Intel’s foundry business is at a pivotal moment. Company leaders have acknowledged that advancing beyond 14A and into future nodes depends on securing substantial external customer demand. Internal consumption alone cannot justify the enormous R&D and capital costs required to sustain process innovation. To remain competitive, Intel must prove it can win large-scale partners.
Recently, a now-deleted Intel video revealed a reference SoC, codenamed “Deer Creek Falls,” fabricated on the Intel 18A process. What makes it noteworthy is that it is based on the Arm AArch64 architecture, not Intel’s traditional x86. The chip integrates:
- Two PCIe controllers
- Four memory channels
- A CPU cluster featuring a heterogeneous core design: 4 efficiency cores, 2 optimized mid-tier cores, and 1 high-performance core
This “big–middle–little” configuration mirrors popular Arm SoC designs in mobile and embedded markets. Its message is clear: Intel wants to prove to external customers that its processes can seamlessly support the Arm ecosystem.
Why It Matters #
Intel is not preparing Deer Creek Falls for commercial release. Instead, the chip serves as a reference platform—a working demonstration that complex Arm-based designs can be successfully taped out on Intel 18A. For potential foundry customers, this is far more convincing than technical documentation or PDKs alone. It directly addresses longstanding concerns about toolchain adaptation, design rules, and ecosystem readiness.
By showcasing an actual silicon proof point, Intel aims to ease customer hesitation and highlight a smoother path to adopting its processes.
18A vs. 14A: Where Intel’s Focus Lies #
Although Deer Creek Falls was built on 18A, Intel’s external foundry push is centered on 14A.
- 18A remains primarily for Intel’s own CPUs, GPUs, and select projects.
- 14A is positioned as the flagship node for external customers, featuring:
- RibbonFET all-around gate transistors
- PowerVia backside power delivery
- Planned High-NA EUV lithography
Together, these innovations are designed to rival TSMC and Samsung in performance, efficiency, and transistor density. But sustaining such a costly roadmap depends on securing anchor customers willing to commit early capacity.
Potential Big-Name Partners #
Industry rumors suggest that Apple and NVIDIA are evaluating Intel’s 14A process. Even limited trial runs from such companies would be a milestone, validating Intel’s process capabilities and strengthening confidence in its foundry business. A full-scale production agreement would be transformative—bringing stable cash flow, expanding packaging and test capabilities, and completing the ecosystem loop.
The Bigger Picture #
Intel’s strategy blends proof of capability with market outreach:
- 18A reference SoC (Deer Creek Falls): Demonstrates technical strength and Arm compatibility.
- 14A process push: Targets external mass production and long-term competitiveness.
For customers, decisions will hinge not just on transistor performance, but also on toolchain maturity, IP ecosystem depth, and production yields.
For Intel, the deeper value of Deer Creek Falls lies in its symbolism: a declaration that Intel is ready to be a serious foundry partner for the Arm ecosystem—and willing to back that claim with working silicon.