YC’s Latest Signals: Software Dominance, Fintech Momentum, and Physical AI

Looking at YC’s last one-year company dataset, the clearest signal is not simply “AI is hot.” That is already obvious. The more useful signal is where AI-native companies are being built.

Across the dataset, YC still looks heavily software-led. But the centre of attention is shifting from generic AI apps toward workflow automation, financial infrastructure, developer infrastructure, and physical-world AI.

1. Software still dominates, while Consumer is not the centre of gravity

In S26, the largest sectors are still Enterprise Software and Developer Tools. Together, they represent almost half of the batch.

That tells us YC remains strongly B2B-oriented. The dominant startup pattern is software that automates business workflows, helps companies run agents, improves engineering productivity, or replaces manual operating work.

Consumer Tech is still present, but it is not the main centre of gravity. The reason is commercial: consumer AI products are easy to launch, but harder to defend and monetize. B2B products have clearer budgets, stronger workflow ownership, and more measurable ROI.

2. Fintech is the clearest growth story

Fintech is the strongest upward signal in the dataset.

This does not look like a return to consumer banking apps. The stronger pattern is financial infrastructure for an AI-native economy: payments that let agents transact, lending and underwriting that run without a human in the loop, and back-office finance automation.

The signal worth noting is that the customer is increasingly software, not a person — money moving between systems rather than between people.

3. The physical economy is becoming more visible

The other important shift is the rise of AI applied to the physical world.

AI is moving beyond screens and dashboards toward drones, factories, satellites, and energy systems. The open question is fit: hardware is capital-intensive and slow to scale, which sits awkwardly with YC's small-cheque, fast-iteration model. Their presence in the batch says founder attention has moved here — it doesn't yet prove the model works at seed.

In S26, Robotics, Defence Tech, Spacetech, and Energy form a meaningful cluster. Robotics is the largest within this group, while Defence Tech has become much more visible.

4. What this means for where capital goes next

Enterprise software tops these batches for a simple reason: it's cheap to build, quick to iterate, and fast to ship. That low barrier is also its problem — everyone starts with the same horizontal tools (generic copilots, agent frameworks, dev productivity), so competition there is brutal. The more intriguing opening is vertical: industry-specific software that hasn't adopted AI yet. Those markets are still open, and that's where a company can differentiate in a way the crowded horizontal layer doesn't allow.

The physical economy looks exciting, but it's capital-intensive and the large players have already arrived — rounds get big early, and small cheques get crowded out. The more realistic edge can be the software layer underneath: AI-native fintech infrastructure, and the tooling that defence, robotics, and energy companies will need to run. Picks and shovels, not the heavy machinery.

Source: YC Directory in here.