Can NVIDIA Corporation's financial report burst the "AI bubble theory"? Citigroup gave a "buy" rating, stating that demand for AI far exceeds supply.

Citigroup released a research report stating that it maintains a "buy" rating on NVIDIA Corporation (NVDA.US) and has raised the target price from $210 to $220. In addition, Citigroup has initiated a "30-day short-term bullish" view on the stock, betting that its upcoming earnings report to be released on November 19 will show strong performance with "revenue exceeding expectations and guidance raised". Analysts Atif Malik and Papa Sylla pointed out in the report that despite doubts about the source of AI investment funds, a more fundamental fact is that due to limited capacity in advanced packaging (CoWoS), supply of AI chips will continue to be lower than demand until 2026. Citigroup expects that NVIDIA Corporation's upcoming performance will easily surpass the general expectations on Wall Street. The report predicts that the company's sales for the quarter ending in October will reach $57 billion, higher than the market average expectation of around $55 billion. Looking ahead, Citigroup predicts that NVIDIA Corporation's sales guidance for the January quarter will reach $62 billion, also higher than the market's expectation of around $61 billion. The report points out that behind this optimistic forecast is the strong momentum of NVIDIA Corporation's Blackwell architecture GPUs. Citigroup analysts believe that the information revealed by NVIDIA Corporation at the GTC Washington conference that 6 million GPUs have been shipped is a strong signal that its performance for the October and January quarters may exceed expectations. Citigroup stated in the report that its model assumes that NVIDIA Corporation's data center sales will increase by 24% and 12% month-on-month in the October and January quarters respectively, while the market expects increases of 19% and 15% respectively. In response to the escalating "AI bubble" rhetoric in the market, Citigroup has put forward a completely opposite view. The main argument of the firm is that the major contradiction in the current AI chip market is supply exceeding demand, rather than insufficient demand. The firm stated, "Despite concerns about the debt and circular financing portfolios surrounding the AI capital expenditure bubble, we fundamentally see that due to CoWoS capacity constraints, AI supply will remain below demand until 2026 and may only catch up sometime in 2027."