The demand for AI is changing the way businesses think about cloud infrastructure. As AI applications become more complex and widely adopted, organizations need infrastructure that can support large-scale workloads while delivering performance, reliability, security, and flexibility.
A recent agreement between Akamai Technologies and Anthropic highlights just how quickly these infrastructure requirements are growing.
On September 24, 2026, Akamai announced a $11.6 billion, seven-year contractual commitment with Anthropic. Under the agreement, Anthropic will use Akamai Cloud’s distributed infrastructure and software to support the growth of its CPU workloads. The relationship could also expand by an additional $9 billion, bringing the potential total commitment to approximately $20 billion.
For businesses watching the evolution of AI infrastructure, this announcement offers an important look at where cloud computing is heading.
What Is the Akamai and Anthropic Agreement?
The agreement represents a significant expansion of the existing relationship between Akamai and Anthropic.
Under the seven-year commitment, Anthropic will use Akamai Cloud’s distributed infrastructure and software to support its growing CPU workload requirements. Akamai says the agreement is designed to help Anthropic build and operate AI workloads at scale.
The deal also includes the possibility of expanding the relationship by another $9 billion, subject to additional cloud service commitments. If fully expanded, the relationship could reach approximately $20 billion over the seven-year period.
Akamai also estimates approximately $5.5 billion in capital expenditures related to the $11.6 billion commitment. The company expects about $1.7 billion of additional capital expenditure in 2026 to secure and pre-purchase critical supply chain components, including memory.
This scale reflects a broader trend: AI companies increasingly need dedicated, long-term access to computing infrastructure rather than relying only on conventional cloud capacity.
Why Does Anthropic Need More AI Infrastructure?
AI workloads are not limited to training large models.
Once an AI model is developed, it needs infrastructure to support applications, APIs, inference, data processing, orchestration, security, and other supporting workloads. As usage increases, these workloads can create significant demand for compute resources.
This is where CPU infrastructure continues to play an important role.
While GPUs are closely associated with AI training and high-performance computing, CPUs remain essential for many supporting workloads. These can include application logic, data processing, networking, orchestration, storage operations, and other components required to run AI-powered applications.
Anthropic’s agreement with Akamai specifically highlights the growing requirement for CPU workloads at scale.
The development also comes at a time when Anthropic is making major long-term infrastructure commitments across the industry. Reuters reported on September 29 that Anthropic’s confidential IPO filing showed plans for at least $518 billion in AI infrastructure commitments over the coming decade, highlighting the scale of computing capacity it expects to require.
Why Distributed Cloud Infrastructure Matters for AI
Traditional cloud infrastructure often relies on centralized regions and large data centers. Distributed cloud takes a different approach by placing computing resources across a wider geographic footprint.
For AI applications, this can have several practical benefits.
1. Bringing Compute Closer to Users
AI applications often need to process requests quickly. When workloads can run closer to users, businesses can potentially reduce network distance and improve application responsiveness.
This becomes particularly important for applications that depend on real-time interactions, including AI assistants, conversational applications, recommendation systems, and intelligent business applications.
2. Supporting Large-Scale Workloads
AI workloads can fluctuate significantly depending on usage.
A distributed infrastructure model can provide organizations with access to computing resources across multiple locations, helping them handle workloads as demand changes.
Akamai says its cloud platform supports a continuum of compute from core to edge and spans thousands of points of presence globally.
3. Improving Application Performance
AI applications are becoming part of customer-facing digital experiences. Slow responses can directly affect user experience.
Distributed infrastructure can help organizations position applications, compute, and supporting services closer to the locations where they are needed.
4. Supporting Global AI Deployments
Enterprises increasingly operate across multiple regions. AI applications may need to serve users in different countries while meeting performance, security, and regulatory requirements.
A globally distributed cloud infrastructure can provide more flexibility when deploying applications across different geographic locations.
What Makes the Akamai-Anthropic Deal Significant?
The size of the agreement is notable, but its significance goes beyond the dollar value.
It demonstrates that AI infrastructure is becoming a long-term strategic requirement.
Akamai had already announced more than $2.8 billion in multi-year Cloud Infrastructure Services commitments across its customer base earlier in 2026. The Anthropic agreement adds substantially to that momentum.
It also shows that AI infrastructure is not simply about access to GPUs.
A complete AI environment requires multiple layers of infrastructure, including compute, networking, storage, application delivery, security, and supporting software.
Akamai’s recent infrastructure investments reflect this broader approach. In March 2026, the company disclosed a $200 million agreement involving a multi-thousand NVIDIA Blackwell GPU cluster and other cloud infrastructure services. Akamai also highlighted the expansion of its global IaaS footprint and its focus on AI inference and distributed computing.
The Anthropic agreement therefore fits into a larger strategy around supporting AI workloads across their lifecycle.
What This Means for Enterprise AI Adoption
The Akamai and Anthropic announcement is also relevant to enterprises that are still building their AI strategies.
Businesses adopting AI need to think beyond the model itself. Choosing an AI platform is only one part of the equation. Organizations also need to consider the infrastructure required to deploy, scale, secure, and manage AI applications.
Some of the key considerations include:
- Compute capacity: Can the infrastructure support increasing workload requirements?
- Performance: Can applications deliver consistent response times?
- Scalability: Can infrastructure expand as AI adoption grows?
- Security: Can sensitive data and AI workloads be protected?
- Global reach: Can applications serve users across multiple regions?
- Reliability: Can critical AI applications remain available as demand increases?
- Cost management: Can organizations scale infrastructure without creating unnecessary operational costs?
These questions become increasingly important as AI moves from experimentation into production environments.
AI Infrastructure Is Moving Beyond the Data Center
One of the broader trends emerging from developments like the Akamai-Anthropic agreement is the shift toward more distributed AI infrastructure.
AI does not always need to run in a single centralized location.
Depending on the workload, organizations may need a combination of centralized cloud infrastructure, distributed compute, edge services, and dedicated AI infrastructure.
Akamai’s distributed cloud approach is designed around this broader infrastructure model. Its platform supports compute from core to edge and is intended to help organizations build, deploy, and operate applications across a distributed environment.
For enterprises, this can create more options for designing AI architectures based on application requirements rather than relying on a single infrastructure model.
What Businesses Can Learn From the Akamai and Anthropic Partnership
The agreement provides several lessons for organizations planning their AI infrastructure strategy.
-
AI infrastructure needs to scale with adoption
AI projects that begin as small pilots can eventually become business-critical applications. Infrastructure planning needs to account for that potential growth from the beginning.
-
CPU and GPU resources both matter
AI infrastructure is not exclusively about GPUs. CPUs continue to support many of the workloads surrounding AI applications, from data processing and application services to orchestration and infrastructure management.
-
Location can influence performance
For applications that require low latency, where computing resources are located can be as important as how much computing capacity is available.
-
Security needs to be part of the architecture
As AI applications interact with enterprise data and business systems, security cannot be treated as an afterthought. Infrastructure, application delivery, data protection, and security need to work together.
-
Infrastructure partnerships are becoming strategic
The size and duration of agreements such as the Akamai-Anthropic deal show that access to infrastructure is becoming a strategic consideration for companies operating AI workloads at scale.
How ZNet Can Help Businesses Prepare for AI-Driven Infrastructure Needs
As organizations adopt AI, they need more than access to individual technologies. They need an infrastructure strategy that connects cloud, compute, security, performance, and application requirements.
ZNet helps businesses evaluate and implement cloud and digital infrastructure solutions based on their business and workload requirements.
Through its ecosystem of technology partnerships and cloud services, ZNet can help organizations explore infrastructure options for workloads that require scalability, performance, security, and reliability.
Whether a business is evaluating AI workloads, modernizing its cloud environment, or preparing its infrastructure for increased digital demand, having the right technology and infrastructure partner can simplify the journey.
Final Words
The $11.6 billion Akamai-Anthropic agreement is more than a large cloud services contract. It reflects the infrastructure demands being created by the rapid growth of AI.
As AI applications move into production and usage continues to expand, organizations will need infrastructure that can support more than model training. Compute, networking, storage, security, application delivery, and distributed infrastructure will all play a role.
The potential expansion of the Akamai and Anthropic relationship to approximately $20 billion further demonstrates the scale at which AI infrastructure requirements can grow.
For enterprises, the message is clear: planning for AI adoption also means planning for the infrastructure that will make that adoption possible.
