In an industry racing toward artificial intelligence (AI) dominance, Akamai is charting a course that differs from traditional hyperscalers – one that combines the power of edge computing with open source flexibility and specialized AI inference capabilities.

In the latest episode of our new show KubeStruck recorded at KubeCon + CloudNativeCon in London, Ari Weil, VP of Product Marketing at Akamai, said that Akamai is positioning itself not as another hyperscaler, but as “hyperscale done differently.” As traditional cloud providers struggle with centralization bottlenecks and businesses seek better returns on their AI investments, Akamai’s distributed approach offers a compelling alternative that could reshape how we think about cloud infrastructure.

Embracing Open Source as a Foundation

Weil says Akamai has long been a proponent of open source technologies, and that commitment continues to strengthen. “We heavily use open source on our platform, and this year, we’re really excited because we’ve been able to give a home to the Linux kernel on the Akamai cloud,” Weil explained.

This commitment extends beyond mere usage—Akamai is actively contributing to the open source ecosystem as a gold sponsor of the Cloud Native Computing Foundation (CNCF), with a $1 million pledge to support CNCF projects. They’re also investing in upstream projects including OpenTelemetry, Prometheus, and Flatcar OS, which features prominently in platform engineering work. Akamai Cloud has also become home to kernel.org that hosts the Linux kernel.

This dedication to open source serves as the foundation for Akamai’s broader vision: providing developers with the flexibility and portability they need while delivering enterprise-grade performance. It’s this foundation that enables their distinctive approach to edge computing and AI.

The Edge Computing Revolution: Responding to Industry Shifts

Weil highlighted a significant transformation occurring in the industry that Akamai is well-positioned to address. “The cloud is not fit for purpose for some of these low-latency use cases because it’s too centralized and lacks expertise in high-performance, high-throughput networking,” he noted.

Central to Akamai’s response is their extensive network of over 4,200 edge locations globally. By integrating compute capabilities into this distributed network, they’re creating an infrastructure that gives developers unprecedented control over data flow and sovereignty, optimized network performance, flexible compute resource deployment, and multi-cloud portability.

This distributed architecture directly addresses the limitations of centralized cloud models, particularly for latency-sensitive applications and regions with data sovereignty requirements. It also sets the stage for Akamai’s innovative approach to AI deployment.

Transforming AI with Edge Inference

Perhaps the most compelling aspect of Akamai’s strategy is their approach to AI inference at the edge. As businesses struggle with the cost and complexity of training large language models in centralized environments, Akamai offers a more practical alternative.

“Businesses are now shifting to ‘I want to use a commercially available large language model. I’ll train it with my data, and then I’ll create a small language model or a fine-tune model that I can now verticalize for my specific use case and for my users, distribute that compute use case out to the edge using a platform like [Akamai],'” Weil explained.

This edge-based approach delivers multiple benefits: high data throughput, microsecond-to-millisecond performance for compute tasks, efficient GPU usage specifically optimized for inference, and—critically—better monetization opportunities through practical AI applications delivered closer to end users.

Democratizing AI Through Resource Flexibility

Addressing the widespread challenge of GPU scarcity and cost, Akamai has developed a flexible approach to compute resources that makes AI more accessible. Through strategic partnerships with companies like Neural Magic, customers can run certain inference workloads on CPUs rather than GPUs, significantly reducing costs while maintaining performance.

For applications requiring greater processing power, Akamai offers VPUs and GPUs that can dramatically accelerate workflows. Weil shared a compelling example where a customer achieved 16x faster encoding and transcoding by using VPUs on the Akamai platform—turning a days-long process into hours.

“We believe that offering choice will ultimately give them the flexibility they need to make Akamai the home for any real-time compute use cases they want to extend or complement the hyperscalers with,” said Weil. This adaptability is particularly valuable as organizations experiment with various AI applications, from video processing to interactive streaming experiences.

Bridging the Gap Between Sovereign Clouds and Hyperscalers

Akamai has positioned itself as an elegant middle ground between highly localized sovereign clouds and complex, expensive hyperscalers. This balance is increasingly important as regions like Europe emphasize data sovereignty and sustainability.

“With Akamai, you’re getting distribution, the ability to localize your data, lower costs and greater predictability,” Weil explained. “We’re also building everything on our platform with an eye toward choice, flexibility, and portability, so we’re not locking you into our platform with powerful but proprietary resources.”

This approach provides the reliability and scale businesses need without the vendor lock-in and complexity that characterize traditional hyperscalers. It allows organizations to maintain sovereignty over their data while still benefiting from global distribution and enterprise-grade reliability.

The Future: Simplification Through Self-Service

Looking ahead, Akamai is focused on making their platform even more accessible through enhanced self-service capabilities. They’re integrating security applications and content delivery network features into the Akamai Cloud platform (formerly Linode) while embracing low-code and no-code development approaches alongside technologies like WebAssembly.

“What we believe will happen moving forward is that people will focus more on front-end coding and providing platform providers like Akamai with their business logic, while we handle the infrastructure components, distribution, scale, and failover,” Weil noted.

This vision aligns with Akamai’s commitment to sustainability as well, with goals to become carbon neutral by 2030. Their distributed approach naturally leads to smaller data centers that pull less power and offer a more eco-friendly alternative to massive hyperscaler facilities.

Akamai’s Distinctive Value Proposition

Akamai’s strategy represents a thoughtful response to converging industry trends: the growing demand for distributed computing resources, the rise of practical AI applications, and developers’ desire for flexibility without sacrificing performance or sustainability.

By leveraging their networking expertise and global presence while embracing open source technologies and offering flexible compute options, Akamai has created a distinct alternative in the cloud marketplace. Their approach provides the distribution, first-party expertise, and community engagement that many organizations need as they navigate the complex landscape of edge computing and AI deployment.

As Weil summarized their philosophy: “We’re not trying to be yet another hyperscaler. We want to be hyperscale done differently.” For organizations looking to balance performance, flexibility, and cost in their cloud strategy—particularly for AI workloads—Akamai’s distinctive approach merits serious consideration.

Guest: Ari Weil (LinkedIn)
Company: Akamai
Show: KubeStruck