New capabilities bring autonomous optimization to AI inference infrastructure, helping organizations balance GPU cost, performance, and efficiency at scale.
New York – June 9, 2026
Akamas, the AI-powered optimization platform for cloud and Kubernetes environments, today announced the launch of AI Infrastructure Optimization, a new set of capabilities designed to help organizations optimize the cost, performance, and efficiency of AI workloads running on GPU-based infrastructure.
As enterprises move generative AI initiatives from experimentation into production, infrastructure teams are facing a growing challenge: rapidly increasing GPU costs, infrastructure complexity, and the need to balance throughput, latency, reliability, and operating expenses. While observability platforms provide visibility into AI systems, identifying the best configuration across infrastructure and runtime layers remains largely a manual and time-consuming process.
Akamas addresses this challenge by extending its optimization technology to AI infrastructure environments, enabling organizations to continuously identify and validate optimal configurations across Kubernetes environments, GPU resources, AI serving infrastructure, and inference runtime parameters.
The new capabilities allow engineering teams to:
- Improve GPU utilization and infrastructure efficiency
- Reduce the cost of AI inference workloads
- Increase throughput and application performance
- Optimize Kubernetes and GPU resource allocation
- Tune AI serving infrastructure and inference runtime parameters
- Balance multiple objectives such as cost, latency, throughput, and reliability
Rather than relying on static best practices, manual experimentation, or trial-and-error approaches, Akamas continuously evaluates thousands of possible configurations to identify the best operating conditions across infrastructure, orchestration, and inference-serving layers.
Built on the same optimization technology already used to improve the performance and efficiency of enterprise applications, cloud environments, and Kubernetes workloads, the new capabilities help organizations make data-driven optimization decisions across the AI stack.
“Everybody is talking about AI models. Very few people are talking about the operational cost of running them at scale,” said Luca Forni, CEO of Akamas. “As organizations move from experimentation to production, optimization becomes just as important as innovation. We’ve spent years helping enterprises optimize applications, Kubernetes environments, and cloud infrastructure. AI infrastructure is simply the next frontier. The challenge is no longer just building AI systems, it’s operating them efficiently, reliably, and economically.”
Forni added: “The industry has invested heavily in AI observability, and that’s a necessary step. But visibility alone doesn’t improve efficiency. At some point every organization needs to answer a different question: what’s the best configuration for this workload, on this infrastructure, under these constraints? That’s the problem we’re solving.”
The launch reflects a broader shift occurring across the industry. As AI workloads become a growing component of enterprise infrastructure, organizations are increasingly looking beyond monitoring and reporting tools toward technologies that can actively improve efficiency, utilization, and operational outcomes.
By extending optimization capabilities to AI infrastructure, Akamas enables Platform Engineering, SRE, Cloud Operations, and AI Engineering teams to apply a consistent optimization approach across both traditional cloud-native workloads and next-generation AI applications.
The new AI Infrastructure Optimization capabilities are available immediately as part of the Akamas platform. The capabilities are being introduced this week at Datadog DASH 2026 in New York, where Akamas is showcasing its latest optimization technologies at Booth #532.
About Akamas
Akamas is an AI-powered optimization platform that helps organizations continuously improve application performance, reliability, and infrastructure efficiency. By automatically identifying optimal configurations across infrastructure and application layers, Akamas enables engineering teams to reduce costs, increase performance, and eliminate manual tuning activities. Akamas supports cloud-native applications, Kubernetes environments, enterprise software stacks, and AI infrastructure workloads through a single optimization platform.