As the drone industry pushes toward greater autonomy, scalability, and integration into shared airspace, its dependence on advanced semiconductors is becoming impossible to ignore. In this guest post, Michelle Duquette of 3 MAD AIR explores why closer collaboration between the semiconductor and aviation communities is essential to address certification, supply chain resilience, and long-term operational safety. Her perspective highlights how aligning chip design with aviation requirements today can help avoid costly barriers to drone deployment tomorrow. DRONELIFE does not make or accept payment for guest posts.
When Chips Take Flight: Building the Semiconductor-Drone Partnership
by Michelle Duquette, Founder and CEO of 3 MAD AIR Consulting.
Last week, I had the privilege of joining a panel at the Northeast Semiconductor Manufacturing Corridor Workshop in New York City-a gathering that brought together semiconductor manufacturers, quantum computing companies, and photonics experts from across the U.S.Northeast and Canada. Canadian businesses,supported by the Canadian Consulate, were particularly active in the discussions. It was apparent that North American semiconductor and aerospace collaboration doesn’t stop at the border. I was the only aerospace professional in the room, and I went specifically to open a conversation between two sectors that desperately need each other but rarely talk.
Panel “Integrated Circuit Design from a Defense Industry Perspective.”
The panel, titled “Integrated Circuit Design from a Defense Industry Perspective,” brought together quantum computing innovators and AI chip designers, and me, an aviation strategist who spends her days helping operators, regulators, and technology providers make remotely piloted and autonomous aircraft operate safely in shared airspace. What became clear over the course of the conversation is that drone aviation and semiconductor manufacturing are two industries that seriously need each other, and that the challenges we’re facing are interconnected.
The Foundational Dependency
Let me be direct: the drone industry cannot scale without trusted, aviation-grade semiconductors. Every capability we’re trying to enable, from autonomous flight controls, AI-powered decision-making, secure communication, or detect-and-avoid depends entirely on chips that can perform under strict Size, Weight, Power, and Cost (SWaP-C) constraints while meeting aviation certification standards. This is about building a truly interoperable National Airspace System from the ground up.
And the timing couldn’t be more critical. The FAA’s Brand New Air Traffic Control System (BNATCS)-a $12.5 billion effort to rebuild America’s air traffic infrastructure by 2028-is being architected right now. If the semiconductor and aviation communities don’t align today, we risk building a system that’s optimized for legacy manned aircraft but can’t natively handle the heterogeneous traffic mix of the 2030s. We need unified data architectures and open standards baked into BNATCS Phase 1, not bolted on later as an afterthought.
And here’s what the semiconductor industry needs to hear from us in the aviation industry: you can’t just take the latest gaming GPU off the shelf and call it flight-ready. Aviation demands provenance, secure-by-design architectures, long-term lifecycle support, and chips that won’t be obsoleted on consumer product timelines. When a public safety drone fleet is grounded because a critical radio chip reached end-of-life with no alternative source, it surpasses inconvenience and immediately becomes a national security vulnerability. Drones deployed for good, like public safety, disaster response, and critical infrastructure inspections can’t afford to fail because of supply chain fragility.
Common Problems, Uncommon Collaboration
As the panel unfolded, I kept hearing echoes of my own interoperability challenges reflected in what the semiconductor folks were describing.
I see the same kinds of gaps every day in my work. I’m helping states build AAM sandboxes that need trusted semiconductor supply chains. I’m working with operators and service suppliers who can’t find domestic aviation-grade chips with predictable lifecycles. I’m connecting technology providers who don’t even realize the other exists. And I’m watching the FAA spend countless hours sorting out safety cases for aircraft whose chip designs don’t quite meet mission requirements, resulting in additional risk assessments, lists of compensating mitigations, and certification-related one-offs that could have been avoided if the semiconductor and aviation communities had been talking early on.
That regulatory burden isn’t just an FAA problem. It literally taps the brakes industry-wide on innovation. Every time a chip is designed without understanding aviation constraints, or an aircraft integrates components without anticipating national security or certification requirements, the FAA has to untangle the implications. Is the processing latency acceptable for detect-and-avoid? What happens when the chip reaches end-of-life mid-certification? How do you validate an AI accelerator’s behavior under all flight conditions? This is our daily reality slowing down approvals and driving up costs for everyone.
• Supply chain fragility: Both industries depend on layered infrastructure-raw materials, fabrication, packaging, testing, integration. A weakness anywhere in that chain breaks the whole system. In semiconductors, it’s rare-earth supply and fab capacity. In drones, it’s aviation-grade components that can’t be single-sourced from adversarial suppliers.
• The valley of death: Promising prototypes that never become certified, deployable products. Semiconductor companies struggle to navigate defense acquisition processes. Drone operators struggle to move from one-off approvals to scalable, repeatable operations. Both industries are stuck in bespoke validation cycles that are too slow and too expensive.
• Export controls and dual-use complexity: High-performance chips, especially AI accelerators, fall under ITAR and EAR restrictions, just like advanced drone radios, sensors, and flight controllers. If you don’t design for compliance from day one, you risk building systems that can’t be deployed internationally or shared with key partners.
• Defense-adjacent vs. defense-only development: Do you build separate systems for commercial and military applications, or design one trusted platform that serves both? The semiconductor industry already knows the answer: they don’t build one-off prototypes for a single customer. They develop scalable, validated processes that serve multiple markets. That’s exactly the model we need in aviation.
Precision Manufacturing Meets Precision Operations
One analogy kept surfacing during the discussion: both industries operate in zero-margin-for-error environments.
In a chip fab, a single contamination event or process deviation can ruin an entire wafer batch. In aviation, a single counterfeit component or assembly error can ground a fleet or cause catastrophic failure. Both require extreme precision, cleanroom-level process discipline, rigorous traceability, and batch testing. The Northeast’s semiconductor heritage of precision manufacturing, defense electronics, and materials science maps directly onto the requirements for building trusted, certifiable drones.
Unfortunately, today we’re optimizing these constraints in isolation, and that’s what breaks us. Semiconductor companies optimize for performance per watt without understanding mission profiles. AI teams optimize for compute without understanding drone flight dynamics. Aircraft integrators optimize for weight and power without understanding what the chips actually need to deliver the mission or the realization they can influence those designs. So we end up with chips that are too power-hungry for the airframe, AI accelerators that demand computing power the platform can’t support, and flight controllers that can’t handle the processing latency requirements for detect-and-avoid.
SWaP-C isn’t just about making a chip smaller or more efficient. It’s about understanding that flight controllers, sensors, radios, batteries, and AI accelerators all compete for the same power envelope and physical space on an airframe, and that every decision in one domain constrains what’s possible in another.
What Real Collaboration Looks Like
From my view, the solution is co-design from day one.
Get chip designers, AI engineers, drone integrators, and operators in the same room. Map the mission requirements first, then work backward to the SWaP-C tradeoffs. That’s how you avoid over-speccing a system that can’t fly or under-delivering a capability that doesn’t meet the mission. Real collaboration means building strategic partnerships where everyone shares a common vision of success. It’s interoperability.
During the panel, we talked about what this could look like in practice for the Northeast corridor-and the broader North American ecosystem:
• Shared sandboxes and qualification pipelines where AI hardware and software get validated together under real aviation conditions and not in separate silos.
• Co-located development where semiconductor designers, AI labs, drone and eVTOL manufacturers, operators, and regulators iterate on certification, safety cases, and mission performance in real time.
• Cross-border consortia where chipmakers, integrators, and aviators jointly map critical components and identify single points of failure before a crisis hits.
• Dual-use test environments that serve as “fabs of the drone ecosystem”: shared infrastructure that reduces barriers to entry, accelerates validation, and enables small companies to compete.
Starting with Real Missions
Here’s my ask to the semiconductor community: Don’t wait for the perfect abstract use case.
Pick a real mission, like a North Country winter storm response, maritime search and rescue off the coast of Maine, critical infrastructure inspection in Quebec, and build the stack together. Prove that AI, semiconductors, and aviation can deliver tangible public benefit and dual-use value in an operational environment with real accountability.
These are actual deployments with real aircraft operators, real airspace constraints, and real consequences if the technology doesn’t work as advertised. This is operational pragmatism at its core, focusing on proven technology that solves real problems.
And when you do this, apply the same discipline that gets you from fab to flight-ready: precision at every layer, traceability, redundancy, and proactive risk management. The aviation community knows how expensive it is to retrofit resilience or interoperability. Design for it from the start.
An Invitation
The Northeast Semiconductor Manufacturing Corridor has a once-in-a-generation opportunity to be the place where two industries that desperately need each other stop talking past one another and start building together.
The challenge: semiconductor companies don’t know which aviation regulations apply to their products or how to map their roadmap to certification pathways. Drone operators can’t articulate SWaP-C requirements in language chip designers understand. States and universities building test sandboxes need both sides at the table but don’t always know how to convene them effectively. And everyone is burning time and money on integration attempts that drag on or fail because no one is translating between domains.
This is exactly the bridge-building work I do. I spent 24 years inside the FAA’s technical architecture, and now I work daily with the drone and eVTOL community scaling real deployments, states standing up interoperable drone and AAM ecosystems, and technology providers trying to navigate certification and supply chain resilience. I translate between regulators, operators, and technology providers so you don’t have to figure it out alone and so the FAA doesn’t have to untangle preventable risks after the fact.
If you’re a semiconductor company trying to understand what “aviation-grade” really means for the drone world, or a drone operator frustrated by chip suppliers who don’t understand your constraints, or a policymaker wondering how to accelerate both industries, let’s come to the table.
The problems are hard, but they’re solvable. And none of us can solve them alone.
Because at the end of the day, chips don’t fly themselves. And drones don’t work without the right chips.
It’s time we figured this out together.
Michelle Duquette is CEO of 3 MAD Air, a consulting firm specializing in drone and advanced air mobility interoperability. With 35 years in aviation-including 24 years at MITRE Corporation resulting in leading the FAA’s UAS and AAM research portfolio-she helps bridge the gaps between operators, regulators, technology providers, and policymakers to enable safe, scalable operations in shared airspace.
Read more:
