Intel’s foundry business is at a pivotal crossroads. Company leadership has acknowledged that moving from 14A to more advanced process nodes depends on securing large-scale external customers. Internal demand alone cannot sustain the massive R&D and capital required for next-generation fabs. To remain competitive, Intel must attract major partners.
Recently, a now-deleted Intel video drew industry attention. It revealed a reference SoC, codenamed “Deer Creek Falls”, manufactured on the Intel 18A process and based on the Arm AArch64 architecture—not x86.
The SoC integrates:
- 2 × PCIe controllers
- 4 × memory channels
- A heterogeneous CPU cluster: 4 efficiency cores, 2 mid-power optimized cores, 1 high-performance core
This “big–medium–little” configuration mirrors common Arm SoC strategies in mobile and embedded markets. Its purpose is clear: to demonstrate that Intel 18A can successfully support Arm designs, lowering barriers for potential foundry customers.
Why It Matters #
Intel does not plan to commercialize Deer Creek Falls. Instead, it serves as a reference platform — a working silicon demo to prove that complex Arm-based SoCs can be fabricated on Intel 18A.
For potential customers, a functioning chip is far more persuasive than PDKs or spec sheets. It directly addresses concerns about EDA tool readiness, design rules, and ecosystem support.
By showing actual silicon, Intel signals that it is ready to manufacture Arm-based designs at scale, if customers commit.
18A vs 14A: Intel’s Strategic Focus #
While Deer Creek Falls validates Intel’s 18A node, the company’s primary foundry offering for external clients is 14A.
- 18A: targeted mainly at Intel’s own CPUs, GPUs, and select projects.
- 14A: marketed as the core node for external customers, featuring:
- RibbonFET gate-all-around transistors
- PowerVia backside power delivery
- Roadmap for high-NA EUV lithography
These innovations aim to position Intel as a peer competitor to TSMC and Samsung. But sustaining the investment requires anchor customers committing to volume production.
Potential Big Customers #
Industry rumors suggest Apple and NVIDIA are evaluating Intel’s 14A node. Even limited trial production would mark a milestone, boosting confidence in Intel Foundry Services.
If either company commits to high-volume manufacturing, it would:
- Provide stable cash flow for Intel’s advanced nodes
- Justify investments in testing, packaging, and ecosystem tools
- Establish Intel as a credible alternative to TSMC
Bigger Picture #
Intel’s dual-track strategy blends technology validation with market outreach:
- 18A “Deer Creek Falls” SoC → Proves Arm compatibility and silicon maturity.
- 14A process focus → Positions Intel as a volume foundry for external partners.
For customers, the decision goes beyond transistor specs. It involves toolchain maturity, IP ecosystem depth, and yield reliability.
For Intel, the takeaway is clear: demonstration chips build credibility. Deer Creek Falls is less about the product itself and more about signaling that Intel is open for business as a foundry partner for the Arm ecosystem.