Based in Toulouse, the start-up Manifest has been developing humanoid robots for the defense sector for a year, with a first very concrete use case: industrial inspection.
Don’t be fooled by his name: Charles-Henri Blanchet is American. But after more than twenty years spent in the Silicon Valley start-up ecosystem, the entrepreneur chose France to launch his new adventure. For the JDN, he looks back on his career, his vision of the humanoid robot market and the challenges of creating a deeptech company in France.
JDN. What exactly does Manifest develop?
Charles-Henri Blanchet. We focus on the needs of defense stakeholders, whether the armed forces or the defense industrial ecosystem. The first use case we are working on with humanoid robots is inspection. This could, for example, involve inspecting parts supplied by subcontractors or monitoring processes at the end of the production line, such as checking a tank before delivery to the customer. On the armed forces side, it is more about inspecting equipment. It is an essential task: checking the condition of grenades, weapons or other equipment before their use.
What was your background before launching Manifest?
I spent about 20 years in San Francisco and the venture capital-funded start-up ecosystem. Manifest is my sixth company. I first worked in the field of data-as-a-service, then in the space industry, notably at ICEYE, a Finnish company specializing in satellite data using radar satellites to observe the Earth. This allowed me to gain experience in both software and hardware. The four companies in which I have held leadership positions have been very successful, with a combined value of approximately six billion dollars.
What convinced you to pursue humanoid robots?
I listened to a podcast with Figure AI founder Brett Adcock. He mentioned the idea of a so-called “brownfield” opportunity. The physical world was already designed for humans. Buildings, hospitals, offices and even our homes are organized around human anatomy and capabilities. Humanoid robots can therefore be integrated directly, without the need to reconstruct the entire environment. This idea seemed very relevant to me. He also talked about the price he was targeting for these robots: around $30,000 per unit. When I heard this, I was surprised. Instead, I imagined machines costing almost a million dollars each. Initially, I didn’t necessarily think about creating a humanoid robot company. It’s a huge ambition, a bit like announcing that you want to launch a car company.
Why did you choose inspection in the defense sector as a use case?
The first thing I did was study the supply chain in Europe. I wanted to know if it was really possible to build this type of robot using European suppliers. In Europe, we know how to manufacture batteries, actuators, motors or energy systems. These are components that the automotive industry already produces, particularly in France and other European countries.
Then I looked at potential customers. I studied different markets, which led me to the inspection use case. Inspection already involves humans moving through industrial or technical environments. You can almost think of this as a mobile sensor platform: a person walking and observing. For a humanoid robot, this is therefore a simpler use case than many others.
At what stage is the development of your robot today?
We are currently building a first prototype. It will be a humanoid torso equipped with arms, hands, a head and several types of sensors. The robot will notably use RGB-D cameras, high-resolution optical sensors as well as thermal sensors. We will also use very advanced hands with multiple fingers to allow precise manipulations. We are currently building the torso, head and chassis, before integrating all of the sensors, vision systems and electronics into the structure. The objective is to carry out a first demonstration in less than six months.
What sets your technological approach apart?
In the defense industry, we have identified approximately twenty different inspection types. Today, humans already use around twenty sensors to carry out these checks. This includes, for example, eddy currents, ultrasonic inspection, optical vision and even thermal imaging. Our idea is to integrate these sensors directly into the robot’s hands. We are also developing an “inspection brain” separate from the humanoid brain itself. This system will merge data from different sensors to produce analyzes and support decision-making.
What will be your economic model for marketing these robots?
We do not want to sell the robots. Our goal is to offer a Robot-as-a-Service model. In this model, we remain responsible for the robot and its performance. Customers pay a subscription to use the technology. Leasing allows companies to not view these robots as a capital-heavy investment, but rather as a service. The cost is expected to be around 60% of the cost of a human employee. But a robot can work almost 24 hours a day, allowing it to produce two to three times as much work.
Why did you choose Toulouse to install Manifest?
I really liked France. The people, the culture… although some aspects can be a little confusing for an American. When I arrived, I already spoke a little French, which helped me integrate. I then traveled to several regions of the country before choosing Toulouse. Toulouse has a very dense industrial fabric. Major players like Airbus are based there and also work on defense-related programs. For me, Toulouse is one of the best cities in the world for developing this type of business.
Is creating a humanoid robotics company in France an advantage or a challenge?
Working in France was a big challenge for me. At the beginning, I arrived with a very American mindset. I thought I could just say, “This is what I’m going to build. Trust me.” In the United States, this often works. But I quickly understood that in France this model does not work in the same way.
So I changed strategy. Rather than raise a lot of capital up front, I chose to build a simple prototype first. I also learned to work with public actors like Bpifrance. Programs like French Tech can play an important role in a company’s credibility. So we are moving forward step by step.
How do you judge the French AI and robotics ecosystem today?
France has a very interesting ecosystem, with both software and hardware skills. Companies like Mistral AI have managed to raise very significant amounts, which shows that there is strong momentum. I think there are many opportunities for collaboration with French companies, whether large groups, SMEs or startups.
