Uber Runs on Ampere®: How the Move to OCI Unlocked Performance, Efficiency and Next-Level Savings
Today at Oracle CloudWorld 2024, Uber’s Kamran Zargahi joined Oracle’s Clay Magouryk to discuss a key modernization effort that has been underway for several months. Over this period, Oracle Cloud Infrastructure (OCI), Ampere and Uber have worked closely to enable the transition of Uber’s production workloads to OCI Ampere A1 and A2 compute shapes. The news serves as a pivotal example of how enterprises can modernize their infrastructure to achieve cost savings, enhance performance, and drive sustainability. Uber’s success story reflects the impact Ampere is having, not only for ride-sharing giants but across industries that demand scalable, cost-efficient, and powerful cloud solutions.
While Uber’s achievements are noteworthy, they are part of a broader movement. Ampere’s cutting-edge AI inference capabilities, security innovations, and collaboration with Oracle’s first-party services continue to reshape the future of cloud computing.
Inside Uber’s Move to OCI Ampere Compute
Uber’s journey of continuous growth requires infrastructure that can scale with its business. As the company adds 1.5 billion trips annually to its platform, balancing rapid expansion with operational efficiency is crucial. Partnering with Oracle and Ampere allowed Uber to modernize its backend infrastructure, delivering substantial cost savings and sustainability benefits.
By migrating to OCI Ampere compute, Uber reduced its infrastructure costs, while cutting power consumption by 30%. These results have been critical for Uber’s long-term growth strategy, enabling the company to maintain profitability while pursuing ambitious expansion goals.
Moreover, Uber’s move to OCI Ampere compute reinforced its commitment to sustainability. In a time when some companies are returning to on-premises solutions, Uber has proven that cloud infrastructure, with the right partners, can deliver both operational and environmental benefits.
AI Inference on OCI Ampere Compute: Driving Innovation Across Industries
While Uber’s experience highlights their own efficiency and sustainability gains, Ampere’s AI capabilities also offer significant benefits to a wide range of enterprises. Ampere’s processors now deliver up to 152% better price-performance for AI inference when using Meta’s llama.cpp models, making Ampere an attractive choice for companies looking to optimize AI workloads.
Ampere’s A1 compute shapes allow businesses to run AI inference and general compute tasks simultaneously, maximizing resource utilization and reducing the need for expensive GPUs. This versatility is key to enabling faster innovation across industries. With OCI Ampere compute, enterprises can adopt advanced AI without compromising on cost or efficiency, ensuring that companies of all sizes can scale AI workloads as their needs evolve.
New Oracle First-Party Services Run on OCI Ampere Compute
In addition to Uber, Oracle and Ampere’s partnership has expanded, with 95% of Oracle's services now running on OCI infrastructure powered by Ampere. This marks a significant step forward for cloud computing and demonstrates Oracle’s trust in Ampere to support its growing portfolio of cloud services.
One of the most notable additions is the Oracle Data Science platform, now running on OCI Ampere compute. This platform plays a crucial role in supporting large-scale AI model development and deployment. By leveraging Ampere’s processors, Oracle offers its customers better price-performance for demanding AI workloads, giving enterprises a powerful toolset to unlock the full potential of data science and AI.
OCI Ampere A1 & A2 Compute Shapes: Available Now with Leading Price-Performance
Uber’s migration to OCI Ampere compute demonstrates how large enterprises can harness the power of Ampere-based compute shapes to scale efficiently. By adopting OCI Ampere A1 and A2 compute shapes, Uber achieved the infrastructure flexibility needed to manage its growing platform while significantly reducing costs. This success story is a testament to the broad appeal of Ampere’s technology for enterprises looking to optimize both performance and sustainability.
Ampere’s A1 and A2 compute shapes are now both available to enterprises through OCI, offering cost-effective solutions for both general compute and AI inference workloads. These shapes strike an important balance between performance, cost, and sustainability, making them an ideal choice for companies with diverse infrastructure needs.
Improved Security with AmpereOne Processor
While Uber benefits from OCI’s robust cloud infrastructure, the AmpereOne processor offers enterprises additional security features that further enhance data protection. AmpereOne’s built-in memory tagging capabilities mitigate vulnerabilities such as buffer overflow attacks and pointer programming errors. For enterprises handling sensitive data or operating in highly regulated industries, AmpereOne provides a strong defense layer to ensure secure operations.
The partnership between Uber, Ampere, and Oracle is setting a new standard for cloud computing, driving growth, efficiency, and sustainability. Uber’s migration to OCI Ampere compute has resulted in significant cost savings, reduced environmental impact, and an infrastructure capable of supporting future growth. This success story showcases how Ampere’s A1 and A2 compute shapes are empowering enterprises to balance cost, performance, and sustainability.compute compute.
Beyond Uber, Ampere’s impact extends across Oracle’s services and a wide range of industries. With a focus on AI innovation, price-performance, and cutting-edge security, Ampere is reshaping how enterprises approach cloud infrastructure and setting the stage for the next wave of digital transformation.
Ampere Computing LLC
4655 Great America Parkway Suite 601
Santa Clara, CA 95054
