Google’s AI Chip Initiative May Pose Greater Risks to Nvidia Than Expected

Key Points

• Google plans to reduce its artificial intelligence (AI) compute costs by up to 30% through its custom Tensor Processing Units (TPUs).

• The company is developing a neocloud business model with a projected 500 megawatts of TPU capacity to rent to other tech firms by 2027.

• Nvidia currently holds 86% of the AI data center market, but could face declining market share and profit margins as competition increases.

Alphabet (NASDAQ: GOOGL, GOOG) is pivoting to its own TPUs, which could signal a significant shift in the AI chip market traditionally dominated by Nvidia (NASDAQ: NVDA). Google aims to utilize its TPUs not just in-house but also for renting purposes, potentially disrupting Nvidia’s market stronghold.

With an anticipated capital expenditure of $190 billion this year, Google is serious about its plans for an expansive TPU capacity. The neocloud model could capture up to 20% of the AI cloud market by 2030, posing a substantial threat to Nvidia’s current 74% gross profit margin as tech companies explore custom processors for cost efficiency.