Blaxel raises $7.3M seed round to build ‘AWS for AI agents’ after processing billions of agent requests

BlexelA specially designed startup setup for artificial intelligence agents was raised to $ 7.3 million seed funding by the infrastructure First Tour CapitalThe company announced on Tuesday. Financing comes only three months after completing y combination Spring 2025 gangPlay in the AI Agent market, which is rapidly expanding investor appetite for infrastructure.

The San Francisco-based company is the current generation of the current creator – betting – Amazon Web Services, Google Cloudand Microsoft Azure – The autonomous EU systems that can move without human intervention are not in the root for a new wave. This AI agents that manage everything from managing the calendars require more than traditional web applications for human users.

“The existing cloud providers are designed for software such as 2.0, service period,” he said. “But the agent is a new infrastructure dedicated to the AI agents, which is believed in this new wave of AI.”

Why didn’t be built for AWS and Google Cloud Autonomous AI agents

Timing, customer service, data processing and operation and processing and operating AI agents reflect a wider range of enterprise calculations to use it. Unlike traditional applications in the predictable patterns of the database in predictable patterns, AI agents have joined the language models in one region and other clouds and answer knowledge bases and answer users immediately.

Blexel During the 16 global districts, it demonstrated a significant traction, processing millions of agents throughout 20 global districts. An agent of more than 1 billion seconds to take millions of videos, which is a scale of 1 billion seconds, which is a scale of AI-first companies showing infrastructure requirements.

“One of our clients process the session replages to ensure how the product managers understand how the product’s behavior is the behavior.” “Every month, millions of sessions must act with the replay. Thus, they represent the sessions in millions of minutes. Use our agent infrastructure to process the replages of this session and provide information for product managers.”

Instead of demanding human managers of AI agents, the company’s approaches that provide the infrastructure of companies. This includes Sandboxed virtual machines, which are included in the APIs designed by automatic scaling and consumption by AI systems by automatic scale and manual systems based on the agent’s action patterns.

How six co-founders with a successful exit plan will take great technology

The structure of the unusual six founder of Blaxel is the team’s shared experience building and a previous company OvehcloudEurope’s largest cloud provider. The company was the whole analytical product of the OVH, which has the first experience of the team with both cloud infrastructure problems and successful performances.

“I know that the unusual, great team sounds. We did not fit exactly on the stage of the demo day,” Sinaï said: “Sinaï said. “But we have already done it. I was six co-founders for my previous company I sold to the Ove cloud.”

In the team Sinaï, for more than 30 years, co-founders include Christophe Plojoux, Nicolas Lecomte, Thomas Crochet and Charles Drapier, who recognized more than 30 years with Mathis Joffre. Their collective expertise, infrastructure, developer instruments and platform engineering – critical expertise to combat technological giants with unlimited resources.

“I think it’s important to be six right now because we have a lot of ambition,” said sinitis. “What we do is build this next generation cloud calculation for this new agent period.”

Blaxel, the competition is separated from each other in the cloud infrastructure market

The cloud infrastructure market is competitive with new players targeting new players, a thousand third-third market share and modal, repetitions and runpodes. Blaxel distinguishes itself with special attention to AI agents than a model result or exercise.

“Most competitors, most of the competitors, the inference solves a very difficult problem – how you can host your model, how many of you can do these models,” he said. “But there are many people working on infrastructure for agents, and that’s exactly what we do.”

The company’s platform includes three main components: Agent Hosting for Placement of AI Systems SERVICE APIS, MCP (Model Context Protocol) Servers To connect agents to external instruments and a uniform gateway To access many AI models. The infrastructure is designed to resolve the variable reserve requirements of AI agents that can require minimal calculation power while waiting for the response but needed important sources during active processing.

Features of enterprise safety and matching the target adjustable industry

Despite the target of small AI-first companies, the Blaxel has implemented enterprise-level security measures, including SOC2 and HIPAA compatibility. The platform offers information residents of information that allow customers to restrict workloads to special geographical areas that criticize workloads in regulated industrial industries.

“For example, we present a policy framework you can add to search for a job, this agent cannot work outside the subsections of the regions,” he said. “You can add a policy to say that this agent cannot work outside the United States, so you are sure that this agent will only process information in the regions of your choice.”

This approach reflects the company’s confidence in the company that the EU has a strong infrastructure experience because they set up tomorrow’s businesses. “We believe that even the best infrastructure with young companies, even the best practices, because they will become enterprises,” he said.

Pay-permissions prices provide 50% cost savings from the traditional server

Blexel After confirming the market demand during the Joint Demand, he accepted a price model to a PER similar to the setting cloud providers. Model customers only accuse the agents when actively processed the tasks, closing the infrastructure in blank times to optimize costs.

“We are only rotating in several milliseconds and present infrastructure that closes in just one second,” he said. “So you have been working for something that your agent is actually working on something, just waiting for something.

The approach has already managed to reduce 50% of the costs compared to typical server-free solutions when processing customers with a client with a customer with a client with a client.

Gartner predicts 75% of apps will use AI agents by 2028

Investment Industrial Analysts come as predicted in the field of an AI agency. Gartner predicts 75% of the application development will cover AI agents by 2028Sinaï believes that the reception of the current enterprise remains mainly experimental.

“Most of the active companies that are currently active in production are mainly smaller companies, not yet enterprise companies,” he said. “Thus, we really focus on to serve them as large cloud providers in the past.”

How to strategy mirrors Amazon Web Services Initially, the start and developer were directed to friendly companies before enterprise customers expand. Blaxel plans to follow a similar way using $ 7.3 million to expand the program platform until potentially specialized and data center optimization.

“Seven million data centers are not enough to build, but I think it is important to go step by step,” Sineï said. “Be sure to provide our best services for our customers, agents, and then we have the best interfaces we will enter in deep infrastructure optimization.”

The company includes features such as Roadmap, agent experience, automatic failure opportunities and more in-depth optimization for the mass scale expected. With the forecasts of hundreds of billion dollars of AI agents in the coming decades, Bladhel sees the opportunity to build an infrastructure for this new calculation paradism.

“We believe that there is a great economy that began around agents,” he said. “There will be a hundred billions of AI agents and today are designed for this new wave.”

Includes the participation of financing tour Y combination, Liquid, Be transferredand angel investors who share the vision of the company’s integrated agent infrastructure. As AI agents transition to experimental products, the production systems related to the production systems, the specialized approach of Blinda can manage the market share that can be converted into the next main category of cloud calculation.