Su previewed the MI500 series accelerators, slated for 2027 and projected to deliver up to 1,000 times the performance of the MI300 series that debuted in 2023.

AMD Targets Edge, Embedded and PC AI With Ryzen AI Platforms

Beyond data centers, AMD announced new Ryzen AI processors for PCs and embedded systems, aiming to run AI models directly on devices. At CES, the company introduced Zen 5-based Ryzen AI Embedded P100 and X100 Series processors, which combine Zen 5 CPUs, RDNA 3.5 GPUs and XDNA 2 NPUs in compact BGA packages for automotive digital cockpits, industrial automation, smart healthcare and humanoid robotics. The P100 Series targets 4–6 core applications such as in-vehicle experiences, operating in a 15–54 watt range and supporting temperatures from -40°C to +105°C, while the X100 Series offers higher core counts and AI TOPS for more demanding tasks. The integrated RDNA 3.5 GPU supports up to four 4K or two 8K displays at 120 frames per second, and the XDNA 2 NPU delivers up to 50 TOPS, which AMD says is up to three times higher AI inference performance than prior models. A unified, open-source virtualization software stack allows multiple operating systems to run securely in parallel, aiming to reduce costs and speed time-to-market for OEMs and industrial customers.

AMD’s AI Strategy, Market Outlook and Valuation Context

AMD’s CES announcements were framed as part of a broader strategy to span AI from cloud infrastructure to embedded edge systems. The company highlighted that AI is now used by an estimated 1 billion people worldwide, with adoption projected to reach about 5 billion as it becomes embedded in work, healthcare, transportation and consumer devices. Su said global AI computing capacity has grown from roughly one zettaflop in 2022 to more than 100 zettaflops, and projected that enabling AI everywhere will require more than 10 yottaflops of compute over the next five years, or 10,000 times more than in 2022. Investment narratives cited in the articles project AMD revenue of $46.2 billion and earnings of $9.0 billion by 2028, implying 18.5% annual revenue growth and a $6.8 billion earnings increase from $2.2 billion today. One fair value estimate referenced put AMD at $283.57 per share, about 28% above its current price, while another discounted cash flow model cited an estimated fair value of $324.80, roughly 31.9% above a recent price of $221.08. Commentators noted that these outlooks assume continued AI-driven growth while acknowledging risks such as Nvidia’s lead in high-end GPUs and potential export controls.

Key Takeaways

Intel is using its 18A process and Panther Lake-based Core Ultra 3 platform to link its AI PC, edge and foundry ambitions to a visible product ramp in 2026.
AMD is extending its AI reach from hyperscale data centers into smaller enterprise, PC and embedded markets through MI440X, Helios and Ryzen AI Embedded chips.
Both companies are positioning around rapidly rising AI compute needs, with AMD explicitly projecting a move from today’s zettaflop-scale infrastructure toward yottaflop-level capacity.