Why Alphabet Stock Popped on Wednesday

Alphabet shares rose 2.8% after Piper Sandler said TSMC, Google's key TPU contract manufacturer, is sold out for advanced AI-chip capacity through H1 2028. Piper forecasts combined GPU and ASIC shipments of 14 million units this year, followed by 50% growth in 2027 and 40% in 2028, with ASICs—including Google's TPUs—accounting for 66% of shipments. Alphabet's TPU sales began generating revenue last quarter, while the company trades at 16.8x trailing earnings despite consensus expectations for 14.6% annual earnings growth over five years.

Analysis

The relevant signal is not a near-term revenue step-up for GOOG, but a potential shift in bargaining power toward leading-edge foundry capacity. If the reported allocation tightness is confirmed in TSMC disclosures, TSM can monetize demand through mix, pricing, and customer prepayments while hyperscalers absorb higher chip costs to protect AI service availability. For GOOG, captive TPU supply is strategically valuable, but constrained wafer allocation can also cap the pace at which it converts AI demand into Cloud revenue or external chip sales.

The market may be over-attributing this to Alphabet rather than to the semiconductor bottleneck. Custom accelerators can pressure merchant-GPU pricing at inference workloads over a 6-18 month horizon, creating a relative headwind for NVDA and AMD multiples if ASIC mix expands faster than expected; however, this does not imply an immediate volume decline because aggregate compute demand remains the dominant driver. The key missing datapoints are whether "sold out" refers to advanced wafers, advanced packaging, or specific customer allocations, and whether Google has binding volume commitments rather than indicative reservations.

Over the next 1-3 months, earnings commentary on GOOG capex, Cloud margin, TPU external revenue, and TSM’s advanced-node utilization should determine whether this becomes investable rather than headline-driven. Falsifiers are a reduction in hyperscaler capex guidance, evidence that capacity additions ease earlier than expected, or TPU economics failing to improve Google Cloud margins; any of these would unwind the scarcity premium. A larger tail risk for TSM remains Taiwan-related disruption, which means capacity scarcity can widen valuation dispersion but does not eliminate geopolitical discount risk.

Key Decisions for Investors

• Accumulate TSM on market weakness rather than chase the headline; target a 6-12 month long with an initial 5-7% risk budget. Upside comes from sustained advanced-node pricing/mix and stronger customer commitments; exit or reassess if TSM signals utilization normalization or accelerated capacity relief.
• Maintain GOOG as a selective 6-12 month long only after the next earnings report confirms that TPU deployment improves Cloud margin or produces measurable third-party revenue. A long initiated solely on supply-tightness headlines has weak risk/reward because incremental capacity can raise capex before it raises monetization.
• Establish a small 6-18 month relative-value basket: long TSM versus short a diversified merchant-accelerator exposure through SMH only if ASIC shipment growth and TPU external adoption are independently corroborated. Keep sizing modest: broad AI demand can still lift both legs, while an NVDA/AMD pricing response or superior GPU roadmap would invalidate the substitution thesis.
• Set an event alert for GOOG capex guidance and TSM advanced-packaging commentary at the next quarterly results. If Google raises capex without concurrent Cloud-margin improvement, reduce GOOG exposure; that outcome indicates supply spending is being capitalized without evidence of economic return.