Most technological breakthroughs begin small. A new material is tested on a laboratory bench. A robot takes its first uncertain steps. A battery completes its first successful cycle. These moments capture the imagination, but the world rarely changes at prototype scale. It changes when an invention can be made reliably, affordably, and by the thousands or even millions.
Part 1 explored why deep technology is attracting renewed attention from venture investors and how artificial intelligence is accelerating innovation beyond software. But invention is only the starting point. A new battery chemistry, robotic system, medical device, or energy technology creates lasting value only when it can make the difficult journey from the laboratory to the factory floor and ultimately to the global market.
Why Manufacturing Matters Again
For much of the past two decades, the venture capital playbook pointed away from the factory floor. The most attractive companies seemed to live on servers rather than assembly lines. Build the software, acquire users, and scale quickly. If manufacturing entered the picture at all, it was often outsourced or left for later.
That approach made sense during an era shaped by the rapid growth of the internet. But as I look at the technologies attracting investment today, the picture is changing. Many of the most consequential opportunities now depend on building something physical and building it well.
Robotics, batteries, semiconductors, medical devices, industrial automation, and advanced energy systems cannot be brought to market through software alone. Each depends on the ability to manufacture a sophisticated product with consistency and precision. In these businesses, production is not the final chapter of the story. It is part of writing the story itself.
A prototype may perform beautifully in the controlled setting of a laboratory and still struggle once it reaches a production line. A material may be too expensive. A component may be difficult to source. A small variation in temperature or tolerance may lower yields or affect reliability. These details can seem far removed from the original breakthrough, yet they often determine whether it becomes a successful company.
This is why the connection between engineering and manufacturing matters so much. The factory does more than reproduce the product. It reveals where the design must improve, where costs can fall, and whether performance can be maintained as volume increases.
One of the clearest examples can be found on the former Weirton Steel site in West Virginia. Form Energy built a factory spanning 550,000 square feet there to produce its iron air battery, which is designed to store electricity for as long as 100 hours. The underlying chemistry is compelling, but the technology becomes truly valuable only if it can be produced at utility scale and at a cost the market can support.
In that sense, the factory is not separate from Form Energy’s innovation. It is the bridge between a promising scientific idea and a product that can serve the electrical grid. Without that bridge, even a remarkable battery remains a laboratory achievement.
The pandemic made this lesson difficult to ignore. Products that once seemed easy to source suddenly became scarce, and supply chains that appeared efficient proved more fragile than expected. Geopolitical tensions have added another layer of urgency, particularly in energy, healthcare, semiconductors, infrastructure, and national security. The ability to invent a critical technology matters. The ability to produce it has become strategically important in its own right.
Artificial intelligence will accelerate this shift, but I do not see it replacing the factory. I see it making the factory smarter. AI can help engineers improve designs, detect production problems earlier, strengthen quality control, and reduce waste. It can shorten the distance between what is imagined in the laboratory and what can be produced in the real world. The next wave of innovation will not be defined by invention alone. It will be defined by the companies that can turn invention into something reliable, affordable, and real.
The Advantage is In the Process
A sheet of glass enters a First Solar factory. Approximately four hours later, it emerges as a finished solar module. What happens between those two moments helps explain why manufacturing can become one of the deepest competitive advantages a technology company possesses.
First Solar does not produce a conventional silicon panel. The company developed its own thin film semiconductor technology along with the continuous manufacturing process used to produce it. Over decades, the product and the process have evolved together. Each improvement in materials, equipment, testing, and production has added to a body of knowledge that competitors cannot easily purchase or recreate.
This demonstrates important differences between software and deep technology. Software can often be copied and distributed quickly. A sophisticated physical product must be produced again and again with the same precision, performance, and reliability. The real advantage may lie not only in what a company makes, but in everything it has learned about how to make it.
That knowledge accumulates quietly. Engineers learn how materials behave under different conditions. Production teams discover how to calibrate equipment, prevent defects, and improve yields. Supplier relationships deepen. Testing data grows. Regulatory approvals and customer qualifications take time to earn. No single element may appear decisive on its own, but together they can form a barrier that is extremely difficult to cross.
First Solar has used this accumulated knowledge to distinguish itself in a global market where many solar products have become increasingly similar. Its advantage is not contained in a single patent or breakthrough. It is embedded across the entire production system and strengthened each time the company manufactures another module.
I think this is where investors need to look beyond the excitement of the prototype. The prototype is a powerful moment. It shows that the science can work and turns an idea into something tangible. The harder test begins when that idea enters production.
Once the product leaves the laboratory, every assumption is tested. The company must reproduce the same performance across hundreds or thousands of units, lower costs as production grows, and maintain quality throughout the process. It must move from a limited run to commercial scale without losing what made the technology compelling in the first place.
That transition is where promising technologies either become durable businesses or remain impressive experiments. A prototype proves possibility. A repeatable production process proves that possibility can become a business.
What This Means for Investors
A technology portfolio can appear diversified while remaining concentrated in the same part of the economy. An investor may own software companies serving different industries, along with cloud computing and artificial intelligence businesses, yet still have relatively little exposure to the physical systems that allow the digital world to function.
Deep technology opens a door to a different part of the market. Advanced manufacturing, robotics, energy, biotechnology, aerospace, materials science, and scientific instrumentation address problems that are essential, global, and difficult to solve. These are not always easy businesses to build, but that difficulty can become part of their value. Once a company earns customer trust and establishes its product inside a critical operation, it can be difficult to displace.
The challenge is that these companies cannot be judged using the familiar software playbook. Their development cycles are often longer, their capital needs are greater, and their progress may not immediately appear in revenue. It may appear in a successful technical test, a higher production yield, a completed certification, or a manufacturing line that is finally ready to scale.
To me, this requires a different understanding of momentum. A software company may demonstrate progress through user growth or subscriptions. A deep technology company may spend years reaching milestones that are less visible but just as important. Customers may need to test the product, certify it, and integrate it into operations where reliability is essential. Commercial adoption can take time because the cost of failure is real.
Rivian makes these demands easier to see. Before the company could deliver electric vehicles at scale, it had to do far more than design an appealing vehicle. It had to validate the product, build a supply chain, equip a factory, and develop capabilities across stamping, body assembly, paint, propulsion, and final assembly. Every step required capital, coordination, and disciplined execution.
That journey shows why a strong physical product is only the beginning. A vehicle may earn attention for its design and performance, but the business ultimately depends on whether it can be manufactured consistently and whether its economics improve as production grows. The same principle applies across deep technology. The breakthrough attracts interest. Execution determines whether that interest becomes lasting value.
For investors, the central question is not simply whether the technology works. It is whether it works outside the laboratory, whether it can be produced repeatedly, and whether it solves a problem important enough for customers to change the way they operate. The management team must understand both the science and the path to commercialization.
Investing in deep technology requires patience, but it should not require blind faith. The milestones are different, not unknowable. The strongest opportunities are often the companies that can turn technical progress into manufacturing progress, customer adoption, and eventually durable economics.
Where Software Meets the Physical World
A thermostat notices that no one is home and quietly lowers the temperature. It is a small action, almost invisible, but it captures a much larger shift in technology.
The Google Nest Learning Thermostat is a physical device connected to a building’s heating and cooling system, yet much of its value comes from sensors and software that learn household patterns, adjust schedules, and manage energy use. The hardware alone would be another thermostat. The software alone could not change the temperature of a building. The value comes from bringing the two together.
I think this combination will become increasingly important. Deep technology is often treated as a separate investment category, but that distinction may matter less over time. Many of the most interesting companies now operate where software, engineering, manufacturing, biology, and materials science meet. The label is not what matters. What matters is whether a company can combine these capabilities to solve a meaningful problem.
For much of the past two decades, technology focused on digitizing information, communication, and commerce. That work transformed the economy and will continue. But the next phase may be defined by what happens when digital intelligence moves beyond the screen and into the physical world.
That means changing how goods are produced, energy is generated, water is managed, healthcare is delivered, and infrastructure is built. These are enormous markets, but they are also essential systems where improvement has often been slow and the consequences of failure are real.
Software alone will not solve energy security, aging infrastructure, fragile supply chains, or the need for more sustainable manufacturing. These challenges require new materials, equipment, physical systems, and production capabilities. Artificial intelligence will be essential, but much of its value will come from making those systems more capable, efficient, and reliable.
For investors, the choice is not between software and deep technology. The real opportunity is to understand where they strengthen one another. A robotics company needs sophisticated software and dependable hardware. An advanced energy company may depend on materials science, manufacturing, and data analytics. A biotechnology company may combine biology, automation, and machine learning. The strongest businesses may be the ones that bring these disciplines together so effectively that they become difficult to separate.
There will be failures. Technical risk is real. Commercialization can take longer than expected, and building physical products often requires more capital than building software. But those challenges do not make the opportunity temporary. If anything, they help explain why the companies that succeed may create such durable positions.
The renewed interest in deep technology feels less like another investment cycle and more like a broadening of what technology can accomplish. The next generation of leaders will not be defined only by the software they develop. They will also be defined by the products they manufacture, the infrastructure they build, and the industries they improve.
Understanding artificial intelligence will be essential. Understanding what happens when that intelligence meets the physical world may prove just as important.
By Jarett Emert
