Partnering with Google to develop custom chips + $12.2 billion potential stake! Marvell Technology (MRVL.US) surged in pre-market trading

According to Zhitong Finance APP, Marvell Technology (MRVL.US) has reached a custom chip cooperation with Google (GOOGL.US) and issued a warrant to Google allowing the latter to purchase approximately $12.18 billion worth of Marvell Technology shares. Boosted by this news, as of the time of writing, Marvell Technology's stock surged nearly 13% in pre-market trading on Wednesday.

In recent years, as companies seek alternatives to NVIDIA's (NVDA.US) expensive graphics processing units (GPUs), the demand for custom chips like Google's tensor processing units (TPUs) for AI workloads has significantly increased. According to a statement released by both Marvell Technology and Google, the two companies will collaborate to develop AI inference accelerators, storage controllers, network interface controllers, memory interface controllers, and near-memory computing technologies.

Additionally, Marvell Technology stated that it has issued a warrant to Google for the purchase of up to 58.97 million shares, with an exercise price set at $206.58 per share. The company signed a commercial agreement with Google on July 29 to provide custom chip products. In the first year following the agreement, 1,360,867 warrants will vest each quarter. The remaining shares will vest based on the "self-purchase" situation from Marvell Technology's third quarter of fiscal year 2027 to the end of fiscal year 2033, with a batch of shares vesting each time the jointly developed products generate $500 million in revenue.

Meanwhile, Marvell Technology's larger competitor Broadcom (AVGO.US) fell over 3% in pre-market trading. For a long time, Google has primarily relied on Broadcom to assist in designing its self-developed TPUs. Both Marvell Technology and Broadcom are capable of helping clients transform chip designs into chips that can be practically produced and provide backend support before entering the wafer manufacturing stage, ultimately handing over production to foundries like TSMC. As large tech companies accelerate the development of self-developed AI chips, this business is becoming an important growth engine for both companies