The AI boom is creating significant opportunities for companies like Alphabet and Broadcom. Their strategic collaboration on Tensor Processing Units (TPUs) has given them a competitive edge in the rapidly evolving AI market.
Alphabet benefits from cost-effective AI model training and inference, while Broadcom secures substantial revenue from manufacturing these advanced chips, positioning both companies as key players in the AI revolution.
The artificial intelligence (AI) market is experiencing explosive growth, benefiting both AI infrastructure and cloud computing providers. Two key players leading this charge are Alphabet (GOOGL, GOOG) and Broadcom (AVGO), who have a history of co-development, notably with Alphabet's Tensor Processing Units (TPUs).
These TPUs, custom AI application-specific integrated circuits (ASICs) developed over a decade ago by Alphabet and Broadcom, form the backbone of Alphabet's cloud platform. Alphabet handles chip design, while Broadcom contributes crucial intellectual property in high-speed Serializer/Deserializer (SerDes) and manages the intricate physical design and packaging.
Alphabet: The TPU Advantage
Alphabet's proprietary TPUs provide a significant edge in the AI landscape. Having utilized these chips for its internal operations for over a decade, the company has meticulously optimized its hardware and software around them. This strategic focus has positioned Alphabet at the forefront of custom AI chip development, with its TPUs now in their eighth generation and thoroughly proven.
This in-house capability offers Alphabet superior economics for its cloud division. It allows the company to train AI models and perform inference tasks at a considerably lower cost compared to rivals predominantly relying on Nvidia's GPUs. Last quarter, Alphabet's Google Cloud revenue soared by 63% to $20 billion, with operating income tripling to $6 billion. Furthermore, Alphabet is now enabling select customers to procure TPUs through Broadcom for external use, opening a new high-margin revenue stream.
Alphabet leverages its TPUs to train its advanced Gemini models and execute inference, reducing costs and strengthening its position in the consumer AI market. By integrating Gemini with its dominant search engine, Chrome browser, and Android operating system, and benefiting from its search revenue-sharing deal with Apple, Google effectively controls internet access. This allows for profitable growth through AI integration across its vast product ecosystem.
As the sole entity possessing both leading AI chips and a premier frontier model, Alphabet is exceptionally well-positioned for sustained success in the AI sector.
Broadcom: The ASIC Leader
TPUs are also a major growth catalyst for Broadcom. As the manufacturer and supplier of these chips, Broadcom directly benefits from the revenue generated from their sales. Alphabet's substantial investments in AI infrastructure translate into significant orders for TPUs. Additionally, Anthropic has placed a massive $21 billion order for TPUs with Broadcom this year, with further commitments for future computing capacity.
The success of TPUs has prompted other hyperscalers to partner with Broadcom for the development of their own custom AI chips. Broadcom anticipates generating over $100 billion in ASIC revenue by fiscal year 2027, with a CitiGroup projection of $180 billion in AI revenue by fiscal year 2028. This represents a dramatic increase from its previous year's total revenue of less than $64 billion, of which only $20 billion was AI-related.
As hyperscalers seek to diversify their supply chains beyond Nvidia, Broadcom is poised to be a primary beneficiary. With significant future growth anticipated, Broadcom's stock presents a compelling investment opportunity.
