Wall Street is abuzz about the "AI closed loop": Bulls say "suppress ASICs, and Nvidia’s bull run will last," while bears say "making loans to customers is just like Cisco back in the day."

``` Nvidia’s $100 billion investment deal with OpenAI is sparking intense debate on Wall Street. On September 24, according to details reported by Reuters, the structure of the deal is that Nvidia will invest up to $100 billion in OpenAI in exchange for non-voting shares, and OpenAI will “use this money” to buy Nvidia’s chips, with plans to deploy at least 10 GW of Nvidia systems. This “vendor financing” model—giving money to customers, who then buy back your own products—is pushing up AI sector stock prices, but is also making some seasoned market participants uneasy. They warn that this model is reminiscent of some companies’ practices before the dot-com bubble burst in 2000: “lending money to customers, just like Cisco back then,” which could hide huge risks. However, bulls believe this move is a strategic step for Nvidia to consolidate its dominance in the GPU market and suppress competition from ASICs (Application-Specific Integrated Circuits). History Echoes: Replay of Cisco’s “Vendor Financing”? For the bears, Nvidia’s deal is filled with a sense of déjà vu. Rich Privorotsky, Head of Goldman Sachs Delta One Trading, called this deal “circular referencing?” in his morning report, and bluntly stated that this was a major feature of the tech bubble era. He recalled that back then, telecom equipment manufacturers like Cisco and Lucent offered loans or equity investments to their customers, who then used the funds to repurchase equipment. History has proven that the outcome of this model was “not very good” for any of the participants. “…Vendor financing was a feature of that (dot-com bubble) era, when telecom equipment makers (Cisco, Lucent, Nortel, etc.) offered customers loans, equity investments, or credit guarantees, and then customers used the cash/credit to buy back equipment… Suffice to say, it did not end well for anyone.” J.P. Morgan technology trader Meyer expressed a similar view, pointing out: “When a company pays customers to buy its goods, it’s usually not a good sign.” The Australian Financial Review also commented that “this is very concerning”: “A company invests $100 billion in another company so it can purchase $100 billion worth of chips made by the ‘financier’. Welcome to the circular economy of artificial intelligence.” Critics argue that this is tantamount to accounting magic, and although the seriousness is not as great as the Enron incident, with growing attention, the market may be approaching a tipping point. Strategic Move: The Long-Term Narrative for GPU Supremacy Despite ongoing skepticism, market bulls have offered a totally different take from a strategic perspective. They believe this is not just a simple financial maneuver, but a critical step for Nvidia to consolidate its dominance. J.P. Morgan’s Meyer, explaining the bullish perspective, said that this move is essentially no different in principle from Nvidia’s previous investments in companies such as CRWV/Lambda. The deal sends a strong signal to the market: “If you want chips, you must order now, or you may not get any supply.” A deeper interpretation is that this deal is seen as OpenAI openly siding with the GPU technology route. Bulls argue that this means OpenAI will reduce or abandon the use of customized ASIC chips, thereby consolidating GPUs’ position as the “winner” in the AI race, adding strong support to Nvidia’s long-term growth narrative. Staggering Energy Bill: One Project Needs Ten Nuclear Reactors Apart from the structure of the deal itself, the massive size of the project has shocked the market. According to the plan, OpenAI will deploy at least 10 gigawatts (GW) of Nvidia systems. Behind this figure is staggering energy consumption. Analyst Lawrence McDonald pointed out that the average capacity of a U.S. nuclear power plant unit is about 1 GW, with 94 reactors in total providing a capacity of 97 GW. This means that just the electricity needed for this single project would equal the total output capacity of ten nuclear reactors. This comparison highlights the huge challenge AI infrastructure poses to global energy supply, forcing investors to consider energy costs and infrastructure feasibility as key variables in assessing the future development of the AI industry. Risk warning and disclaimer The market involves risks; investment requires caution. This article does not constitute individual investment advice, nor does it take into account the special investment objectives, financial situation, or needs of individual users. Users should consider whether any opinions, viewpoints, or conclusions contained herein are suitable for their specific situation. Invest accordingly at your own risk. ```