|
Listen to this story
|
NVIDIA’S recent quarterly report presents a curious picture of the company — a unique combination of record profit and loss. In this earnings call for the last quarter, the company reported a loss of $1.83 billion in gaming, which has long been its primary focus market. However, even as their gaming section saw a downturn, the company’s revenue rose to a record $15.01 billion in the data centre sector, representing a growth of 41% over the past year.
While NVIDIA has been dedicating resources towards developing new solutions for their enterprise customers, it seems like gaming is taking a back seat. Last year alone, it released 7 new products for the enterprise while releasing only the RTX 40 series for the gaming market. However, according to NVIDIA chief Jensen Huang, gaming is recovering from the post-pandemic downturn, with gamers enthusiastically embracing the new Ada architecture GPUs. Compared to the last quarter the company has seen an increase of 16 percent in the gaming sector.
Gaming reality
Gaming has been NVIDIA’s bread and butter since inception. As per the information from the January 2023 Steam Hardware Survey, NVIDIA reigns supreme in the field. However, there is a twist in the story – users are still on old hardware.
The green giant boasts 75.03% market share in PC gaming, but the list of most-used GPUs has a telling lack of the latest 40-series GPUs. The gaming community ditched the 40-series GPUs, mainly due to pricing issues and miscommunications on the capabilities of the chips. Most gamers are still using generations-old technology, with the top 5 dominated by 10-series and 20-series cards. The closest to a ‘latest GPU’ on the line comes in at number 3, a position taken by the RTX 3060 laptop GPU.
NVIDIA’s last big win in gaming was the launch of ray-tracing enabled GPUs in 2018, termed the ‘RTX’ line of GPUs. When combined with accompanying technologies like DLSS 3.0 and Tensor cores, these chips are currently the de-facto standard for high-fidelity gaming. However, the company’s downturn in the field may represent a saturation in the gaming market, where gamers are unwilling to upgrade for minimal improvements.
The upgrade cycle for consumer GPUs hovers around 3-4 years, with the mean lifecycle for enterprise components pegged around 5.5 years. NVIDIA’s gaming GPUs are usually released in a 2-year cycle, but the pace of research and developments in the enterprise sector has crept up in the past few years. NVIDIA still caters to the consumer upgrade cycle, but it has circumvented the enterprise upgrade cycle by catering to cloud service providers.
“Our strategy and goal is to put the DGX infrastructure in the cloud. By having NVIDIA DGX and NVIDIA’s infrastructure full stack in their cloud, we’re effectively attracting customers to the CSPs. This is a very, very exciting model for them…we’re going to be the best AI salespeople for the world’s clouds,” said Jensen Huang, the CEO of NVIDIA.
Gaming revenue used to be neck-and-neck with enterprise revenue at NVIDIA, but their enterprise pivot has now come into focus. Along with the lacklustre 40-series launch, it signed an agreement with Microsoft to bring Xbox games to GeForce NOW, their cloud gaming platform. The company seems to have shifted towards AI hardware where they stand to gain a lot from doubling down on their already-pervasive industry presence.
Why enterprise?
NVIDIA’s release schedule has been dominated by enterprise products for the last year, such as the Hopper architecture and DGX H100 systems. From its CUDA software stack to its partnership with major cloud service providers, to providing training compute for GenAI like ChatGPT, Stable Diffusion, and more, NVIDIA has a finger in every slice of the pie.
Moreover, NVIDIA holds a near-monopoly in the training hardware space. A 2020 study pegs NVIDIA’s market share for AI accelerators at a whopping 80.2%. What’s more, the company has bigger plans — one of a future where every AI model is trained on the NVIDIA hardware, no matter the cloud service provider. In Huang’s own words, NVIDIA AI is “essentially the operating system of AI systems today”.
Besides the partner program with cloud service providers, NVIDIA is actively working with companies from other industries. Last year, it joined forces with Deutsche Bank, Mercedes, Dell, Lockheed Martin, and Foxconn. What’s more, we are currently in the midst of a growth boom for AI chips, as seen by this study which predicts that this field will be valued at $263.6 billion by 2031 at a CAGR of 37.1%. With its market leader position, NVIDIA is poised to become the face of AI accelerator chips — the crown jewels of their enterprise strategy.
When looking at gaming’s reduced revenues, a slowdown in the pace of new gaming products, Huang’s renewed enthusiasm towards AI compute, and a lack of significant partnerships with other companies in the gaming field, it seems as though NVIDIA is content with its crown for now. The installed base for NVIDIA’s hardware is extremely high, but this base might start eroding if the company goes in search of greener pastures in the arms of the enterprise sector.
