Nvidia has recently introduced an innovative business model for its AI accelerators, targeting smaller cloud data center operators, known as Neoclouds. This new approach allows these operators to purchase Nvidia’s hardware, which can then be rented by startups and other end users. The significant innovation here is a “credit-support model,” where Nvidia commits to renting back unused computing capacity if the operator struggles to attract sufficient customers. In return, these operators share a portion of their revenue with Nvidia. However, it remains unclear whether this model comes with discounted hardware prices.
### The Emergence of Neoclouds
Neoclouds are designed to help operators acquire the latest hardware with reduced financial risk. Initial partners in this venture include Australian companies Sharon AI and Firmus Technologies. Sharon is reportedly developing a data center with a capacity of up to 72 megawatts and around 40,000 GB300 boards. Firmus, on the other hand, aims to establish a campus in Indonesia that can scale up to 360 megawatts with 170,000 Nvidia GPUs.
### The Advantage of Customized AI Data Centers
A considerable advantage for Nvidia in this model is that the AI data centers created under this initiative will be built according to Nvidia’s specifications. This means that in addition to AI accelerators, Nvidia can also sell processors, switches, and other essential components, further enhancing their revenue streams.
### A Consistent Revenue Stream
Nvidia’s Chief Financial Officer, Colette Kress, describes the credit-support model as a “recurring, usage-based revenue stream.” This program serves as an additional method to fill their order books. Previously, Nvidia made significant investments in AI companies, which often resulted in those companies purchasing or renting Nvidia hardware. However, this closed-loop model primarily benefited larger customers like OpenAI.
The new credit-support model aims to offer flexibility without the need for large capital investments. End users will not have to worry about building a data center while still accessing the necessary computing power.
### Implications for the Industry
The implications of Nvidia’s new model could be substantial. By lowering the barrier to entry for smaller operators, Nvidia not only diversifies its customer base but also encourages innovation within the AI ecosystem. Smaller companies can now participate in the AI race without the substantial upfront costs usually required for infrastructure.
Furthermore, this model fosters competition among cloud service providers, as those that offer Nvidia hardware can create more attractive pricing and service packages for startups and other organizations needing AI capabilities.
### Conclusion
In summary, Nvidia’s introduction of a revenue-sharing model paired with its AI accelerators signifies a pivotal shift in the cloud computing landscape. As more players enter the field through Neoclouds, the market will likely become increasingly competitive and innovative. Overall, this strategy could enhance Nvidia’s position in the AI market, driving growth not just for themselves but for the entire industry.
