At the 2025 Cannes Lions global creative and advertising industry conference, NVIDIA is showcasing a suite of AI-powered advertising and marketing tools built with six partner companies, designed to automate repetitive enterprise ad operations and improve campaign measurement accuracy for large-scale marketing teams operating across global markets. The partner suite addresses four core adtech pain points identified by enterprise marketers: causal AI for measuring marketing return on investment (ROI), real-time AI bidding for programmatic ads, autonomous campaign management, and content-aware ad placement. Participating partners include Alembic, Amazon Web Services (AWS), Criteo, Higgsfield, KERV.ai and Taboola, all leveraging NVIDIA’s GPU and inference infrastructure to scale AI workloads for high-volume enterprise use cases. NVIDIA’s official announcement of AI ad tech tools at Cannes Lions
NVIDIA AI Marketing and Advertising Tools at Cannes Lions: Causal AI for Unbiased ROI Measurement
Traditional marketing measurement tools rely on correlation-based analysis that frequently misattributes sales lifts to incorrect campaigns, creating quantifiable wasted ad spend for enterprise marketing teams, per NVIDIA’s event announcement. Alembic’s causal AI approach eliminates that bias by modeling direct cause-and-effect relationships across all marketing touchpoints, a process that requires processing massive, rapidly changing datasets without reducing them to correlation-based assumptions. Unlike correlation-based tools that only identify relationships between marketing activities and sales, Alembic’s models isolate the specific impact of each individual campaign, even when multiple campaigns run simultaneously across channels.
Alembic is the first causal AI provider to deploy NVIDIA DGX Vera Rubin SuperPODs for enterprise-scale modeling, enabling it to process 3x more cross-channel, cross-market variables per run than earlier GPU configurations. The deployment runs on Equinix data center infrastructure paired with World Wide Technology’s secure deployment tools, keeping sensitive enterprise marketing data on-premises to meet strict regulatory requirements for industries including finance and healthcare, where data residency rules restrict cloud storage of customer and campaign data.
For example, enterprise finance sector clients using the setup have reported 22% reductions in wasted ad spend within the first two quarters of deployment, per NVIDIA’s event data. The setup delivers a single source of unbiased, audit-ready data for executives to identify which marketing initiatives drive actual business outcomes, and where budget is being wasted, eliminating the guesswork from marketing budget allocation for global enterprise campaigns.
NVIDIA AI Marketing and Advertising Cannes Lions: Real-Time Bidding and Content Intelligence
For programmatic advertising, where bid decisions must be made in milliseconds to win ad auctions in high-volume digital ad marketplaces, AWS is offering a production-ready reference implementation for AI-powered bidding that runs directly inside the auction pipeline, powered by NVIDIA Triton Inference Server. The stack combines AWS cloud infrastructure, pre-trained foundation models and NVIDIA GPU-accelerated computing to let demand-side platforms, supply-side platforms and adtech vendors replace static rules-based bidding logic with AI models that optimize bid prices, audience targeting and deal scoring in real time, adapting to changing market conditions faster than human-operated or rules-based systems.
The implementation reduces bid decision latency by 40% compared to static rules-based systems, per NVIDIA’s event details, allowing adtech vendors to process more bid requests per second and win more auctions without increasing bid costs. NVIDIA’s AWS real-time bidding reference implementation details
Criteo, which operates one of the largest product recommendation networks in digital advertising, collaborated with NVIDIA to optimize its adtech workloads on NVIDIA GPUs, enabling 25% faster model iteration for its recommendation systems. The performance gains from this collaboration allow Criteo to run additional model experiments to improve ad relevance for its publisher and advertiser clients, leading to higher click-through and conversion rates for ad campaigns.
Separately, Taboola is using NVIDIA GPUs to power its DeeperDive AI answer engine, extending the infrastructure to conversational AI platforms and chatbots to generate advertising revenue from AI-powered user interactions. The setup handles 10x more conversational queries per second than prior CPU-based infrastructure, per NVIDIA’s partner showcase data, allowing Taboola to scale its ad delivery to support the growing volume of AI chatbot user interactions without performance degradation.
What autonomous campaign management tools are on display at Cannes Lions?
For end-to-end marketing operations, Higgsfield AI is using the NVIDIA Agent Toolkit to power its Higgsfield Supercomputer agent platform, which manages the full marketing lifecycle from initial ideation and creative production to posting and autonomous optimization in a single unified interface, eliminating the need for marketers to switch between multiple disconnected tools for different campaign stages. The platform coordinates multiple image, audio, and video models alongside large language models, including Higgsfield’s in-house proprietary Soul and Soul 2.0 generative models, both built on NVIDIA’s Blackwell GPU architecture, to generate on-brand creative assets tailored to specific audience segments.
This infrastructure reduces end-to-end enterprise campaign creation time from 14 days to 48 hours for standard social and video ad assets, per NVIDIA’s event announcement, allowing marketing teams to launch time-sensitive campaigns faster and iterate on creative based on real-time performance data. NVIDIA’s agentic AI ad tech partner showcase details
NVIDIA NemoClaw blueprints and the OpenShell secure runtime are being integrated to provide enterprise trust guardrails including auditability and role-based permissioning for agent workflows, addressing common enterprise concerns about unregulated AI agent actions that could lead to off-brand content or policy violations. The platform is already used to create marketing campaigns for enterprise clients across industries, with use cases spanning social media content, video ads and cross-channel campaign assets.
How does NVIDIA power content-aware ad placement for higher engagement?
For contextual ad placement, KERV.ai uses NVIDIA’s multimodal AI stack to evaluate signals across video and media assets, identifying individual scenes, objects and products to match ad creative to content for higher engagement, rather than relying on broad keyword targeting that often results in mismatched ad placements. The system is designed to deliver real-time content recommendations for advertisers, reducing mismatched ad placements that lead to low user engagement and wasted ad spend. NVIDIA testing shows the system delivers 30% higher user engagement rates than industry-standard contextual ad placement tools, per the company’s Cannes Lions showcase materials.
Bottom line: Enterprise adtech teams evaluating AI tooling at Cannes Lions should prioritize NVIDIA partner stacks that offer on-premise deployment options for sensitive regulated data (per Alembic’s Equinix and World Wide Technology setup built for finance and healthcare compliance requirements, which delivers 22% reductions in wasted ad spend for finance clients), proven low-latency inference for real-time bidding workloads (per AWS’s Triton Inference Server-powered production reference implementation, which cuts bid decision latency by 40% compared to static rules), and integrated agent guardrails for autonomous campaign workflows (per Higgsfield’s NemoClaw and OpenShell integration, which reduces end-to-end campaign creation time from 14 days to 48 hours).
