Can AI Help America Solve Its Engineering Shortage?

In April 2019, the DSV Limiting Factor, built by Triton Submarines, descended to roughly 35,800 feet below sea level—the deepest point humans have ever reached. It touched down at Challenger Deep, the deepest known part of the ocean.

That is nearly seven miles beneath the surface, deeper than Mount Everest is tall. Yet even that extraordinary achievement represents only a fraction of what remains to be explored. According to NOAA’s April 2026 figures, more than 70% of the ocean floor still awaits high-resolution mapping.

Exploring the remaining depths is easier said than done. Before engineers can send people or machines into these environments, they have to design vehicles and systems capable of surviving enormous pressures. At the deepest parts of the ocean, pressure exceeds 1,000 times that of the atmosphere at sea level.

The ocean is only one of humanity’s frontiers. Space presents another. Reaching either one requires physical systems that can operate reliably under extreme conditions.

That is where simulation and digital twins are becoming increasingly important. Engineers can use them to study how a vehicle, component, or entire system might behave in dangerous environments before putting people, expensive equipment, or prototypes at risk.

As McKinsey explains, a digital twin is a digital representation of a physical object, system, person, or process that is placed within a digital version of its operating environment. The idea is simple: simulate reality before committing resources to the real thing.

That approach could do more than help humanity explore places it has never reached. It could also help the United States rebuild its engineering and manufacturing capabilities at a time when the country is trying to make more advanced products domestically.

Simulation Could Help America Build Again

This is the opportunity Masha Petrova sees in modern simulation technology.

“Simulation is sort of the hidden area, but it’s critical in order for us to manufacture again. Without that work, nothing happens, or we keep manufacturing things that are outdated, and innovation doesn’t occur.”

Petrova is CEO of Nullspace, a deep-tech company developing electromagnetic simulation software for mission-critical radio-frequency (RF) and quantum computing applications.

Her point goes beyond simply making engineering calculations faster. Simulation can change how companies approach product development in the first place.

Instead of building a physical prototype, testing it, discovering a problem, redesigning it, and starting over, engineers can test many possibilities digitally before manufacturing anything. That can reduce wasted materials, shorten development cycles, and allow smaller teams to explore more design options.

But there is another obstacle standing in the way of America’s manufacturing ambitions: people.

America Faces a Growing Engineering Talent Shortage

The United States already faces a shortage of engineers, and the problem becomes more significant as engineering systems grow increasingly sophisticated.

In a 2023 analysis conducted with SAE International, Boston Consulting Group estimated that the U.S. would need approximately 400,000 new engineers every year. The analysis warned that the next-generation skills required by those engineers are in short supply, potentially leaving nearly one in three engineering positions unfilled each year through at least 2030.

That creates a difficult equation.

Companies need more engineers, but the systems those engineers are being asked to design are becoming more complicated. At the same time, experienced engineers cannot simply be produced overnight.

Antennas offer a good example.

For many people, an antenna still brings to mind the rabbit-ear television antennas of decades past. Today, antennas are embedded in smartphones, vehicles, laptops, thermostats, security systems, satellites, drones, and advanced communications equipment.

As Petrova puts it:

“Now that we have satellites in space and drones in the theater of war, and advanced communications, design challenges increase.”

Designing systems that can operate across all these applications requires specialized knowledge built over years of experience. The challenge, therefore, is not simply hiring more graduates. Companies also need tools that allow smaller engineering teams to make better use of the expertise they already have.

“We can design better engineering tools to empower leaner teams,” Petrova says.

Test Before You Build

Petrova describes an R&D process in which engineers simulate a design before building physical prototypes.

A useful comparison is the way a restaurant chain might test a new menu item.

Imagine a chain with 2,000 locations considering a new barbecue chicken sandwich. Rather than immediately adding it to every restaurant, the company could launch the sandwich at five locations in different parts of the country.

If customers love it, the company has evidence to support a larger rollout. If the product fails, the financial and operational damage is limited.

Engineering simulations follow a similar principle. Instead of committing significant resources to a physical design before knowing whether it will work, engineers can test different configurations digitally first.

That is particularly valuable when the physical prototype is expensive, difficult to manufacture, or potentially dangerous to test.

Faster Antenna and Radar Simulation

Nullspace’s software is designed to help engineers predict how antennas and radar systems will perform and interact with their surrounding environment before the physical product is built.

Petrova points to the company’s work with a major Japanese manufacturer as one example.

The manufacturer had previously used another simulation tool that required approximately six hours to model the performance of a satellite antenna.

Using Nullspace’s software, that simulation was reduced to about two minutes, while maintaining the accuracy of the results.

The difference is substantial. Cutting a simulation from hours to minutes does not simply save time on one calculation. It can allow engineers to test significantly more design variations within the same development cycle.

That means more opportunities to identify problems early, compare alternatives, and improve the final product before manufacturing begins.

AI Could Make Engineering Simulation Even Faster

AI adds another layer to this process.

Rather than replacing the engineer, AI can take over portions of the workflow that consume time without requiring the engineer’s highest-level judgment.

“AI can make workflows much faster,” Petrova told me. “For example, an AI agent can help set up a simulation, change design parameters, perform variations and organize results for an engineer to review.”

The distinction is important.

The simulation software still performs the underlying physics calculations. The engineer remains responsible for interpreting the results and determining whether they make sense in the real world.

In that model, AI acts less like an autonomous engineer and more like an engineering assistant—handling repetitive setup and analysis tasks while leaving critical decisions to the person with the relevant expertise.

That could become increasingly valuable as engineering teams become smaller relative to the complexity of the systems they are expected to deliver.

Agentic AI and the Future of Lean Engineering Teams

The idea of using AI to make scarce technical talent more productive extends beyond simulation.

I recently explored a similar theme for Forbes, looking at how agentic AI could support organizations operating with limited personnel.

Beehive Industries offers another example of a company rethinking how engineering and manufacturing should work.

The company manufactures jet engines for drones and defense applications. Instead of relying on a traditional network of separate suppliers, factories, machining operations, and lengthy development processes, Beehive is pursuing a more integrated approach.

According to Beehive Industries, more than 90% of the components in each engine it sells are produced using 3D printing, with design, machining, assembly, and testing handled in-house.

The advantage is not simply that 3D printing changes how components are manufactured. It can also reduce the time between designing a product and getting a working engine into testing.

As Tom’s Hardware reported, the approach could reduce the time required to design, test, and deploy an engine while also lowering production costs. That matters to the U.S. military, particularly when expensive missiles are sometimes used to destroy relatively inexpensive drones.

Beehive’s Approach Is Already Producing Results

The strategy has attracted attention from the U.S. Air Force.

According to Air & Space Forces Magazine, the Air Force awarded Beehive Industries a $29.7 million contract in April 2026 to complete work on a new disposable jet engine designed to power drones and munitions.

The company’s CEO, Mohammad Ehteshami, sees the engineering challenge in broader terms.

“We need more engineers, but we also need to stop wasting the expertise we already have. An experienced engineer should be solving design problems, not spending their day working around outdated processes.”

That idea connects closely with Petrova’s argument about simulation.

The answer to an engineering talent shortage may not be to simply hire more people. It may also be to eliminate the inefficient processes that prevent existing engineers from spending their time on the problems that actually require their expertise.

Better Tools Could Multiply Engineering Talent

These examples point toward a broader shift in how engineering work gets done.

Simulation allows teams to test designs before building them. Digital twins allow engineers to study how systems might behave in real-world environments. AI can automate repetitive parts of simulation and analysis. Advanced manufacturing methods such as 3D printing can shorten the distance between design and production.

None of these technologies eliminates the need for skilled engineers.

In fact, their value depends on having engineers who understand the underlying physics, constraints, tradeoffs, and real-world requirements. The technology simply gives those people more leverage.

That distinction matters as America tries to rebuild its industrial and engineering capabilities.

If the country needs hundreds of thousands of additional engineers but cannot produce them quickly enough, increasing the productivity of the engineers already in the workforce becomes critical.

Can AI Help America Build Again?

America’s ambitions extend from the deepest parts of the ocean to space and, closer to home, to rebuilding advanced manufacturing capabilities.

Whether the goal is designing vessels capable of exploring unmapped ocean floors, spacecraft capable of operating far beyond Earth, advanced communications systems, or jet engines for defense applications, the engineering challenge remains the same: build complex systems that work reliably under demanding conditions.

Simulation can reduce the number of physical prototypes required. AI can accelerate repetitive engineering workflows. Digital manufacturing can shorten production cycles. Together, these technologies can give engineers a faster path from an idea to a tested, manufacturable product.

The larger opportunity is not to replace human ingenuity with AI. It is to give skilled people better tools so they can accomplish more with the time and expertise available.

If America wants to build at the scale it once did, the country will need more engineers. But it will also need to make every engineer more productive.

That may be where AI, simulation, digital twins, and modern manufacturing have their greatest role to play.