(Source: Financial News)
JPMorgan Chase指出,英伟达财报需要关注四大变量。国内H200芯片获批出货有望上调营收指引。市场更关注毛利率指引以及成本转嫁信号。
On August 26, according to overseas financial news, Nvidia will announce its second-quarter (for July 5–7) results on the 26th local time. Compared with the financial report data itself, the market is paying even more attention to the signals conveyed during the earnings conference call.
Data from prediction platform Polymarket shows that, as of the 25th, the probability that Nvidia’s adjusted earnings per share (EPS) for the second quarter exceeded expectations is 96%; the probability that the gross margin falls in the 74%–76% range is 93%; and the probability that revenue for the data center segment surpasses $80 billion (approximately 111 trillion KRW) is as high as 95%.
JPMorgan analyst Harlan Cell said that, in addition to revenue guidance, Nvidia’s earnings report should focus on four key variables: 1) its plan for emerging competitors in inference-focused segments such as ASIC custom chips; 2) how to define the nature of the $500 billion AI infrastructure financing plan agreed with Wall Street; 3) the initial shipment progress of H200 chips for the China market; 4) whether the specifications of the Rubin and Bella next-generation chips are adjusted, given tight memory supply. It is reported that some domestic customers have already received approved H200 chips, which could become a positive factor for upward revisions to revenue guidance in the fourth quarter.
The dispute over round-based financing is also a key area of observation. Nvidia has provided the highest payment guarantee of up to $105 billion for a 20-year lease contract for OpenAI’s Ohio data center, and along with Apollo Global and others launched a third-party capital plan with a total scale of $500 billion. Brian Malvy, CEO of Jax Investment Management, warned of the risks of being highly dependent on AI business. CEO Jensen Huang rebutted that this project is not round-based financing, but a long-term investment.
Morgan Stanley estimates that only the Rubin chips will generate $9 billion in revenue in the third quarter. UBS analyst Timothy Akury noted that a shortage of memory supply may ultimately benefit Nvidia, and the meeting could bring news of a significant increase in order backlog.
For Korean memory makers such as Samsung Electronics and SK Hynix, the signal conveyed by the gross margin outlook is more valuable than revenue and profit. If Nvidia itself absorbs cost pressure and lowers its gross margin forecast, it would indirectly confirm an improvement in the memory suppliers’ bargaining power. If Nvidia maintains the original gross margin of around 75%, it would indicate that cost pressure is being passed on to downstream customers. The market cares more about gross margin guidance and commentary on storage raw-material costs than revenue or EPS.