For most of us, drone delivery still sits somewhere between science fiction and a clever marketing demo. We imagine drones being hit by birds, knocked down by stones, stolen after landing, or simply falling out of the sky when the internet disappears.
I had the opportunity to unpack these questions with Ankit Kumar, Founder and CEO of Skye Air Mobility, a pioneer in India’s drone logistics ecosystem. Ankit and his team have completed millions of drone deliveries, building autonomous delivery infrastructure for healthcare, e-commerce, and quick commerce use cases across India.
What came through clearly in our conversation is that drone delivery is not just about putting a package on a flying machine. It is an operating system made up of aircraft, airspace management, communication redundancy, delivery infrastructure, safety protocols, and partnerships. The drone is only one component.
Table of Contents
The Real Questions Behind Drone Delivery
Myth 1: A Stone Can Knock Down a Delivery Drone
Myth 2: A Drone Can Be Stolen at the Delivery Address
Myth 3: A Drone Is Lost When Internet Connectivity Fails
Bird Strikes Are Real, and So Is Learning From Them
What Happens When Another Drone Enters the Flight Path?
Why Drone Delivery Can Be Cheaper Than Rider Delivery
India’s Position in Global Drone Delivery
The Future Is an Ecosystem, Not a Single Company
Drone Delivery Will Become Mainstream When the Invisible Infrastructure Works
The Real Questions Behind Drone Delivery
I wanted to begin with the questions any ordinary customer would ask. Not polished investor questions. The practical, slightly sceptical ones.
Can somebody throw a stone at a drone? Can it get stolen? What happens when the mobile internet fails? Do bird strikes happen? Can two drones collide?
These questions may sound simplistic, but they get to the heart of whether autonomous logistics can work in the real world. Ankit’s answers made one thing clear: safety in drone logistics is not based on assuming nothing goes wrong. It is based on designing multiple responses for when something does.
“Many times you do not even realise that the drone is flying over your head because it is so silent.”
Ankit Kumar
Myth 1: A Stone Can Knock Down a Delivery Drone
According to Ankit, this has not happened in Skye Air’s operations. The reason is straightforward. Delivery drones fly at an altitude of around 120 metres, far above what someone on the ground can casually reach with a thrown object.
To hit one at that distance, a person would need to be an exceptionally capable athlete. More importantly, these drones are not constantly hovering low over streets or landing at random locations. Their routes, altitudes, and landing zones are part of a planned system.
This is a useful correction to the popular image of drone delivery. We tend to imagine a machine navigating between buildings at head height. In a real logistics operation, the flight occurs much higher, within designated corridors, and the final handoff is designed for controlled infrastructure.
Myth 2: A Drone Can Be Stolen at the Delivery Address
This concern disappears once we understand the delivery model. Skye Air does not land a drone at a customer’s doorstep and leave it exposed. The drone lands only at authorised locations operated within the network.
At the destination, it drops the package into a secure mailbox or skypod and returns. The drone itself does not wait on the customer side for someone to collect it.
That distinction matters. The delivery endpoint is not just an address. It is a purpose-built, secure node in the network. This is precisely why drone logistics requires infrastructure, not just better flying hardware.
Myth 3: A Drone Is Lost When Internet Connectivity Fails
Drone operations cannot depend on a single communications channel. Skye Air works with a layered connectivity approach:
Primary communication: 4G and 5G connectivity.
Secondary communication: Radio frequency communication.
Third option: SATCOM, or satellite communication, in locations where the other two options have limited use.
The system automatically attempts to regain the primary connection for 60 seconds. If that fails, it switches to the secondary communication layer. In more challenging territories, including some operations in Himachal Pradesh, satellite communication has been necessary.
The Redundancy Framework for Autonomous Delivery
I found this to be one of the most important operational lessons from Ankit. Reliability does not come from a single perfect technology. It comes from designing fallback layers.
Layer 1: Connectivity
4G and 5G enable day-to-day communication.Layer 2: Backup control
Radio frequency takes over when primary networks fail.Layer 3: Remote resilience
SATCOM supports operations where terrestrial networks are unreliable.Layer 4: Automated response
The system switches communication channels without waiting for manual intervention.
That is the broader principle for any autonomous system. Do not build for the happy path alone. Build for failure, detection, switching, and recovery.
Bird Strikes Are Real, and So Is Learning From Them
Bird strikes are not a myth. Ankit was candid that Skye Air had a bird incident during medicine deliveries in Mandi, Himachal Pradesh. A flock of eagles attacked the drone, and that is not an event that can be fully controlled in the moment.
But the learning after the incident was more valuable than pretending the risk did not exist. The team began studying bird migration patterns, particularly during morning and evening hours. They looked at how flocks moved across specific areas and where they returned in the evening.
That intelligence now informs the design of flight tunnels and air corridors, helping ensure routes do not intersect with bird migration zones.
“The corridors that we make are made in a certain way that they do not cross bird migration zones.”
Ankit Kumar
There is an important masterclass here for founders building in the physical world. A difficult incident is not only a problem to solve. It can become operational data that makes the system stronger. Since that major bird incident roughly two to two-and-a-half years ago, Ankit shared that Skye Air has not had another one.
What Happens When Another Drone Enters the Flight Path?
Collision risk is another serious question because airspace is not empty. Skye Air experienced a near miss when one of its drones encountered an unauthorised DJI drone in the same corridor. Someone had apparently launched it for aerial imaging, and it came into the path of an active logistics operation.
This is where traffic management becomes essential. A delivery drone uses proximity sensors and LiDAR to detect obstacles. When an obstacle is detected, the drone holds its position for up to 120 seconds.
If the obstacle clears, the drone continues its route. If it does not, the drone diverts to an alternate recovery or landing site. It lands safely and triggers an alarm, after which pilots take action to recover the aircraft.
India’s larger challenge is that unauthorised drone activity is still difficult to monitor. As Ankit pointed out, there is no complete mechanism today for identifying every drone that may be flying in a particular area. That makes a robust traffic-management layer central to the future of drone logistics.
Why Drone Delivery Can Be Cheaper Than Rider Delivery
The assumption that drone delivery must be expensive is perhaps the biggest misconception of all. Ankit’s point was simple: cost follows efficiency.
A drone flight has a relatively fixed operating cost. If that same flight carries more shipments, the cost per shipment comes down. The economic question is not only what the drone costs. It is how many deliveries the system can produce per hour, per pilot, and per flight.
The economics change quickly when delivery capacity, batching, and return time are considered together.
Quick Commerce: The Efficiency Comparison
Consider a typical quick commerce rider. Without batching, a rider may need 10 minutes to go to the customer and 10 minutes to return. That means roughly one delivery every 20 minutes, or a maximum of around three deliveries an hour before accounting for breaks.
Over a 10-hour day, practical output can fall to around 25 deliveries once regular breaks are factored in.
Now compare that with a drone. Ankit explained that a drone can travel out in roughly two minutes and return in another two to two-and-a-half minutes. Even after allowing a four-minute turnaround between deliveries, the operation can complete roughly one delivery every 10 minutes.
That is around six deliveries an hour per drone. One pilot can manage two drones, which creates a significant improvement in human productivity as well.
Batching Changes the Equation
The more interesting insight is that consumers do not necessarily need a delivery in three minutes. If the expectation is 10 minutes, a drone operation can wait for 60 to 90 seconds and batch orders within a dense local cluster.
Instead of one shipment, a single flight may carry two or three shipments. In six flights per hour, that can mean nine or 10 deliveries, versus the rider’s three deliveries over the same period.
“If your efficiency of the system is higher, your cost will be on the lower side.”
Ankit Kumar
Without scale, Ankit estimates drone delivery can be 10% to 15% cheaper. At scale, the advantage can rise to 20% to 25%, while also being faster and more reliable.
The Drone Logistics Efficiency Framework
Shorter travel time: Air routes can reduce delivery time substantially.
Faster return cycles: The drone comes back quickly for the next flight.
Battery swaps instead of rest periods: The machine does not sleep, it is recharged and redeployed.
One pilot, two drones: Human oversight scales across aircraft.
Dense-cluster batching: Multiple packages on one flight lower per-order cost.
India’s Position in Global Drone Delivery
India began later than markets such as the United States, China, and parts of Africa. China is currently the largest market for drone deliveries, followed by the United States, including US companies operating drone delivery networks in Africa.
Yet India has accelerated rapidly. Ankit credits policy changes and a more streamlined regulatory environment for helping the sector move faster without excessive bureaucracy.
His view is bold but grounded in the opportunity ahead: India is already the third-largest country in the world for drone deliveries, despite operations being concentrated in only a small number of cities and with very few companies operating at scale.
The potential becomes much more significant when drone delivery reaches 20 or 25 cities. India has dense urban clusters, growing quick commerce demand, challenging terrain in several regions, and an increasing need for efficient last-mile logistics. Those conditions can create a powerful operating environment for autonomous delivery.
The Future Is an Ecosystem, Not a Single Company
A particularly valuable part of my conversation with Ankit was his philosophy on collaboration. Skye Air has partnered with Arrive AI, a NASDAQ-listed American company, and Autonomy. Together, the companies are part of the NVIDIA Inception Program.
Skye Air already operates skypods, but future logistics needs multi-compartment units. As quick commerce, food, healthcare, and multiple brands use the same delivery infrastructure, each participant needs secure access to its own packages.
Imagine separate compartments for different brands. A delivery worker enters an OTP, and only the correct compartment opens for the intended pickup. Skye Air does not need to own every last touchpoint. It can provide the drones, airspace access mechanisms, and systems integration while specialist partners contribute smart delivery points, rovers, and autonomous technology.
The next delivery network will connect aircraft, secure handoff points, and ground systems into one coordinated flow.
The integrated model looks like this:
Skye Air: Drones, flight operations, and unmanned traffic management systems.
Smart arrival points: Secure, connected mailboxes and multi-compartment handoff infrastructure.
Rovers: Ground-based movement systems for the next stage of fulfilment.
AI and compute: GPU-driven intelligence that helps connect and coordinate the layers.
Brand partners: The companies whose products and workers interact with the delivery network.
“Let them do what they are best at. Let us do what we are best at.”
Ankit Kumar
This is a lesson beyond drones. Founders often confuse owning the full stack with building the best solution. Ankit’s point is that if a partner has spent six years building a strong capability, the better strategy may be to integrate it rather than rebuild it internally.
Drone Delivery Will Become Mainstream When the Invisible Infrastructure Works
The future of drone logistics will not be decided by whether a drone can fly. That part is already happening. It will be decided by whether the full system can operate safely, reliably, and economically across cities.
That means protected air corridors, traffic management, backup communications, secure handoff points, battery operations, smart batching, regulatory clarity, and an ecosystem of partners doing what each does best.
The biggest takeaway I had from Ankit is that the strongest technologies become ordinary only after the complexity underneath them becomes invisible. A customer should only see a fast, dependable delivery. Behind that simple experience sits a deeply engineered autonomous logistics network.
I am Saurabh Agrawal and we come with a new episode on Dilse omni talks every fortnight and cover different aspect of omnichannel with amazing speakers.
