When we founded Cloudar in 2014, we made a choice that may have seemed radical at the time: we would build exclusively on AWS. No multi-cloud story, no hedging between platforms. AWS, and nothing else. Twelve years later, Cloudar stands as the first local Benelux partner with Anthropic Authorized Reseller status. A coincidence? I don’t think so. After all, it is the result of the same logic.
AWS: the choice that determined everything
In the cloud industry, “focus” is a dangerous word. It sounds great in a pitch deck, but in practice it means leaving opportunities on the table. Customers running on Azure, projects you have to refer elsewhere, RFPs where you know from the start you won’t win. We accepted that over the years. Our drive was (and still is) simple: by concentrating everything on one platform, we can become the best possible partner on it. Not broad, but deep. That translates concretely into certifications, into architectural knowledge, into the relationships you build with the AWS team… and ultimately into the trust that customers place in you for their most critical workloads. Becoming an AWS Premier Partner was a milestone in that journey. But it was never the endpoint, let alone the ultimate goal.
With the rise of generative AI, the role of cloud architecture has fundamentally changed. The question is no longer “how do we migrate our workloads to the cloud?” but “how do we design infrastructure that can run AI models at scale, securely, cost-efficiently and with the right governance?” That is an architectural challenge of an entirely new order. Inference costs, latency optimisation, data residency, IAM configurations compatible with foundation models, VPC setups that withstand enterprise compliance: these are the problems our customers face. And these are the problems for which you need to know AWS to the bone.
AWS Bedrock has become the platform in that landscape where enterprise AI adoption takes place in practice. Not because it is the only option, but because it is the option that takes security, compliance and integration with existing AWS workloads seriously. For organisations that have been running on AWS for years, Bedrock is the most logical bridge to production-ready AI.
What Anthropic Authorized Reseller status concretely means
I’ll be honest: it took quite some effort. The Anthropic Authorized Reseller status is not a badge you apply for and receive in the post a week later. It requires knowledge, skill, a demonstrable pipeline and navigating an approval process in which Anthropic itself decides whether you clear the bar. A handful of AWS partners worldwide manage that. Most do not. And we are now the first local Benelux player to join that group.
What this concretely means for the Benelux market is perhaps less well known than it should be. Since October 2025, AWS requires that a customer’s managing partner be officially recognised by Anthropic, if that customer wants access to the latest Claude models via Bedrock. This applies to Claude 4, Claude Code, and all future frontier models. Without a recognised partner: no access.
In practice, this meant that large organisations in the Benelux were faced with a closed door for months. They wanted to experiment with the most capable models on the market, had the budget, had the use cases, but had no AWS partner with the right status. For CTOs, CIOs or enterprise architects who take AI seriously, there are a few practical implications worth mentioning.
- First: access to frontier models has become a partnership question, not just a technical or commercial one. The architectural choices you make today – which AWS partner manages your account, how your Bedrock setup is configured – partly determine which AI capabilities you can deploy tomorrow.
- Second: enterprise seats for Anthropic are not the same as API access. It is about onboarding, governance, user management and monitoring data flows within your own AWS environment. That requires someone who knows both the Anthropic side and the AWS infrastructure.
- Third: Claude Code is a relevant example of where this is heading. AI-driven code assistance running directly on your own AWS infrastructure, without your code leaving the perimeter of your own environment. That is a different proposition from a SaaS tool with a subscription.
In 2014 we chose AWS because we believed that focus leads to expertise, and expertise leads to trust. In 2025 and 2026, AI has not changed our conviction. The complexity of production-ready AI at enterprise scale requires precisely the kind of deep platform knowledge we have invested in over the past twelve years.
The Anthropic Authorized Reseller status is, in that light, not just a commercial milestone. It is the validation that our approach works. And for our customers, it is the guarantee that they can get started with the most advanced AI models on the market, guided by a local team that knows the full stack: from infrastructure to model deployment to governance. But as has been said before, for us this is far from the final destination.
