This piece leans on numbers for a simple reason: in any industry, progress only matters when it’s measurable. Competition pushes innovation, innovation reshapes the status quo, and those shifts are what move markets. Now, let’s see what the data says.
Why are legacy industries breaking, and why now?
Legacy industries are under pressure. Industries such as construction, mining and logistics, which have long been known as laggards when it comes to technology innovation, are now encountering economic and operational pressures that require them to change in ways they’ve never had to before. Not to mention that labor shortages persist and are forcing chronic overtime.
Current trends in human resource projections
Recent workforce projections indicate these challenges will become even more pronounced in the future. Industry estimates reveal the construction sector in the U.S. is expected to seek 456,000 new entrants in the workforce by the year 2027. Each incremental $1 billion in construction spending supports the creation of 3,450 new jobs.
This challenge is deepening as hyperscale technology companies ramp infrastructure spending faster than labor capacity can adjust. Hundreds of billions of dollars in annual capital expenditures by technology companies in data centers and infrastructure could potentially become a labor constraint issue.12
Image source: Courtesy of the author; field deployment image provided by Under Control Robotics.
To take one example, the U.S. construction industry was short some 439,000 workers as of late last year, especially in skilled trades like electricians and pipefitters.1
Mining and logistics tell a similar story: a 2023 global survey found over 3 million truck driver positions unfilled (about 7% of total) across 36 countries and warned that the driver shortage could double by 2028 without intervention. 2 In mining, 71% of industry leaders say talent scarcity is already impeding on project delivery. An aging workforce (the average trucker or heavy machine operator is nearing retirement age, with only ~12% of truck drivers under 25) and declining interest in these tough jobs portend further strain.3
When demand is exploding, why does execution keep failing?
Financial strain and execution risk amplify the labor shortage. Major construction projects routinely run over budget and behind schedule, eroding already thin margins. Fewer than half of projects (only 47.9%) finish within their initial budget, with average cost overruns around 65%. A mere 8.5% of projects manage to come in on-time and on-budget, and a vanishing 0.5% hit the trifecta of on-time, on-budget, and meeting owner expectations.4
Volatile demand shocks haven’t helped: the rapid expansion of data center construction, for example, has led tech firms to pay construction workers a premium (up to 30% higher wages) to staff projects, driving up project costs across the board. All these factors, labor shortfalls, frequent cost overruns, and demand outpacing capacity, squeeze legacy industries to a breaking point.
What changed that finally makes automation make sense?
Luckily, just as the need for modernization reaches urgency, the toolkit for doing so has matured dramatically. Recent breakthroughs in artificial intelligence, robotics, and automation promise solutions that simply weren’t viable a decade ago.
A prime enabler has been the plummeting cost of AI computation, especially for “inference” (i.e. running AI models in the field). According to Stanford’s 2025 AI Index, the cost to prompt a state-of-the-art AI model dropped 280× in two years, from about $20 per million queries in 2022 to just $0.07 by late 2024.5
In practice, this means advanced vision or language models can now run on affordable edge devices or cloud instances, enabling real-time decision support on job sites and in vehicles. Cheaper, more powerful AI lowers the barrier for small players to deploy smart automation.
At the same time, open-source and “foundation” AI models have proliferated, accelerating innovation in unstructured environments like mines and construction sites. State-of-the-art large language and vision models, most with publicly available weights, bring unprecedented semantic understanding and high-level reasoning capabilities to robots operating in messy real-world conditions.6
This brings about improvement in robotic perception and autonomy: whereas earlier industrial robots were confined to tightly controlled settings, new AI brains allow machines to understand complex scenes, navigate changing environments, and even understand human instructions with far greater robustness.
Together, these innovations mark an inflection point, with intelligence, economics, and autonomy aligning to enable deployment beyond controlled environments.
Is this still experimental; or already inevitable?
The adoption of technology in construction sites and infrastructure projects is fast moving from novelty to necessity. Drones and robotic surveyors are also gaining prominence in the field of site mapping, progress surveying, and inspections. The construction drone’s system possesses a high market growth rate of above 20%.
This market has the potential to reach the market value of $7-8 billion in the current decade. And most importantly, construction firms are now seeing the actual evidence of the advantages of drone surveying, which is helping firms save up to 60% of their manual surveying time, up to 30% in labor costs for conducting the survey, and up to a whopping 25% gain in investment returns, as per the latest research carried out in the industry sector.7
In mining, the march of automation is even further along. What used to be isolated pilots of self-driving haul trucks are now large-scale fleet conversions. As of the end of 2024, Caterpillar alone had 690 autonomous haul trucks operating at customer sites worldwide. (For perspective, Caterpillar had only about 100 autonomous trucks in 2017, a nearly 7× increase in seven years.) Now the company expects over 2,000 autonomous trucks to be rolled out by 2030, not just targeting the big iron mines in Australia but even smaller quarries and construction aggregates.8
Real-world results support this optimism, with some mines recording an increase in production output, and in the absence of collision-related incidents, vindicating the creation of this technology in harsh environments.
Each drone flight and robot truck deployed in the job further reinforces the ROI, which is a feedback loop of sorts.
What happens when ROI starts compounding instead of stalling?
For these under-digitized industries, the underlying business case for modernization for enterprises ultimately depends on ROI, and that ROI case gets better and better. The basic idea behind it is that new technologies address the often-acknowledged inefficiencies with traditional projects.
Every construction worker is aware of this pain point. There is rework to be done, delays waiting on information, materials to be supplied, and teams potentially being idle due to poor coordination. All of this leads to a waste of valuable time and money. Digitization and automation directly address this problem.
Every construction worker knows the potential benefits of technology and how it can deliver significant results despite being a relatively simple solution. In fact, a 2024 report on technology trends within this industry demonstrated that adding one more technology or software to a contracting business directly results in a 1.14% increase in business revenue. Although this seems like a marginal statistic, for a $100 million contractor, this equates to a $1.14 million increase in revenue alone.7
Is safety the most overlooked return on investment?
No discussion of automation is complete without considering the people on the ground. These industries have historically been dangerous: construction remains one of the deadliest occupations (the U.S. recorded 1,075 construction fatalities in 2023, the highest of any industry), and mining is rife with hazards from blasting to toxic gases.
Automation offers a chance to drastically reduce these grim statistics. By taking humans out of the most perilous tasks, whether it’s scaling high steel, handling explosives, or driving 200-ton trucks in pit mines, robots and AI can literally save lives.9
What made investors take legacy industries a lot more seriously, all at once?
The convergence of need and technology in construction, mining, and logistics has not gone unnoticed by investors and entrepreneurs. In fact, many see this modernization cycle as a venture-scale opportunity on par with the biggest tech waves of previous decades, albeit with some unique characteristics. Consider the sheer TAM (Total Addressable Market): construction alone is a ~$13 trillion global industry (about 7% of global GDP)10, yet it has been historically under-digitized (with firms investing only ~1-2% of revenue in IT, versus 3-5% in most industries)7.
Mining and logistics add trillions more in economic value and have similarly lagged in tech adoption. Even a modest penetration of software, robotics, and AI into these domains translates into hundreds of billions in potential value. Investors have taken note.
From a market point of view, the prospect for category leaders in this modernization push is huge, with potentially tens of years of retrofit and modernization to come. Projections estimate that by 2040, the U.S. could potentially deploy 8 million + humanoid or general-purpose robots in the workforce, which would have a $357 billion impact on wages (i.e., robots would be doing work that humans currently would be paid for). By 2050, these numbers would escalate to 63 million robots and $3 trillion in wage value.11
Why is this modernization wave different from every failed one before it?
Years of false starts are finally leading the long under-digitized worlds of construction, mining, and logistics into what appears to be a genuine renaissance of modernization. And the catalysts are plain: critical labor and cost issues face a moment when AI and robotics are finally advanced and viable enough to provide solutions that offer ROI benefits. Early adopters enjoy the benefits of increased efficiency, safety, and competition that were previously unavailable with conventional techniques.
Most importantly, the process of this transformation has not been led, at least in terms of hype, from Silicon Valley, but from operators and engineers in various sectors who are realistically applying technology into the way they work. This gives these trends longevity and legitimacy that the past “revolutions” did not.
Challenges remain, of course. Companies must manage the human side of the equation, preparing their workforce for new roles and mitigating displacement where they can.
Throughout the process, they’ll unlock immense economic value and perhaps even rebrand these legacy industries from laggards into leaders of innovation.
