Intel’s Bold ASIC Strategy Could Redefine Its Future in the AI and Foundry Race

Intel’s Next Big Move: Building a Custom ASIC Empire to Reinvent Its Foundry and x86 Future

Intel is once again trying to redefine its future, this time by going deep into the ASIC and custom silicon business. After years of losing ground to NVIDIA and AMD in the AI and high-performance computing race, the company’s leadership – headed by CEO Lip-Bu Tan – has unveiled a bold new direction: to transform Intel into a leading design and manufacturing powerhouse for purpose-built chips serving external clients. And this isn’t just another business pivot. It’s an attempt to merge Intel’s foundry, IP, and design ecosystems into one unified strategy that could finally unlock the company’s long-promised potential.

At the core of this transformation is Intel’s new Central Engineering Group (CEG), a division meant to bring together every bit of engineering muscle across the company. Led by Srini Iyengar – formerly of Cadence Systems – the CEG will lead Intel’s renewed push into ASICs and custom design services. Iyengar’s background, combined with Lip-Bu Tan’s deep history in design tool ecosystems, makes this move more than just corporate restructuring. It’s a strategic play to place Intel at the center of the next big wave in semiconductor demand: custom silicon for AI, networking, and data centers.

Intel’s CEO put it clearly during the Q3 earnings call: “Our new ASIC and design service business will deliver purpose-built silicon for a broad range of external customers. This will not only extend the reach of our core x86 IP but also leverage our design strength to deliver a full spectrum of solutions – from general-purpose computing to fixed-function designs.”

In simpler terms, Intel wants to build chips not just for itself but for the entire industry. While NVIDIA and AMD dominate AI accelerators, there’s a growing number of companies – think hyperscalers, cloud providers, automotive firms, and edge-computing startups – who want tailor-made silicon. And that’s the market Intel wants to capture. The company’s advantage? It owns the entire stack: world-class manufacturing, packaging, and IP design resources, all under one roof. Unlike Broadcom, Marvell, or Alchip, Intel can provide an end-to-end ASIC service – from architecture to wafer production – within its own facilities. That’s a powerful differentiator in a world where most design houses still rely on TSMC or Samsung for fabrication.

This initiative also reflects a cultural shift within Intel. By consolidating engineering under CEG, the company has reduced the friction between design and production teams, potentially speeding up development cycles and improving yields. It’s a structure built to eliminate the silos that have slowed Intel in the past. If successful, it could restore the kind of agility that allowed the company to dominate the industry decades ago.

However, optimism alone won’t save Intel. Its product lineup still lags behind competitors in key areas, particularly AI. The much-anticipated Jaguar Shores rack-scale accelerator, designed to compete with NVIDIA’s GPU dominance, won’t arrive until 2027. Until then, Intel’s best bet is to monetize its engineering expertise through these new ASIC partnerships. The hope is that while others chase the AI gold rush with prebuilt chips, Intel can sell the tools, know-how, and silicon blueprints to those building their own AI ecosystems.

Lip-Bu Tan’s appointment makes this move especially interesting. Known for his time at Cadence, Tan is synonymous with IP licensing, design automation, and ecosystem building. His fingerprints are all over this new approach – Intel is essentially adopting the Cadence playbook but on a manufacturing scale. If Intel executes well, it could become what industry insiders call a “system foundry”: a company that doesn’t just fabricate chips but helps conceive, design, and optimize them from the ground up. That model could bring Intel a steady stream of design fees, production revenue, and long-term customer relationships.

Yet, challenges abound. Competing ASIC providers are becoming more efficient and more specialized. Broadcom and Marvell already dominate telecom and networking silicon, while startups are rapidly innovating in low-power AI hardware. Intel will need flawless execution to reclaim credibility in a market where “good enough” is no longer enough. Still, if the CEG-led strategy succeeds, it could be the company’s most meaningful reinvention in decades – turning Intel from a chipmaker fighting to stay relevant into a full-fledged silicon services powerhouse.

In short, this new ASIC strategy isn’t just another expansion. It’s Intel’s shot at rewriting its legacy – transforming from a struggling CPU giant into the ultimate partner for the next generation of custom computing.