Snowflake soars 30% on earnings beat and plan to spend $6 billion on Amazon cloud

Amazon Web Services secured a $6 billion, five-year spending commitment from Snowflake, reinforcing AWS momentum in AI and custom silicon adoption. Snowflake also beat fiscal Q1 expectations with adjusted EPS of $0.39 on $1.39 billion revenue versus $0.32 and $1.32 billion consensus, and guided to a 12.5% fiscal Q2 adjusted operating margin on $1.415 billion to $1.420 billion in product revenue. Shares rose as much as 30% after hours on the earnings beat and the strategic AWS deal.

Analysis

This is more than a customer win for AWS; it is evidence that the economics of AI infrastructure are shifting from GPU-only thinking to a hybrid stack where CPU efficiency matters just as much as accelerator supply. The second-order implication is that Amazon is monetizing not just compute demand but architectural migration toward its own silicon, which should improve AWS gross margin mix over time as Graviton penetration rises and reduces dependence on higher-cost third-party x86 capacity.

For Snowflake, the deal reinforces that large enterprise AI workloads are becoming an infrastructure-arbitrage game: buyers will commit long-dated spend to secure capacity, performance, and unit economics before agentic use cases scale. That helps de-risk Snowflake’s growth story for the next 12-24 months, but it also signals a heavier reliance on cloud partners for margin leverage, so any delay in AI workload adoption would be more visible in bookings quality than in near-term revenue.

The market is likely underestimating the competitive read-through for Intel and, to a lesser extent, AMD. If agentic workflows are structurally CPU-intensive and hyperscalers keep steering volume to Arm-based custom silicon, the incremental demand pool that traditionally would have supported x86 refresh cycles may never fully materialize. Nvidia remains insulated near term because training and premium inference still anchor the GPU cycle, but this deal suggests the next leg of AI capex growth may be less GPU-exclusive than consensus assumes.

The contrarian takeaway is that the headline is bullish for AMZN, but not because of near-term revenue upside; the real upside is strategic lock-in and margin expansion over multiple years. Snowflake’s strong print may already have pulled forward some optimism, so the cleaner expression is not chasing SNOW after a gap higher, but owning the infra beneficiaries with the deepest operating leverage and shorting the laggards most exposed to custom silicon substitution.

Key Decisions for Investors

• Long AMZN vs. short INTC for a 3-6 month horizon: AWS gets credit for AI platform share and silicon monetization, while Intel faces the cleanest risk from continued migration away from x86; attractive if the pair stays anchored to margin and adoption rather than multiple expansion.
• Add to AMZN on any post-event consolidation over the next 2-4 weeks: the catalyst is not the contract value itself but evidence of durable AWS share gains in AI; upside is multi-year operating leverage, downside is limited unless AI capex pauses materially.
• Avoid chasing SNOW after the print; instead consider a short-dated put spread if the stock re-rates above post-earnings levels in the next 1-2 weeks: good fundamentals, but the market may be overpaying for a deal that is mostly validation rather than incremental surprise.
• Long ARM as a second-order beneficiary over 6-12 months: if hyperscalers keep standardizing on Arm-based custom chips, ARM’s royalty and ecosystem leverage can compound even without direct exposure to any single cloud contract.
• Tactical underweight NVDA relative to AMZN/META over the next quarter: not a bearish call on AI, but a view that the next wave of AI spend may tilt toward CPUs and orchestration, compressing the relative scarcity premium in GPUs.