Alphabet and Broadcom are capitalizing on the booming AI market through their co-developed Tensor Processing Units (TPUs). These custom AI chips are driving significant growth for both companies, with Alphabet leveraging them for its cloud services and AI models, and Broadcom benefiting from increased chip sales.
Alphabet’s TPUs offer cost advantages over competitors, boosting its cloud revenue and enabling the development of its Gemini models. Broadcom, as the physical producer of these chips, is set to see massive revenue growth, especially as other hyperscalers look to diversify their AI hardware suppliers away from Nvidia.
The artificial intelligence (AI) market is experiencing explosive growth, with significant advancements in both AI infrastructure and cloud computing. Two key players at the forefront of this revolution are Alphabet ((GOOGL 1.03%) (GOOG 1.04%)) and Broadcom ((AVGO +2.94%)), companies that have collaborated to develop Alphabet's highly acclaimed Tensor Processing Units (TPUs).
These TPUs serve as a crucial link between Alphabet and Broadcom, underpinning their recent successes. Co-developed over a decade ago, these custom AI application-specific integrated circuits (ASICs) are now the backbone of Alphabet's cloud platform. Alphabet spearheads the chip design, while Broadcom contributes its proprietary intellectual property in high-speed Serializer/Deserializer (SerDes) technology, alongside managing the physical design and complex integration of packaging.
Let's delve into why these two AI stocks are exceptionally positioned for sustained long-term growth.
NASDAQ: GOOGL
Key Data Points
Alphabet: The TPU Advantage
Alphabet's proprietary TPUs provide a significant competitive edge in the AI landscape. The company has leveraged these chips for its internal operations for over a decade, meticulously optimizing its entire hardware and software ecosystem around them. This strategic focus has given Alphabet a substantial head start in the custom AI chip market. Its TPUs are now in their eighth generation, having undergone extensive real-world testing and refinement.
Possessing its own cutting-edge AI chips allows Alphabet to achieve superior economics within its cloud division. It can train its advanced AI models and execute inference tasks at a considerably lower cost compared to rivals predominantly reliant on Nvidia's graphics processing units (GPUs). In the most recent quarter, Alphabet's cloud computing revenue soared by 63% to $20 billion, with operating income tripling to $6 billion. By the close of Q1 2026, the company reported an impressive $460 billion Google Cloud backlog. Furthermore, the high regard for its TPUs has led Alphabet to enable select customers to procure them through Broadcom for both internal and external use within Google Cloud, creating an additional high-margin revenue stream.
Alphabet also utilizes its TPUs for training its flagship Gemini frontier models and for running inference. This dual capability not only reduces operational costs but also strategically positions the company in the consumer AI market. By leveraging its vast distribution network and established advertising ecosystem, Alphabet can effectively drive growth. Through ownership of dominant platforms like the Chrome browser and the Android operating system, coupled with a search revenue-sharing agreement with Apple, Google serves as a primary gateway to the internet. The company has adeptly integrated Gemini into its search functions and other products, fostering profitable expansion.
As the sole entity boasting both leading-edge AI chips and a world-class frontier model, Alphabet is exceptionally well-positioned to be a long-term victor in the AI sector.
Image source: The Motley Fool.
Broadcom: The ASIC Leader
NASDAQ: AVGO
Key Data Points
The TPUs are also a significant catalyst for Broadcom's growth. As the direct supplier of these chips, Broadcom is the entity that records the revenue from their sales. Alphabet's substantial investments in AI infrastructure are channeling considerable funds into TPU development and procurement. Furthermore, Anthropic has committed to a massive $21 billion TPU order with Broadcom for the current year, with plans to acquire an additional 3.5 gigawatts of computing capacity in the future.
The success of TPUs has prompted other hyperscalers to engage Broadcom for the development of their own custom AI chips. The company anticipates generating over $100 billion in ASIC revenue by fiscal year 2027. Citigroup has projected that Broadcom could produce $180 billion in AI revenue by fiscal year 2028, a dramatic increase from its less than $64 billion in total revenue last year, of which approximately $20 billion was attributed to AI.
As hyperscalers increasingly seek to diversify their supply chains away from Nvidia, Broadcom is strategically positioned to be a major beneficiary. With substantial growth on the horizon, the stock presents a compelling investment opportunity.
