Can Nvidia beat sky-high expectations as AI worries pile up?
AI chip firm Nvidia is set to announce its quarterly numbers on 17 November against perhaps the most difficult backdrop it has endured since the launch of ChatGPT lit the touch paper on the whole artificial intelligence story back in 2022.
The company has faced significant hits to sentiment during that period, not least the launch of the DeepSeek model in China early last year which suggested AI models could be trained at a much lower cost.
Facing multifaced challenges
It shook that challenge off, but the difficulties facing the company now look more multifaceted.
The broader sector is contending with fears the Democratic blue wave in the US midterm elections might prompt a backlash against data centres. Concern is mounting about the huge costs involved in the artificial intelligence push and now governance worries can be added to the list.
The board has launched an independent investigation into CEO Jensen Huang and whether he has been targeted in a foreign espionage effort. For now, the company says there is no evidence that company secrets or US national security have been compromised.
As its market value has grown, the pressure on Nvidia to not only match but beat expectations with its results and guidance has ramped up. In the current circumstances that pressure is likely to be even greater. Even if, on a price to earnings basis, Nvidia has not looked this cheap for more than a decade.
Investors will be looking for evidence that surging data centre sales, which topped $89 billion in the second quarter, are retaining their momentum, with the minimum expectation being that Nvidia hits consensus forecasts for $109 billion revenue and $2.49 in earnings per share.
An increasing number of participants in the AI race are producing their own custom chips, usually to accompany rather than replace Nvidia’s GPUs but the market will want reassurance on this potential competitive threat. Plus, attention will be drawn to adoption of Nvidia’s next generation Vera Rubin architecture.
Is Nvidia being paid on time?
Pressure on margins from the growing costs of components is another factor to watch and given doubts about the sustainability of AI spending, days sales outstanding (DSO) will also be worth monitoring.
This metric, which measures the average time it takes to collect payment after a sale, has been creeping higher. Moving from year 45 in the first quarter of the current financial yar to 59 in the second quarter.
