Valendra joins NVIDIA Inception to advance QDivZero
Published 14 September 2026 by Marta Barea Sepúlveda3 min read
Today marks an important step for Valendra Tech SL and QDivZero. We have been accepted into NVIDIA Inception.
We are joining at a particularly relevant point in the product's development. QDivZero is being built to remove the complexity of deploying AI models, while the program supports startups as they move from prototype to production within the accelerated computing ecosystem.
What we are building with QDivZero
QDivZero lets developers run open-weight AI models without having to understand or manage the infrastructure behind them.
A developer chooses a model from Hugging Face or our catalogue. QDivZero determines its hardware requirements, finds compatible compute across infrastructure providers, deploys the model and exposes it through a single OpenAI-compatible API.
The platform handles everything underneath. This includes GPU compatibility, provisioning, inference engines, quantization, model serving, scheduling, scaling and the deployment lifecycle. Each workload is matched with suitable infrastructure to achieve an efficient balance between price and compute.
QDivZero also offers a different economic model from inference APIs that charge per token. Customers pay for the active compute capacity running their models. This provides dedicated capacity, makes costs easier to predict and can reduce execution costs when inference volume is high.
The application remains separate from the infrastructure. A team can change models, hardware or providers without rebuilding its integration because the API contract stays the same.
We are building the compute abstraction layer for open-weight AI. Developers choose what they want to run, and QDivZero determines how and where to run it efficiently.
What NVIDIA Inception is and why it fits QDivZero
NVIDIA presents Inception as a free program that guides AI startups from prototype to production. Members can access technical training, developer tools, preferred pricing on selected hardware and software, and offers from partner companies. The program also includes community activities and, where the relevant requirements are met, opportunities connected with events and investors.
That journey from model to production is directly related to QDivZero's work. Our platform aims to make infrastructure complexity invisible, but automating it requires a deep understanding of the systems underneath.
Every deployment must connect a model's size and architecture with available memory, GPU type, inference engine and runtime configuration. Capacity then needs to be provisioned, the model served, its operation monitored and its cost controlled. QDivZero turns that process into a simple experience for the developer.
This is why we see a practical fit with NVIDIA Inception. The program's training and tools may provide useful context in areas at the technical core of QDivZero. Its business ecosystem may also help us compare decisions with other companies bringing AI products into production.
The value does not lie in adopting every available technology. It lies in evaluating each one more effectively and deciding whether it contributes to broader compatibility, more reliable deployments or more efficient execution.
How we want this to translate into the product
We want this membership to support better-informed technical decisions. A deeper understanding of how models behave across GPU configurations may help us refine automatic infrastructure selection and the balance between price and compute.
It may also broaden our perspective when working with inference engines, quantization, scheduling and scaling. Users should not have to manage these layers, but QDivZero must coordinate them correctly to provide a reliable experience.
From a business perspective, the program places us within a community focused on AI and accelerated computing. We want to use that environment to learn, share what we are building and open conversations that create tangible value for the product and its users.
Keep exploring QDivZero
To learn more about how we are turning this vision into a product, continue with these resources.
