The new Puzzle dashboard is live with charts on every metric card. Read more about it here.
YC's latest batch shows the future: 87% of the startups on the list are AI companies. This completely changes how founders build, raise, and scale. Integrating AI is no longer optional—it’s foundational.
AI is changing every layer of the startup tech stack. Companies are rebuilding developer tools, solving complex industry problems, and automating core operational systems like ERP and customer support with AI.
This isn’t a subtle trend:
These are seismic shifts, not incremental changes. “AI is a catalyst unlike anything we’ve seen in a long, long time; what the internet did for startups in the 90s, and smartphones did in the 2000s, AI is doing once again,” YC said. This is the reality of building a startup in 2025.
The AI focus in YC’s latest batch isn’t generic. It’s specific and targeted. These startups are building AI to solve concrete problems and drive real results, not just chase trends.Across the latest cohort, three patterns stand out:
Startups are zeroing in on industry-specific problems. Sandra AI automates customer interactions for car dealerships, Asha Health optimizes clinic operations, and Tiny brings AI-powered ERP systems to factories. Each company targets distinct operational pain points.
YC's latest Request For Startups includes AI-first vertical solutions. Tax accountants, compliance tools, and customer service platforms now use AI to fundamentally reshape how work gets done. The top companies build specialized AI models on industry data, automate complex workflows, and boost accuracy through AI-guided decisions.
As YC’s Jared Friedman puts it, “While many founders are already working on these ideas in the most obvious categories, we think it's still relatively undiscovered compared to the size of the opportunity, and there are many large categories still untouched.”
Rather than replacing humans, AI is giving them superpowers. Afternoon operates as an AI accountant for small businesses, while Relvy streamlines debugging processes for developers. The focus here is on making people more productive, not obsolete.
This is especially clear in software development. AI writes code, but engineers aren’t going anywhere. Their job is evolving: managing AI-driven development, debugging edge cases, and optimizing performance. As Pete Koomen puts it, “We’ll need more human software engineers in the future because software is going to run almost everything.” The best startups won’t replace humans, they’ll give them more leverage.
AI is pushing professionals toward higher-value work. The startups that win in this space will be the ones that find the right balance between AI automation and human expertise.
Scaling AI is a technical challenge. Companies like Ryvn and Regatta Storage are solving these problems, offering tools that allow other startups to deploy and scale AI reliably. This infrastructure development is crucial as AI becomes central to modern startup operations.
Startups trying to fundraise in 2025 have to integrate AI from the start. Investors back startups that use AI to automate, streamline, and scale. Efficiency is the expectation, not a bonus.
But as AI adoption accelerates, a new bottleneck is emerging: inference-time compute costs. YC’s latest RFS highlights the need for startups that can optimize inference-layer tooling, reduce GPU workloads, and rebuild the AI stack for cost-effective scaling. As Diana Hu from YC puts it, "Until recently, compute spend went into pre-training foundation models. But now with DeepSeek R1 and OpenAI o1 and o3, there is a new scaling trend that suggests we’ll need far more compute at inference time when AI apps actually use these models."
The cost of running AI—rather than just training it—is becoming a challenge. The next wave of AI infrastructure companies won’t just focus on enabling deployment but on making AI economically viable at scale.
For founders, the message is clear: if you’re not building with AI from day one, you’re already behind. The market is evolving, and investors are paying attention.
Founders hoping to join YC or raise funds elsewhere need to align with these expectations.
YC’s latest Request For Startups signals where AI is heading next.
They highlight high-impact AI opportunities across infrastructure, automation, and industry-specific solutions. 2024 made AI mandatory. 2025 will determine which AI innovations win. We're seeing new categories of startups emerge, not just AI features bolted onto existing products.
Startups that deeply embed AI into their foundations—whether through infrastructure, automation, or domain-specific intelligence—will define the next wave of innovation.
Founders need to ask themselves: once intelligence is effectively free, what's your competitive edge? How will AGI reshape your product and team strategy?
Startups that win won’t treat AI as an add-on. They’ll rebuild their foundations to take full advantage of it.
For our team at Puzzle, these questions forced us to rethink everything. We shifted from focusing on owning intelligence to building systems that thrive in a future where intelligence is abundant and accessible. Instead of bolting AI onto existing systems, we rebuilt the foundation from scratch. We created an AI-native general ledger that’s faster, smarter, and designed for today’s digital-first businesses.
Seema Amble, a partner at a16z, explains: "By being trained on contextual data — including internal and external signals — the next generation of AI-powered software could become a system of record that the user can live in." Puzzle's AI is trained specifically on startup data to offer insights tailored to the unique challenges founders face.
Our platform simplifies complex financial workflows by automating repetitive tasks like data categorization and reconciliation. This means founders spend less time navigating their financial systems and more time focusing on strategy. While AI handles the heavy lifting, people ensure accuracy and apply critical judgment where it matters most.
Founders need to rethink their foundation as intelligence becomes more accessible than ever. The challenge isn’t whether to adopt AI, but how to do it meaningfully.
Those who figure this out will lead the next wave of successful companies.
Founders need to start rethinking their foundation to fit a world where intelligence is more accessible than ever before. The challenge isn’t whether to adopt AI, but how to do it meaningfully. Those who figure this out will lead the next wave of successful companies.
Stay at the forefront of innovation and streamline your finances with Puzzle - use the code YCALUM at checkout for 50% off our Starter Plan for 12 months!
