Anduril Industries has introduced two autonomous aerial vehicles with AI capabilities, Bolt and Bolt-M, as the defense contractor expands its lineup of military and surveillance technology. The company presents both systems as easier to operate than manually piloted armed FPV drones, with artificial intelligence handling much of the flight and tracking workload.
The announcement places the focus on a practical military problem: how much attention an operator must spend flying the drone, and how much can instead be spent on mission decisions. Anduril says its AI approach automates "thousands of manual flight inputs required to effectively track objects of interest, enabling operators to focus on key decision-making over navigation."
What Anduril says Bolt is built to do
Bolt is the reconnaissance version of the new drone family. According to the source article, it has a range exceeding 20 kilometers, can remain airborne for over 45 minutes, and weighs about 5.4 kilograms.
Anduril says Bolt can be used for reconnaissance as well as search and rescue missions. That makes it the non-ammunition member of the pair, designed around observation, tracking and mission support rather than direct attack.
The key feature is not only that Bolt can fly at distance or stay in the air for a meaningful period of time. The company is emphasizing that the system uses AI to reduce the manual effort involved in tracking an object of interest. In a drone mission, that distinction matters because operating the aircraft and interpreting what it sees can compete for the operator's attention.
By presenting Bolt as easier to operate, Anduril is framing autonomy as a way to lower the burden on the person controlling the system. The source article also notes that the drones are designed for safe handling, quick deployment and effective operation with minimal training.
How Bolt-M changes the role
Bolt-M is the ammunition variant. Anduril describes it as a system meant to provide ground troops with "simple, flexible and lethal precision firepower." It shares similar range and endurance specifications with Bolt, but its weight depends on payload, ranging between 5.9 and 6.8 kilograms.
The company says Bolt-M can attack from any angle, including directly from above. Anduril states that this allows the drone to strike weak points of targets in complex environments with high precision.
That positioning separates Bolt-M from Bolt in both purpose and stakes. Bolt is presented as a reconnaissance and search and rescue platform. Bolt-M is presented as a weaponized option intended to deliver precision firepower for ground troops.
The two drones therefore appear to share a common autonomy and tracking foundation, while serving different mission categories. One is focused on seeing and tracking. The other is focused on using those capabilities as part of an ammunition system.
AI-assisted tracking is the core claim
Both Bolt and Bolt-M feature advanced computer vision and machine learning software. The source article says this enables operators to track targets from user-defined standoff positions and maintain tracking even when the target is obscured.
Those claims are central to how Anduril is marketing the drones. The company is not only describing new airframes or payload options. It is describing a shift in the operator's role, where AI assists with navigation and target tracking so the operator can concentrate on decisions.
The source article highlights several related capabilities:
- Bolt has a range exceeding 20 kilometers.
- Bolt can stay airborne for over 45 minutes.
- Bolt weighs about 5.4 kilograms.
- Bolt-M weighs between 5.9 and 6.8 kilograms, depending on its payload.
- Both drones use computer vision and machine learning software.
- Both are presented as easy to handle, quick to deploy and usable with minimal training.
In plain terms, the company is trying to make the case that AI can reduce the difficulty of operating drones in demanding missions. Instead of requiring constant manual inputs for flight and tracking, the system is designed to automate a large share of that work.
Where Bolt and Bolt-M fit in Anduril's portfolio
The source article places Bolt and Bolt-M within Anduril's broader history of military and surveillance systems. It notes that the company has developed the Anvil drone, which autonomously identifies and attacks targets at high speeds.
It also notes that Anduril's Ghost surveillance drone is already in use by the US military. In addition, Anduril is involved in the U.S. Air Force's Collaborative Combat Aircraft program, also called CCA.
The CCA program aims to develop highly autonomous drones that work alongside manned aircraft as part of the Air Force's future strategy. That context matters because Bolt and Bolt-M are not isolated products in the source article. They are part of a broader push toward drones that rely more heavily on autonomy, computer vision and machine learning.
Anduril's message is consistent across the new drones: autonomy should make aerial systems simpler to use, faster to deploy and more capable in tracking targets. With Bolt, that message is tied to reconnaissance and search and rescue. With Bolt-M, it is tied to precision firepower.
The bigger takeaway
The unveiling of Bolt and Bolt-M shows how AI is being presented inside defense technology: not as an abstract software layer, but as a way to change how operators fly, track and make decisions during missions.
The source article does not provide independent testing results for the drones. It reports Anduril's specifications and claims, including range, endurance, weight, tracking capability and intended roles.
Based on those claims, the central development is clear. Anduril is packaging AI-assisted autonomy into small aerial systems meant to be deployed with limited training, while separating the product line into reconnaissance and ammunition variants. Bolt is built around observation and mission support. Bolt-M adds the firepower role, with Anduril emphasizing flexible attack angles and precision against weak points in complex environments.