Everyone wants to learn from NVIDIA Corporation's "chip swap financing." Alphabet Inc. Class C (GOOGL.US) and AMD (AMD.US) both want to support "AI cloud."

NVIDIA Corporation (NVDA.US)s computing power - finance closed loop constructed with the support of CoreWeave is really tempting, forcing Alphabet Inc. Class C (GOOGL.US) and AMD (AMD.US) to feel their way across the river and attempt to establish their own AI chip ecosystem. On February 20, according to media reports citing informed sources, Alphabet Inc. Class C is exploring the use of its strong financial resources to build a broader AI ecosystem, with the core strategy being to provide funding support to data center partners on the condition that they must use Alphabet Inc. Class C's TPU chips. At the same time, Huashijianwen wrote that AMD is also aggressively promoting its AI chips by providing guarantees for loans to customers. This strategy is seen by the market as a direct replica of NVIDIA Corporation's "CoreWeave model" - that is, by supporting "emerging clouds" manufacturers, bypassing traditional cloud giants that have already been occupied or developed their own chips by NVIDIA Corporation, and establishing their own main force. Following the NVIDIA Corporation route: injecting funds into "new cloud" and mining enterprises Sources revealed that Alphabet Inc. Class C is in talks to invest about $100 million in cloud computing start-up Fluidstack. This transaction values Fluidstack at about $7.5 billion. This is not just a financial investment, but also a strategic binding. Companies like Fluidstack are important variables in the current AI computing market, as they specialize in providing computing services for AI companies. Previously, NVIDIA Corporation was able to tear open a market by supporting CoreWeave, which hoarded a large number of GPUs. Now Alphabet Inc. Class C also wants to do as the Romans do. Sources said that Alphabet Inc. Class C hopes to "help amplify Fluidstack's growth potential and encourage more computing providers to use its AI chips" by injecting funds. In addition to direct funding, Alphabet Inc. Class C is also reaching out to former cryptocurrency mining companies in transition. Reports indicate that Alphabet Inc. Class C has provided financing support (backstopped financing) for projects involving companies such as Hut 8, Cipher Mining, and TeraWulf, which have existing data center infrastructure and are eager to transition into AI computing factories. The funding support provided by Alphabet Inc. Class C will result in their acceptance of TPU chips. AMD's aggressive "bet": if I can't sell it, I'll rent it out Compared to Alphabet Inc. Class C's direct funding, AMD's approach is more aggressive and risk-seeking. According to reports, AMD will provide substantial guarantees for a $300 million loan to data center start-up Crusoe. This funding from Goldman Sachs Group, Inc. will be used to purchase AMD's AI chips. The most notable aspect is the "safety net clause." Sources revealed that if Crusoe cannot find customers to use these chips, AMD has agreed to rent the chips from Crusoe. This means that AMD becomes the "last tenant," eliminating concerns on the demand side for customers. While this approach can boost sales in the short term, it also exposes chip manufacturers to significant risks when AI demand slows down. Once the market cools down, AMD's balance sheet will directly bear the impact. Why take the detour? Alphabet Inc. Class C and AMD have chosen this detour because the main road has been blocked. For Alphabet Inc. Class C, although star start-ups such as Anthropic are increasing the use of TPUs, traditional cloud competitors (such as Amazon.com, Inc. AWS, Microsoft Corporation Azure) show little interest in TPUs. Industry insiders pointed out: "The interest of major cloud service providers seems lukewarm, in part because they see Alphabet Inc. Class C as a competitor." In addition, Amazon.com, Inc. AWS itself is actively developing its own AI chips. Therefore, supporting neutral third-party "emerging clouds" has become the best choice for Alphabet Inc. Class C and AMD to break the blockade. Internal power struggle and production bottlenecks In order to accelerate the commercialization of TPU, Alphabet Inc. Class C has even considered restructuring internally. According to sources, some senior management in Alphabet Inc. Class C's cloud division recently reopened a long-standing internal debate: "Whether to reorganize the TPU team into an independent department." This split plan may allow Alphabet Inc. Class C to bring in external capital to expand investment opportunities. However, this proposal was denied by official representatives of Alphabet Inc. Class C. An Alphabet Inc. Class C representative explicitly stated that "there are no plans to restructure the TPU department" and emphasized that "maintaining close integration between the chip team and other departments of the company (such as the Gemini model development team) has advantages". In addition to architectural issues, the more realistic obstacle is production capacity. Although Alphabet is currently working with Broadcom Inc. (Broadcom) to design TPUs and outsourced to Taiwan Semiconductor Manufacturing Co., Ltd. Sponsored ADR for production, in the current global surge in AI demand, Taiwan Semiconductor Manufacturing Co., Ltd. Sponsored ADR's advanced production capacity is struggling. Sources said, "Taiwan Semiconductor Manufacturing Co., Ltd. Sponsored ADR may prioritize its largest customer NVIDIA Corporation, rather than Alphabet Inc. Class C." In addition, the global shortage of essential HBM memory chips for AI chips is also a hard problem that Alphabet Inc. Class C must face in expanding TPU shipments. This article is reprinted from Wall Street News, GMTEight editor: Zhang Jinliang.