The New Industrial Revolution: Why Atoms Are the New Bits

The move toward hard tech, such as physical infrastructure, robotics, and defense, is a fundamental change in the venture landscape. As AI agents lower the cost of software development and scientific research, the main challenge for founders has shifted from "can I build this?" to "do I know what to build?" This transition favors experienced operators who have the industry knowledge to spot high-value problems and the management skills to direct AI agents. For builders, this reveals a clear advantage: the most lucrative opportunities are now in the atoms layer, such as manufacturing, energy, and defense, where legacy companies are too slow to compete and software-driven automation provides a scalable moat.

The Death of the Software-Only Moat

For years, the venture capital playbook favored B2B SaaS because it was capital-efficient and scalable. But as the YC team notes, the market is responding to a new reality: software is becoming a commodity, and value is moving to full-stack, end-to-end solutions.

The traditional SaaS model, which acts as a system of record to track data, is increasingly vulnerable. If a product does not actually perform the work, it risks being replaced by an AI harness that does.

"If you are a system of record you either will be preyed upon, you will release an MCP and then maybe like the data, you lose your moat around the data. The data goes elsewhere becomes very trivial to switch or you kind of have to be a harness."

-- Jared

This forces a choice for founders: evolve into an active agentic workflow or face obsolescence. The companies growing fastest today are those that do not just track the job but complete it. This shift from tracking to doing is why median revenue during the YC batch has jumped from $8k to $20k in monthly recurring revenue.

Why Hard Tech is Scaling at Software Speeds

The most striking insight is the resurgence of hard tech, including robotics, defense, and power infrastructure. Historically, investors avoided these sectors due to supply chain complexity and high capital requirements. However, the rise of AI-driven research and code generation has changed the economics of building in the physical world.

The need for a thousand-person engineering team to build complex hardware is vanishing. With AI agents handling the heavy lifting of coding and scientific simulation, small, highly technical teams are achieving breakthroughs that were previously impossible.

"The super smart models that we have now are actually accelerating scientific research and making it possible for startups to have bigger research breakthroughs earlier and therefore these deep tech companies will actually work better."

-- Jared

When these startups enter the market, they find a massive vacuum. Legacy defense primes and industrial manufacturers are sleepy businesses, often unable to match the speed of modern, AI-integrated startups. This creates a feedback loop: new defense tech startups need metal, and they prefer to buy from agile, tech-forward suppliers like Knox Metals rather than legacy vendors. This is how a manufacturing company begins to grow at software growth rates.

The Resurgence of the Experienced Founder

The 20-year-old coding prodigy archetype is being challenged by a new reality: the AI-pilled veteran. Because the barrier to writing code has dropped, the advantage has shifted to those who understand the specific, high-stakes problems that actually require solving.

Managing AI agents is similar to managing human teams. Founders with decades of experience in operations or management are finding they can direct these agents more effectively than those who lack that context. This is why we are seeing a spike in solo founders and older, experienced founders; they are not just building software, they are applying deep domain expertise to automate complex, real-world workflows.

"There is just so many classic gate kept things that happen. It is like, oh, you know, you have to have a co-founder or you need like a certain set of cool investors to be into you. And like now it is just less and less true. It is actually like you need to know what to build."

-- Garry

Key Action Items

• Audit your product’s doing ratio: Move your roadmap away from passive systems of record toward active agentic harnesses. If your software tracks work but does not perform it, you are vulnerable to displacement. (Immediate)
• Leverage AI for scientific R&D: If you are in hard tech, stop treating software as a secondary concern. Use AI-driven research and code generation to accelerate your physical prototyping. This is your primary speed advantage over legacy incumbents. (Immediate)
• Prioritize domain expertise over raw coding speed: If you are hiring or looking for a co-founder, prioritize those who understand the atoms of your industry. The what to build is now significantly more valuable than the how to code it. (Next 3 to 6 months)
• Target the sleepy supply chain: Look for industries where incumbents are slow and rely on legacy processes. There is a massive opportunity to build the Stripe of X for these sectors by providing better, faster service to the new wave of tech-forward startups. (Next 6 to 12 months)
• Invest in proprietary data loops: As models become commoditized, your moat will be your proprietary data. Whether it is robotics footage or industrial manufacturing data, build systems to capture and fine-tune models specifically for your vertical. (12 to 18 months)