Amadeus and NVIDIA Deepen AI Collaboration to Explore Faster Flight Search

A new memorandum of understanding expands GPU testing, travel-specific AI research and cloud optimisation as automated travel shopping grows.

KEY TAKEAWAYS

• Amadeus and NVIDIA have signed a memorandum of understanding to deepen research into accelerated flight search and travel-specific AI.
• Initial benchmarking on NVIDIA Hopper and Blackwell GPUs suggests potential improvements in search response times and scalability.
• The work spans GPU and CPU optimisation, specialised foundation models and cloud infrastructure efficiency.
• The collaboration remains exploratory, with no commercial launch date or quantified performance gains announced.

Amadeus and NVIDIA are extending their technical collaboration to explore how accelerated computing and specialised artificial intelligence could handle the increasing complexity of flight shopping. The companies have signed a memorandum of understanding aimed at developing faster and more scalable search technology, while preserving the accuracy and bookability of travel offers.

The agreement follows initial assessments using NVIDIA’s Hopper and Blackwell computing platforms. According to Amadeus’s 21 October announcement, the work remains exploratory: early benchmarks have indicated potential improvements, but the partners have not announced a commercial launch date or quantified performance gains.

Why AI-driven flight shopping puts pressure on search systems

Conventional flight searches already require travel technology systems to evaluate combinations of routes, availability and fares. As AI-powered assistants begin handling more elaborate requests on behalf of travellers, they could generate additional searches and demand more processing capacity. For travel sellers, the challenge is to respond efficiently without presenting flight options that cannot actually be booked.

The newly formalised work covers three technical areas. First, Amadeus is examining how graphics processing units (GPUs) and central processing units (CPUs) can be used together to improve flight-search workloads. Using NVIDIA CUDA, its GPU computing platform, Amadeus has adapted a portion of the search process so that calculations can be carried out in parallel. Initial testing on the Hopper and Blackwell architectures indicated scope to improve speed and scalability; broader testing is planned.

Travel-specific AI models and cloud efficiency

A second strand concerns the potential development of foundation models designed for travel-related tasks. The partners are considering collaboration with other industry participants to address challenges that may not be covered as effectively by general-purpose AI systems. This work could also help improve existing specialised models used across the travel sector.

The third area focuses on cloud and infrastructure efficiency. Amadeus and NVIDIA are reviewing approaches to resource planning and workload optimisation that could make more effective use of computing capacity while sustaining reliable service as travel-search demand changes.

These three elements are connected: additional processing capability may help manage the calculations required for complex itineraries, while tailored AI models and infrastructure planning could support the next generation of digital travel agents. However, the companies are evaluating possibilities rather than announcing that these capabilities are already available to customers.

What the partnership could mean for travellers and sellers

Lutz Vorneweg, SVP Strategic Partners at Amadeus, said the need for quick and accurate searches at global scale calls for different technical approaches as AI changes the travel ecosystem. He positioned Amadeus as a technology layer connecting providers, sellers and digital assistants.

David Hogan, NVIDIA’s Vice President of Enterprise EMEA, highlighted the new demands AI assistants are placing on computing infrastructure and described the collaboration as an opportunity to combine NVIDIA’s computing technology with Amadeus’s travel-industry expertise.

For travellers, the stated objective is faster access to flight options that reflect available and bookable inventory. Travel sellers could benefit from systems better able to cope with complex requests and rising search volumes. Any practical improvements will depend on subsequent development, testing and deployment.

The memorandum adds to wider efforts to prepare airline distribution and travel retailing systems for AI-mediated customer journeys. Further details of Amadeus’s technology programme can be found on its artificial intelligence and cloud innovation pages.