The first thing you notice about Safe Pro Group is the unit of account. It is not a valuation or a revenue multiple. It is a count of images, a tally of hectares, a number of detections. The company’s homepage ticks them off: over 1.6 million drone-based battlefield images processed, more than 36,000 landmines and pieces of unexploded ordnance identified [Safe Pro Group, 2026]. In this case, the machine is an AI trained to spot concealed explosives from the sky, and the market is a grim one: the estimated 110 million landmines still buried in over 60 countries [Safe Pro Group, 2026].
Based in Aventura, Florida, and publicly traded on the Nasdaq under SPAI, Safe Pro Group sells a suite of hardware and software to defense, homeland security, and humanitarian clients. Its core bet is that proprietary computer vision, trained on that massive and growing image library, can detect threats faster and safer than human-led field surveys. The company wraps this AI in several products: the SPOTD (Safe Pro Object Threat Detection) software platform, the NODE-X miniaturized edge processor for drones, and a ballistic protection division called Safe-Pro USA [Safe Pro Group].
The dataset as a defensive moat
In any AI-driven business, the quality and scale of the training data is the first line of defense. Safe Pro’s claim to over 1.6 million analyzed images and 36,000 confirmed detections is not just a marketing metric; it is the foundation of its technical pitch. The company says its models can identify more than 150 types of mines and UXO with over 90% accuracy [Safe Pro Group, 2026].
The data did not come from a lab. It was gathered through real-world operations, including work by its Airborne Response subsidiary, which holds a U.S. Department of State contract [Yahoo Finance, May 2025]. Each new image from a conflict zone or a former battlefield improves the model. This creates a feedback loop that is difficult for a new entrant to replicate quickly.
Why the government is the first customer
The sales motion here is inherently institutional. The buyers are governments, militaries, and large NGOs tasked with demining and counter-IED operations. Safe Pro has structured itself to meet them. The company holds a U.S. General Services Administration Multiple Award Schedule contract, a crucial vehicle for selling to federal, state, and local agencies [The US News]. It has also entered a Master Teaming Agreement with an unnamed U.S. government prime defense contractor to pursue AI-powered solution deals [Globenewswire, 2026].
Early traction signals are emerging from this channel. In 2026, the company landed a $1.3 million Army subcontract to integrate its AI landmine detection onto autonomous vehicles [South Florida Business Journal, 2026]. The team, led by CEO Daniyel Erdberg, is reportedly enriched by professionals from the Intelligence Community and Department of Defense [Globenewswire, 2026].
Funding the public company pivot
Safe Pro’s journey to the public markets was an early one, with an IPO in 2024 that raised approximately $5.1 million [Stockanalysis, 2026]. But its recent growth capital has come from strategic private placements. In 2025, it closed a $14 million strategic investment led by ONDAS Holdings, with participation from drone company Unusual Machines [Safe Pro Group, May 2025]. This was followed by another undisclosed round with ONDAS in August 2025 [Safe Pro Group, August 2025].
| Metric | Value |
|---|---|
| 2024 IPO | $5,100,000 |
| 2025 Private Placement | $14,000,000 |
| 2025 Strategic Round | Undisclosed |
Where the wheels could come off
For all its concrete metrics, Safe Pro Group operates in a high-stakes, slow-moving, and fiercely competitive arena. The risks are not subtle.
- The validation gap. While the company cites 90%+ accuracy, independent, third-party validation of its detection rates in varied, real-world conditions is not publicly available.
- The integration burden. Selling AI software to the military means integrating with legacy hardware, secure networks, and existing doctrine.
- The competitor set. Safe Pro does not name direct competitors in its materials, but the space for drone-based ISR and counter-IED tech is crowded with both private startups and defense primes.
The company’s answer to these risks seems to be its partnership strategy and its focus on a hardened, edge-deployable product like NODE-X, which processes footage on the drone itself without needing a constant data link [Globenewswire, 2026].
The next twelve months
The key milestones for Safe Pro are likely to be commercial, not technical. Watch for announcements of follow-on contracts from the Army autonomous vehicle pilot or new teaming agreements stemming from the Master Teaming Agreement. The company will also need to demonstrate that its ballistic protection division, Safe-Pro USA, can achieve meaningful scale and synergize with its AI business, rather than remaining a disparate product line.
Traditional manual clearance can cost between $300 and $1,000 per mine, and a single deminer can clear 20 to 50 square meters per day. If Safe Pro’s AI can help a drone survey team accurately map threats ten times faster, the cost savings in personnel time and risk avoidance are immense. The company’s real competition is not another software startup. It is the status quo: men and women in protective suits, probing the ground with metal sticks, meter by terrifying meter.