Intel is preparing to ship its first x86 processors that integrate NVIDIA’s RTX graphics technology directly onto a single chip, with an internal roadmap listing a launch window in the first quarter of 2028. The leak, first published by Igor’s Lab and corroborated by VideoCardz and Wccftech, reveals a project codenamed “Serpent Canyon” that would combine Intel CPU cores with NVIDIA GPU IP—a historic first for the two companies that have long been both partners and rivals in the PC ecosystem.

What the Leak Tells Us About Intel’s 2028 Plans

The leaked document is a product planning roadmap that outlines Intel’s upcoming processor families for client computing. According to the reports, a new category labeled “RTX SoC” appears for the first time, targeting a Q1 2028 release. This suggests that engineering work is already well underway and that Intel and NVIDIA have reached a licensing or design agreement to embed NVIDIA’s graphics architecture within an Intel processor.

Details are scarce—we don’t yet know manufacturing nodes, core counts, or which RTX GPU generation will be used. But the name “Serpent Canyon” aligns with Intel’s tradition of using geographic codenames for its platforms. Notably, Intel already sells a “Serpent Canyon” NUC mini-PC that pairs a 12th Gen Core CPU with an Intel Arc discrete GPU, so the reuse of the name likely signals a similarly high-performance target, but this time with deeply integrated NVIDIA graphics.

If the 2028 chip follows the chiplet trend that both companies have embraced, it’s plausible that Intel would use its own process node for the CPU chiplets while contracting NVIDIA’s GPU chiplets from another fab—likely TSMC, where NVIDIA already manufactures its discrete GPUs. A multi-chip module (MCM) design would allow Intel to combine its x86 prowess with NVIDIA’s graphics and AI accelerators without sharing proprietary x86 intellectual property, sidestepping one historical barrier to such a deal.

What This Means for You: Gaming, AI, and Everyday Computing

The integration of an RTX GPU directly into the main processor carries practical implications for several types of Windows users.

For Gamers

Imagine a thin-and-light laptop that can run Cyberpunk 2077 with ray tracing enabled at 60 frames per second, all without the bulk and power drain of a discrete graphics card. Today’s gaming laptops must sacrifice portability or battery life to house a dGPU; a System-on-Chip (SoC) with built-in RTX cores could deliver genuine AAA performance in ultrabooks. Existing solutions like NVIDIA’s Optimus technology attempt to switch between integrated and discrete graphics, but they add latency, driver complexity, and rarely match the efficiency of a truly unified design.

Unified memory is another game-changer. In current systems, the CPU and GPU each have their own dedicated memory pools, forcing games to copy data back and forth over the PCIe bus—a bottleneck that costs both performance and power. An RTX SoC with shared memory would eliminate that overhead, potentially reducing frame latency and enabling new features like instant game resume or seamless GPU compute without explicit data transfers.

For Content Creators and AI Users

NVIDIA’s CUDA cores and Tensor cores on an x86 SoC would bring powerful AI acceleration to mainstream Windows laptops. From real-time video upscaling in Adobe Premiere Pro to running local large language models (LLMs) for coding assistants, an integrated RTX GPU could democratize workflows currently limited to high-end workstations. Developers would also benefit from a single programming model where CPU and GPU share the same memory space—simplifying application design and potentially unlocking new classes of performance-sensitive software. For researchers and data scientists, having an RTX-capable x86 chip could mean training small neural networks locally on a laptop, rather than relying on cloud instances.

For Everyday Users

Even non-gamers and non-creators will notice the difference. Windows continues to integrate AI features—Copilot, Studio Effects, real-time captioning—that rely on neural processing units (NPUs). While Intel’s current chips have an NPU, their performance is modest. An RTX SoC would bring desktop-class AI processing into the chip, enabling faster and more responsive AI experiences while preserving battery life. Features like NVIDIA’s DLSS upscaling and frame generation might also percolate down to integrated graphics, improving visual quality and fluidity for everyday applications and video playback.

For IT Professionals and Enterprise

Enterprises deploying collaborative tools, virtual desktops, or local AI models for data-sensitive environments could see a significant performance-per-watt improvement. Maintaining a fleet of laptops without dGPUs reduces heat, fan noise, and potential points of failure, while still providing the necessary compute for demanding workloads. An RTX SoC could also simplify endpoint security, as drivers and firmware for a single chip are easier to manage and update than a hybrid integrated/discrete solution.

How We Arrived at an Intel-NVIDIA x86 SoC

The path to this rumored partnership has been years in the making. Intel and NVIDIA have had a complicated relationship: they compete fiercely in GPUs (Intel Arc vs. GeForce) and AI accelerators, but they’ve also cooperated for decades, with NVIDIA GPUs serving as the standard companion to Intel CPUs in hundreds of millions of PCs. The PC industry has been moving toward tighter integration of CPU and GPU ever since AMD launched its first Accelerated Processing Unit (APU) in 2011, merging x86 cores with Radeon graphics on a single die. Today, AMD’s APUs power the Steam Deck, the Asus ROG Ally, and the latest gaming consoles.

Intel, meanwhile, developed its own integrated graphics over many generations, culminating in the Arc-based “Xe-LPG” architecture found in Meteor Lake and upcoming Lunar Lake processors. While these iGPUs have improved dramatically—doubling gaming performance in some scenarios—they still lag behind NVIDIA’s discrete GPUs in both raw speed and critical features like DLSS and advanced ray tracing. Intel’s push into high-performance Arc discrete GPUs has also been rocky, with uneven driver support and a limited market share.

NVIDIA’s first foray into CPU integration is its Grace ARM-based server chip, which pairs with Hopper GPUs using a high-speed interconnect. But an x86 integration was long considered impossible due to Intel’s tight hold on the x86 license and NVIDIA’s reluctance to share its GPU crown jewels with a direct rival. That calculus changed when market dynamics shifted: the AI PC boom—catalyzed by Microsoft’s Copilot+ requirements and Apple’s M-series silicon—demanded ever-stronger integrated AI performance. Intel needed a differentiator to counter AMD’s rising mobile market share, and NVIDIA wanted to extend its CUDA ecosystem into the vast installed base of x86 laptops. A deal began to make strategic sense.

The first credible rumor emerged in late 2024, when analysts at Wedbush and other firms speculated about a collaborative project. Then in May 2025, a leaked Intel document hinted at “Halo” products that would incorporate NVIDIA RTX into future mobile platforms. Today’s roadmap disclosure, with a specific codename and quarter, adds significant weight to those earlier whispers.

What You Should Do Now (and What You Shouldn’t)

Given the 2028 timeframe, this leak doesn’t require immediate action for most consumers, but it can inform longer-term decisions.

  • If you’re in the market for a laptop today: Do not wait three years. Current high-end laptops with discrete RTX 40-series or upcoming 50-series GPUs will deliver excellent gaming and creative performance for the foreseeable future. An integrated RTX SoC may not outperform a discrete card on raw specs at first—it will excel in efficiency and form factor.

  • If you plan a hardware refresh cycle for your business: Consider that 2028 might be a tipping point for AI-capable, thin-client laptops. Investing now in standardized, easily upgradeable docking solutions could future-proof your fleet without committing to bulky mobile workstations. If your workflow already leverages CUDA-accelerated software, the migration path will be smoother.

  • For developers and IT decision-makers: Keep an eye on unified memory architectures and NVIDIA’s CUDA roadmap. Software optimized for shared CPU-GPU memory could run exceptionally well on future SoCs. Familiarizing your team with DirectML and CUDA programming today will pay dividends when this hardware arrives.

  • For investors and industry watchers: This leak could signal a deeper alliance between Intel and NVIDIA, potentially reshaping the competitive landscape for AMD and Qualcomm. Monitor official financial filings and product announcements for confirmation or denial.

A crucial caveat: semiconductor roadmaps change frequently. What’s penciled in for early 2028 could slip, be cancelled, or evolve into a completely different product. Treat this leak as a signal of industry direction, not a binding promise.

The Outlook: What to Watch Next

The first official confirmation might come at an Intel Architecture Day or NVIDIA GTC later this year. Intel’s next major mobile platform, Nova Lake, is expected in 2026, so the RTX SoC would likely belong to the following generation—tentatively called “Panther Lake” or another codename. We’ll also watch for regulatory hints; any deep collaboration of this scale might require antitrust clearance, given the combined market power of the two companies.

Meanwhile, AMD is not standing still. Its rumored Strix Halo APU, set for 2025, aims to bring high-performance RDNA 3.5 graphics to ultra-enthusiast laptops, potentially closing the performance gap with discrete GPUs. The RTX SoC leak suggests that the PC industry is accelerating toward a future where the boundary between CPU and GPU blurs, delivering powerful, efficient computing in ever more portable and affordable devices. For Windows users, that’s nothing but good news.