Startups Shift from Software to Hard Tech

AI capabilities now enable solo founders to achieve what was previously considered impossible for a single individual. This shift results in rapid development cycles, where ideas can be transformed into working models within a month, pushing the boundaries of what small teams can accomplish.

The proportion of hard tech companies within YC batches has surged from 8% to 20% over the past 12-18 months. This indicates a significant industry-wide shift towards tangible physical products and hardware. Concurrently, the median revenue for YC companies at the end of the batch has increased from $8,000 to $20,000, underscoring both accelerated market adoption and a higher valuation for these new ventures.

Detailed analysis of hard tech growth reveals significant increases across several key sectors. Robotics, for instance, has grown from 1% to 7% of YC batches, while industrial manufacturing, focusing on domestic production, saw an increase from 4% to 10%. Additionally, defense startups expanded from 1.5% to 5%, semiconductor and photonics companies grew from 1% to 4%, and power infrastructure rose from 1% to 3%, highlighting broad diversification within the physical hardware domain.

In the current YC summer batch, one in six founders holds a PhD, a significantly higher proportion than historically observed. This trend reflects the increasing complexity and specialized nature of deep tech fields, such as silicon photonics, which demand extensive research backgrounds. Such highly technical founders are demonstrating disproportionate success, leveraging their expertise to tackle ambitious problems.

This advancement means that developing full-stack hardware no longer necessitates massive engineering teams comparable to those in Big Tech. The economic landscape for hardware startups has fundamentally shifted, lowering the barrier to entry and fostering innovation.

Three overarching macro trends are propelling the growth of atom-based companies. First, the success of ventures like SpaceX has inspired a new generation of founders to build in the space sector. Second, there's an increased focus on sovereign and space-based infrastructure, exemplified by companies like Exosat and Beyond Reach Labs. Lastly, heightened geopolitical awareness and modern conflicts are spurring significant interest and investment in defense innovation, driving demand for advanced hardware solutions.

While the median starting revenue for startups entering YC remains near zero, there has been a significant increase in their revenue by the end of the batch. End-of-batch median monthly recurring revenue (MRR) has jumped from $8,000 to $20,000. This growth is largely driven by startups that perform full end-to-end workflows, delivering complete solutions rather than partial ones.

Customers are increasingly prioritizing software that completes entire jobs, rather than just tracking them. Consequently, companies are willing to pay more for products that offer higher value propositions through comprehensive, job-completing capabilities.

Juicebox, an AI-powered recruiting tool, has evolved from a basic LLM-powered search function to a sophisticated agent-based product. This advanced agent can now not only identify potential candidates but also contact them and schedule interviews, automating much of the early recruitment process. Recruiters benefit significantly from these agents as they offload repetitive tasks, allowing human talent acquisition specialists to concentrate on higher-value activities such as assessing cultural fit.

This unprecedented acceleration in revenue growth represents a stark contrast to previous norms, where reaching such a scale typically required 18 months or more. This rapid maturity is driven by startups' ability to solve real-world problems effectively and their efficient use of coding agents.

This sector was practically non-existent at the time Scale AI was founded in 2016. Y Combinator has recently funded over a dozen such companies, each generating more than $10 million in annual revenue, with some exceeding $100 million.

Companies are heavily investing in benchmarking models to optimize their effectiveness, recognizing that this process is costly but essential. Big labs are reportedly investing approximately a billion dollars into these critical RL developments, solidifying its place in the AI landscape.

This effort aims to bridge the gap between AI and robotics, enabling AI systems to operate effectively in real-world environments. Startups like Practis Robotics and DeepReach are at the forefront, collecting industrial and local entrepreneurial data, establishing physical interaction as a critical category for AI development.

Tools such as River AI or Tinker enable companies to train models out-of-the-box, significantly streamlining the development process. Custom fine-tuning of these models allows smaller, more focused AI systems to outperform even the most advanced frontier models in specific tasks, demonstrating that ownership of proprietary data provides a powerful competitive advantage.

There is a notable resurgence in the market for experienced founders, with many successful entrepreneurs now in their 40s and 50s. This demographic shift highlights the increasing value of deeper industry expertise and seasoned judgment in navigating complex startup challenges. The long-held myth that YC, and the startup ecosystem in general, is exclusively for young, first-time founders is being actively challenged by these successful, older entrepreneurs.

AI is reshaping the necessity of co-founders, allowing entrepreneurs to prioritize product direction over mere technical execution. Knowing what to prompt and what to build has become significantly more valuable than the CTO's ability to simply code a solution. While co-founders still offer substantial benefits for startup success, the entry barrier for solo founders has been significantly lowered, as evidenced by historically successful companies like Instacart and Coinbase that started with solo founders.

Experienced founders are increasingly outperforming in the current AI era, leveraging their historical industry context and refined business acumen. These seasoned entrepreneurs possess a better understanding of market risks and opportunities, often referred to as knowing "where the dragons are." The traditional "gatekept" path to startup success is diminishing as market knowledge and strategic insight become primary drivers. Consequently, their opinions and industry experience now carry more weight than needing a conventional co-founder setup.

AI agents are now capable of handling complex bugs by referencing historical chat contexts, streamlining the development process. This era allows for rapid iteration and significant technical progress within remarkably short timeframes, making it an opportune moment to start a company.