Could Intel be the next AI play? And what to do with Nvidia?

Why Intel Might Be a Buy — The Hidden AI-Play

• According to a recent report, Intel is “probable” to get orders from Google for advanced packaging services for Google’s next-gen TPUs. (Wccftech)
• The core of this is Intel’s proprietary packaging technologies (like EMIB / Foveros) — something the report positions as a U.S.-based alternative to foreign supply-chain dials, which is becoming more valuable given geopolitical pressure on chip supply chains. (Wccftech)
• If that deal materializes (and potentially more from other big-tech players), it could ignite a real resurgence for Intel’s foundry & packaging business — a business-line that’s underappreciated today, partly because past earnings have reflected legacy CPU weakness rather than future AI infrastructure upside.

Thus, Intel seems to be in a classic “underpriced optionality” phase: the core business may look pedestrian — but there’s latent upside if the AI-infrastructure trend swings in its direction.

Why NVIDIA Could Be a Short (or at Least “Richly Valued Relative to Risk”)

• NVIDIA’s current P/E ratio (TTM or trailing) remains elevated relative to historical norms for a broad semiconductor benchmark. (World PE Ratio)
• As reported, some valuation analyses argue that NVIDIA commands a premium comparable to 35× the valuation of Intel — a gap that seems hard to justify if rivals like Intel start gaining share in AI-chip supply chain components. (Benzinga)
• Meanwhile, the AI-deployment landscape may be shifting. With alternatives (like Google’s TPUs) gaining traction, NVIDIA may face increasing competition — which could compress future expectations baked into its stock price. Indeed, some industry observers consider TPU-style accelerators from Big Tech a genuine challenge to GPU-first AI infrastructure dominance. (Wccftech)

In short: if growth disappoints, or if alternative AI-hardware platforms gain traction faster than expected, NVIDIA’s premium could feel fragile — making it vulnerable to a re-rating.

Why This Could Make a Compelling Pair Trade — Long Intel / Short NVIDIA

• Valuation divergence: Intel looks relatively cheap (and arguably undervalued) given its potential upside; NVIDIA looks expensive if you assume pressure on its dominant position.
• Asymmetric upside vs downside: Intel’s potential gains — from securing high-profile AI-infrastructure contracts — could be substantial if just a few deals go through. Meanwhile, the downside for NVIDIA (a re-rating or loss of dominance) would likely be sharp given its high multiple.
• Sector-level rebalancing: If AI-hardware supply chains begin to diversify, capital may flow from “pure GPU” plays to more diversified or infrastructural chip firms — a structural shift that favors companies like Intel.

Hence, a trade that is long Intel and short NVIDIA might capture both valuation reversion (NVIDIA) and optionality (Intel) — offering a contrarian but potentially lucrative angle.

Risks & What Could Go Wrong

• The deal between Intel and Google is still rumored, not confirmed. If it doesn’t materialize, Intel’s “AI upside” might remain just that — upside.
• NVIDIA still dominates many segments of AI infrastructure. Their software ecosystem, customer relationships, and performance advantages are nontrivial barriers to displace — so NVIDIA might still deliver strong growth and justify its premium.
• Macro risks (e.g., broader semiconductor slowdown, AI infrastructure demand weakening, capital markets repricing) could hit both companies — undermining the pair-trade thesis.

Conclusion: Watch Intel — With a Contrarian Microscope

If I were investing on conviction rather than consensus, I’d be watching Intel — because the pieces seem to be quietly aligning for a rebound tied to the AI infrastructure wave. Meanwhile, the premium on NVIDIA feels increasingly like a crowded long — especially as alternative architectures start gaining legitimacy. As a result, the long-Intel / short-NVIDIA pair trade might offer a compelling way to hedge macro risk while expressing a view on structural change in AI hardware.