news

한어Русский языкEnglishFrançaisIndonesianSanskrit日本語DeutschPortuguêsΕλληνικάespañolItalianoSuomalainenLatina

author | chai xuchen

editors: zhou zhiyu, huang yu

it seems difficult for nvidia to create another miracle. investors who are used to the high-growth days with "old huang" have smelled some danger signals, and they are no longer buying it this time.

on august 28, eastern time, after nvidia released its second quarter financial report for fiscal year 2025, its stock price briefly rose by 2.2% before plummeting, falling by 8% at one point. market anxiety was amplified, and investors began to sell off technology stocks, chip stocks, and ai concept stocks, causing u.s. stocks to fall across the board. before the market opened on august 29, nvidia's stock price still fell by more than 3.6%.

judging from the financial report, nvidia's performance data is not bad, with revenue of us$30 billion, a year-on-year increase of 122%; net profit increased by 168% to us$16.6 billion, both in line with market expectations, but there are no more surprises.

the "great leap forward"-style growth of this trillion-dollar giant has gradually died down.

over the past year, nvidia's revenue has doubled with the surge in data center business. however, the year-on-year growth rate of this main segment's revenue has dropped from more than 400% in the first two fiscal quarters to 154% in this quarter, and the quarter-on-quarter growth rate has also slowed down for four consecutive quarters; nvidia's total revenue growth has also resonated with the same frequency, ending with a 200% growth in this fiscal quarter.

the slowing trend may continue. for the third fiscal quarter, nvidia gave a revenue guidance of $32.5 billion, with the year-on-year growth rate falling to about 80%. this also means that its six consecutive quarters of "triple-digit" growth have come to an end.

previously, under the analysts' continued bullish expectations, "master huang" always brought bigger surprises.

westpark capital analyst kevin garrigan commented, "for investors with high expectations, nvidia's q3 revenue outlook being just $2 billion above the market average is not enough."

the market began to "demystify" nvidia, and investor sentiment rapidly diverged under heavy holdings, and the ai ​​bubble theory became rampant again.

as the leader of the ai ​​industry, nvidia's performance can be said to be a barometer of the entire ai industry. if the demand for its ai chips is strong, the ai ​​"wealth-making carnival" will continue.

originally, everyone still expected it to exceed expectations and lead the market out of volatility. a few weeks before the q2 earnings report was released, many institutions such as goldman sachs and morgan stanley, which were optimistic about nvidia, rushed to buy it, and about half of the world's top ten hedge funds also had heavy positions in nvidia.

the latest 13f documents disclosed by the sec show that as of the end of the second quarter of this year, global asset management giants such as jpmorgan chase, blackrock, and vanguard have all increased their holdings significantly.

a huge amount of funds poured in, and the market sentiment was also extremely tense. according to statistics from stockwe, a big data website for us stocks, before the financial report, nvidia's implied volatility was 10%, which means a large fluctuation of more than 300 billion us dollars.

but now that the cover has been turned over, nvidia's board of directors has approved an additional $50 billion in stock repurchases to appease shareholders.

but this still cannot make investors truly reassured, because at the earnings conference, "master huang" failed to answer the question that the market is most concerned about: can nvidia maintain its super-strong technological competitiveness and build a high "moat"; can the technology giant really make money from ai and continue this feast.

before the earnings conference, the industry already knew that the blackwell chip, which huang had high hopes for, had a "difficult birth". this may have an impact on major customers including meta, google and microsoft, because huang renxun once boasted at the q1 earnings conference that blackwell would contribute a lot of revenue to the company this year.

that evening, cfo colette kress acknowledged that there were production problems at blackwell and said the company was improving the design to increase production, and that shipments had been delayed until the fourth fiscal quarter.

huang renxun revealed that the demand for blackwell is extremely strong and will contribute "billions of dollars" in revenue in q4. as for whether these billions are incremental,"master huang" did not give a direct answer, which caused its stock price to fall during the conference call.

however, nvidia still has the support of many institutions such as goldman sachs, morgan stanley, bank of america, and deutsche bank. they believe that even if the delay of blackwell will cause short-term fluctuations, nvidia's solid fundamentals will act as the ballast for its stability in the upward channel, and the overall demand trend will not be overly affected.

wells fargo predicts that nvidia's data center revenue can rely on hopper chips (h100, h200) for a longer period of time: the current strong demand for h100, the start of shipment of h200 chips, and the company can launch a simplified blackwell version as a stopgap measure.

however, as the backlog of hopper orders is cleared by the end of the year, if blackwell fails to take over and contribute more growth, nvidia will face the situation of old rivals such as intel and amd catching up, and its "god-making" story will come to an end.

this is not the issue that investors are most worried about. after the "100 model wars", technology giants have become more rational, and the outside world hopes to see from nvidia's financial report whether customers can make real money from the ai ​​track.

in response, huang renxun still insisted that investment in computing infrastructure is the project with the best return on investment (roi) today. nvidia has just begun to reshape the global data center, and this is a trillion-dollar opportunity.

well-known investment bank wedbush also predicted that a $1 trillion ai spending "wave" is underway, and technology companies are still in the early stages of investing in ai hardware. analyst dan ives said, "for every $1 invested in nvidia gpus, the multiplier of technology stocks will increase by $8-10."

in other words, the deterministic opportunities of ai still remain at the infrastructure level. it is uncertain whether the subsequent capital expenditures of major customers such as cloud vendors can support the high growth of nvidia's data center business, which really tests investors' patience. next, if the commercialization of customers is delayed, nvidia may also be unable to escape the cycle.

in the world of technology, the only constant is change itself. standing at the top of the industry and being swept up by the trend, huang renxun, who has experienced many ups and downs, must start the battle to defend ai infrastructure.