Have Google & Nvidia Solved AI’s Data Compliance Problem?

According to McKinsey & Co., 78% of businesses around the world are now using AI to fulfil at least one business function, whether that's for data analysis, customer service or even for writing emails.

For some industries, AI has been like the ideal new hire, slotting in seamlessly. For others, integrating AI has been more difficult. 

In highly regulated industries like healthcare, finance or civil service, strict governance requirements prevent AI technologies from being used so freely.

This is simply because of the sensitivity of data at stake in these sectors. For companies in these fields, the risk of compromising private data is simply too high to introduce AI.

Nvidia and Google have recognised the opportunity here, though. The two tech heavyweights have expanded their strategic partnership in an attempt to make the deployment of AI in highly regulated workplaces easier and more compliant.

How Google and Nvidia plan to improve AI access

Highly regulated workplaces are also known as 'air-gapped environments', which means that they operate without external network connections in order to maintain isolation from cyberattacks or hackers.

For Google and Nvidia, the plan is to deploy the former's Gemini AI models using the latter's Blackwell GPU architecture to improve the level of cybersecurity that Google can offer these kinds of clients.

The two Silicon Valley powerhouses believe that by joining forces, their AI offering will be compliant with the rules and regulations of the air-gapped sectors, which could open doors to countless new customers.

But how does it work? Google has a system known as Distributed Cloud, which is a disconnected, sovereign, private cloud solution, allowing companies with sensitive data to set up a local network of nodes, creating an airtight network only accessible from the site where it is housed.

Meanwhile, Nvidia's Blackwell system has confidential computing capabilities that can provide an additional layer of security for clients.

Its confidential computing system uses hardware-based security features that encrypt data whilst the system remains in use, preventing unauthorised access even by administrators or cloud providers.

"By combining engineering excellence, open-source leadership and a vibrant developer ecosystem, the companies are making it easier than ever for developers to build, scale and deploy the next generation of AI applications," wrote Uttara Kumar, Senior Product Marketing Manager at Nvidia, in a recent blogpost.

A strategic partnership

Google's Gemini model family represents a sophisticated approach to multimodal AI systems, capable of processing text, images and various other data types within a unified model architecture.

These models excel in complex reasoning tasks, code generation and understanding intricate relationships between different information types.

To maximise the effectiveness of these advanced capabilities, Nvidia and Google have implemented targeted performance optimisations ensuring Gemini inference workloads operate at peak efficiency on Nvidia GPU infrastructure, particularly within Google Cloud's Vertex AI platform.

These enhancements enable Google to handle substantial volumes of user queries for Gemini models across both Vertex AI and Google Distributed Cloud environments.

The collaboration also extends to the Gemma family of lightweight, open models, which have been specifically optimised for inference using Nvidia's TensorRT-LLM library.

This acceleration framework enhances large language model inference by optimising neural network operations for Nvidia hardware architectures.

These optimised models are planned for release as Nvidia NIM microservices, which package AI models as containerized applications to simplify deployment processes.

The result is deployment flexibility across diverse architectures, from enterprise data centres to local Nvidia RTX-powered personal computers and workstations, allowing developers to select infrastructure that aligns with their specific performance requirements and deployment constraints.

This expanded partnership positions both companies at the forefront of enterprise AI adoption, bringing AI access to sectors that have thus far had to remain disconnected.

Explore the latest edition of Technology Magazine and be part of the conversation at our global conference series, Tech & AI LIVE.

Discover all our upcoming events and secure your tickets today.

Technology Magazine is a BizClik brand