YC Application — What to Prepare Now for the Next Batch

This page covers exactly what to prepare in the weeks before that deadline opens, what has changed in YC's stated priorities heading into the 2026 cycle, and the specific preparation sequence that gives you the best shot at a strong application rather than a rushed one.

YC now runs four batches a year rather than two. Understanding the current calendar prevents the most common timing mistake — missing a deadline because you were tracking the old two-batch system.

The Spring "X" batch is a relatively new addition to YC's calendar — prior to 2025, there was no separate spring cohort. If you are planning your preparation timeline, check ycombinator.com/apply directly for the exact current-cycle dates, since YC adjusts deadlines by a few weeks from cycle to cycle.

Weeks 1-2: Get your traction story precise

Before touching the application form itself, get your numbers exact. Current MRR or revenue, exact customer count, retention rate with the specific cohort and window you're measuring, and your month-over-month growth rate for the last 3 months individually. If you do not yet have meaningful traction, spend these two weeks talking to users — aim for at least 20-30 structured conversations if you have not already done this depth of research.

Weeks 3-4: Write and rewrite your core fields

Draft your 50-character description, your "what does your company do" answer, your insight, and your "why now" answer. Do not submit your first draft. Write each field, set it aside for 48 hours, then rewrite it with fresh eyes. The fields that benefit most from multiple revision passes are the insight field and the unfair advantage field — both require you to identify the single most important, most specific thing you know that a generic competitor would not.

Weeks 5-6: Record and re-record your demo video

If your application requires a product demo or founder video, budget more time for this than you expect. Multiple short, focused takes beat one long, rambling take. Show the actual product working with real data wherever possible rather than narrating over static slides.

Weeks 7-8: Submit early, then keep building

Submit at least 48 hours before the deadline — YC's servers experience high traffic in the final hours and early submission avoids that risk entirely. After submitting, do not stop building. If you are not immediately invited to interview, YC's update feature lets you report new progress on your existing application. A meaningful update — a new paying customer, a jump in retention, a feature shipped — can move a borderline application from "maybe" to "yes" weeks after submission.

Based on YC's recently stated Requests for Startups and partner commentary heading into the 2026 cycles, several priority shifts are worth knowing before you write your application:

AI positioning has tightened significantly. Thin interfaces on top of foundation model APIs — informally called "GPT wrappers" — are explicitly deprioritized. YC partners are looking for vertical AI applications and autonomous agents with genuine workflow depth, not general-purpose AI chat interfaces. If your product touches AI, your application needs a clear, specific answer to "what happens when the underlying model gets better at this?"

YC is actively seeking startups beyond pure software. Manufacturing, logistics, defense technology, and healthcare applications using specialized models and robotics are explicitly named priority areas. If you are building in one of these categories, lead with that positioning rather than trying to fit a generic SaaS narrative onto a hardware or operations-heavy business.

Outcome-based and usage-based pricing models are favored over seat-based SaaS. As enterprise buyers move away from per-seat licensing, YC has signaled interest in companies pricing on outcomes, usage, or compute consumption rather than traditional per-user SaaS pricing.

Energy infrastructure for AI compute is a stated priority. Given the scale of data center energy demand driven by AI growth, nuclear fission and fusion, battery chemistry, and grid efficiency startups are explicitly called out as areas of interest for the 2026 cycles.

Crypto positioning has shifted toward infrastructure. Speculative NFT-style projects are not a current priority. Practical cross-border payment rails and stablecoin infrastructure are the specific crypto/web3 angles YC has signaled interest in.

The most common mistake founders make when preparing for an upcoming YC deadline is treating preparation as a content-writing exercise — polishing the application fields — rather than a traction-building exercise. The application fields document what is true about your company. If you spend 8 weeks polishing prose around weak underlying evidence, you will submit a well-written application describing a weak company.

The better sequencing: spend the first 4-6 weeks of any preparation window primarily on the business itself — talking to more users, shipping the feature that addresses the most common objection, signing the customer that proves your distribution channel works — and only the final 1-2 weeks on the application form itself. The strongest applications are not the ones with the most writing effort; they are the ones describing the most real evidence, written clearly enough that the evidence is immediately legible to a partner reading hundreds of applications.

Competition has intensified, not relaxed. Recent batches have seen roughly 20,000-25,000 applications per cycle, with acceptance rates holding in the 1-1.5% range historically. The barrier to building a product has lowered thanks to AI tooling, which has increased application volume without proportionally increasing the bar for genuinely differentiated companies. This means the relative advantage of real traction and a sharp, specific insight is larger now than it was several years ago — generic, AI-assisted applications are easier to produce and therefore less differentiated by default.