Artificial Intelligence & Machine Learning
Autodesk and AWS Are Driving the Next Generation of AI-Powered Design Innovation
Design is the silent force that brings ideas to life. From building cities to developing transformative products and creating inspiring visual narratives, design is at the core of innovation. Yet, as the challenges in modern innovation grow increasingly complex, so do the tools and processes required to address them. How can technology extend human creativity, enabling designers to work with greater precision, speed and sustainability?
See Also: How to Take the Complexity Out of Cybersecurity
For Autodesk, a leading design technology company, the answer lies in generative AI. At AWS re:Invent 2024, Raji Arasu, CTO and executive vice president at Autodesk, unveiled the company’s innovative vision for using AI to redefine design and production. This vision is powered by advanced AI models developed in collaboration with Amazon Bedrock and SageMaker, enabling Autodesk to address industry challenges while reimagining what’s possible.
A Bold Leap Forward: Project Bernini
At the core of Autodesk’s gen AI strategy is Project Bernini, a billion-parameter foundation model designed to transform 2D and 3D CAD geometries. This technology converts diverse inputs – text, sketches and voxels – into high-fidelity models with exceptional accuracy.
The implications are transformative. Architects, engineers and creators can now push the boundaries of their imagination, whether for designing a single building or visualizing an entire city. Project Bernini acts as a tool as well as a collaborator, bridging the gap between the physical and digital worlds. It simplifies complex workflows, allowing designers to focus on creativity while AI handles the technical intricacies.
AWS: Powering Autodesk’s Ambitions
Autodesk’s leap into gen AI is fueled by its 15-year partnership with AWS. This collaboration provides the cloud, data and AI infrastructure required to make Autodesk’s bold ambitions a reality. Together, they address some of the toughest challenges in design innovation:
- Processing massive data volumes: Modern design requires managing petabytes of data, from intricate building schematics to extensive infrastructure plans. With AWS solutions such as Amazon DynamoDB, Autodesk ensures seamless access to critical data with zero latency.
- Optimizing data quality: High-quality datasets are crucial for effective AI models. Autodesk uses AWS tools such as Amazon EMR, Amazon EKS and AWS Glue to refine its data for better model performance and accuracy.
- Accelerating model training: Training billion-parameter models such as Bernini demands immense computational power. AWS services such as Amazon SageMaker and Elastic Fabric Adapter help reduce training times by 50%, thereby reducing costs and accelerating innovation.
- Delivering real-time AI inference: AI outputs must be fast and precise. SageMaker ensures Autodesk’s models achieve high throughput and low latency, enabling seamless real-world applications.
These advancements have boosted AI productivity by 30%, halved foundational model costs and significantly accelerated design workflows. Beyond efficiency, Autodesk is embedding sustainability into its processes, ensuring today’s designs contribute to a greener future.
Strategic Insights for Tech Leaders
“What if Autodesk could work five times faster? With AWS, we’re exploring that possibility,” said Arasu, encapsulating the excitement surrounding Autodesk’s journey.
Looking ahead, Autodesk is already experimenting with Amazon SageMaker Unified Studio, further pushing the limits of gen AI to create tools that redefine 3D CAD design. For tech leaders, Autodesk’s approach offers critical insights:
- How to align AI strategies with core business objectives;
- The importance of leveraging partnerships to drive innovation at scale;
- The strategies for integrating sustainability into cutting-edge technologies.
Autodesk’s innovations are redefining design, blending the physical and digital worlds to shape a bold, AI-driven future.