Welcome back to Inc.’s 1 Smart Business Story, where we’re counting down our favorite stories of the year.  

Today: The leading startup accelerator, Y Combinator, regularly puts out a call for the kinds of companies it wants to see apply for its program, and the one it released last summer was a doozy for what it had to say about the state of AI and the future of startup creation and company building. Read on to learn:   

Not too long after Inc. published this analysis, we ran another piece that explored how the AI bubble was eating Y Combinator. So … which side are you on? Email me at [email protected] to tell me what you think.  

One idea the accelerator has is supporting founders who want to build the first 10-person $100 billion company.

BY BEN SHERRY, STAFF REPORTER 

Y Combinator, the famed Bay Area startup accelerator, has released its quarterly request for startups for this fall. These requests give a sense of the types of companies that YC partners are looking to invest in, and this quarter, it’s all about AI. 

All six of the firm’s requests for startups are for companies built with AI at the core. They’re looking for organizations focused on everything from retraining employees so that their skills are more relevant to the AI economy to growing a company to $100 billion in revenue with just 10 people. 

Here’s what Y Combinator is looking to invest in this fall. 

YC managing partner Harj Taggar wrote that he’s looking for startups creating “a new kind of vocational school for the AI economy.” As Taggar explains, to keep the AI revolution going, massively ambitious infrastructure projects are needed to create data centers and chip foundries.

“While we’re focused on the race for AI talent,” Taggar wrote, “we also have a shortage of skilled tradespeople—the electricians, the HVAC technicians, the welders—who are essential to building this physical infrastructure.” This need for more skilled tradespeople was echoed in the Trump administration’s recently released AI action plan, which called for the federal government to fund retraining programs aimed at creating these kinds of jobs. 

Taggar wants to fund startups that use personalized AI to train people in trade-based skills in months rather than years, and then charge employers to hire those newly trained workers. 

Taggar imagines that multimodal AI, which can process and generate various types of media, including images, video, and audio, could be a pivotal factor in making such a company work. “For example,” he wrote, “maybe a voice AI could coach someone through these tasks. Or perhaps some combination of AR/VR could let people practice the work in simulation with an AI tutor using vision models to watch them and give feedback.” 

David Lieb, a general partner at YC, is looking for startups that are taking advantage of the rapid technological progress of AI-generated video. Lieb wrote that AI-generated video, especially videos created by Google’s Veo 3 model, can be incredibly lifelike while costing only a few dollars, and it’s just going to get cheaper going forward. 

Because of the power and low cost of AI video, Lieb sees several potential ways for startups to create new kinds of businesses using the tech. Lieb offered several examples of what this could look like in practice, from creating “a brand-new season of your favorite canceled TV series” to generating a digital double that can model clothes while you shop online, and even giving people the ability to hold video calls with AI versions of deceased relatives.

Aaron Epstein, a general partner at YC, wrote that he believes that AI tools have enabled small, high-agency teams to “build multibillion-dollar companies with as little as just $500,000 in funding from YC.” 

Just as the rise of cloud computing made it easier for people to build businesses by eliminating the need to spend big on physical server infrastructure, Epstein says, AI is making it “easier for ambitious founders to scale with far fewer people.” 

According to Epstein, the most high-agency companies of the current day will take advantage of this ability to efficiently scale by making “revenue by employee” their most important business metric. 

“With smaller, efficient teams at scale, they won’t get bogged down with the politics, excessive meetings, and lack of focus that grind huge companies to a halt,” Epstein wrote. “They can just focus on winning with better speed and execution.” 

Pete Koomen, a general partner at YC, is looking to fund startups developing tools that make it easier to build, run, and maintain a fleet of AI agents that can work in tandem. AI agents are essentially AI systems that can take actions on a user’s behalf, like navigating websites, sending emails, scheduling meetings, and completing digital workflows. 

Systems that can run multiple agents at once can be incredibly useful, wrote Koomen, but there’s a problem: They’re incredibly difficult to build and maintain, can be incredibly expensive to run, and often aren’t reliable enough to use on their own. 

Koomen is looking to fund startups that are actively building tools to make it easier to manage a digital army of agents. “If you want to make operating fleets of agents as routine and reliable as deploying a web service or running a Spark job,” he wrote, “we’d love to hear from you.” 

Andrew Miklas, a general partner at YC, is on a search for startups building AI-native enterprise software services. He wrote that a new era for B2B SaaS is approaching, and just like the rise of Salesforce and ServiceNow, which took advantage of the cloud computing revolution to win over much more established competitors, Miklas sees a once-in-a-generation opportunity for founders to build new enterprise systems with AI “embedded deeply and thoughtfully throughout.” 

This next generation of B2B SaaS would enable employees to do their work faster and more accurately, Miklas wrote, comparing such a system to popular vibe coding app Cursor. Just as Cursor allows engineers to write and edit code faster, an AI-powered enterprise system could help people in sales, HR, and accounting. 

“Just as before, today’s incumbents will struggle to rebuild their product around this new technology, giving today’s startups the time they need to win,” wrote Miklas. “History doesn’t repeat itself, but it sure does rhyme.” 

Gustaf Alstromer, a general partner at Y Combinator, is looking for startups that are using large language models to create virtual consultants, specifically consultants for the United States federal government. 

According to Alstromer, the U.S. government spends over $100 billion per year contracting consulting firms like Deloitte, McKinsey, and Accenture to assist with various projects and initiatives, but that could change very soon with the Trump administration’s ambitions to make major cuts to the federal government’s spending. 

“Every part of the government now runs on software, but usually custom software built by a consulting company,” Alstromer wrote, “and anyone who has used this software knows we can do a lot better,” especially with the rise of LLMs. Alstromer wants to fund startups that are building AI-powered software that can do an improved job delivering this crucial government consulting work while costing a fraction of the price. 

“Here’s What Y Combinator Is Looking For in AI Startups Right Now” first appeared on Inc.com.