Y Combinator opens applications for the fall cohort, with one of the startup themes focusing on government waste reduction.
The theme of “Using LLMS in place of government consulting” reflects the mission of the Department of Government Efficiency, a Trump administration's initiative to reduce spending at the federal level.
Although the well-known accelerators did not explicitly mention Doge, YC said on its website they wanted to fund startups building large language model-driven software to take over the work that big giants like Deloitte and Accenture do for the public sector.
“There's political pressure to reduce wasteful consulting and spending,” writes Gustaf Alstromer of YC. “The US government spends over $100 billion a year on consulting. As you can imagine, this is not the most efficient or innovative part of our economy.”
Alstromer said, “Currently LLMs are so good that they can already do the work of many consulting companies,” but “startups” can do better with building software for government use.
He added that he is either supporting businesses that support YC already certified to sell to the government, or that he is using AI to ensure the legality of new laws and policies.
The Y Combinator did not respond to requests for comment from Business Insider.
The accelerator pitch will put pressure on government spending and modifying the slow moving bureaucracy.
Earlier this year, the Trump administration called on the federal government's highest-paid consulting firm to justify spending on contracts using a language that “a 15-year-old can understand.”
Doge was led by Elon Musk until April.
This is one of six themes outlined in YC's Fall 2025 “Startup Requests.” This is a wish list showing where Silicon Valley can see the waves of the next breakout company.
Application for an accelerator that will be closed on Monday evening during Pacific time.
Here's what YC bets:
Related Stories
Business Insider tells the innovative stories you want to know
Business Insider tells the innovative stories you want to know
AI training for blue-collar workers
YC wants to tackle talented bottlenecks in trading rather than technology.
AI Racing focuses on hiring engineers and researchers, but there is a growing shortage of skilled merchants like electricians and welding machines. This requires building the infrastructure behind AI, such as data centers and semiconductor fabs.
“I think we can use AI to create personalized training programs so that people can be hired in months rather than years,” writes Harj Taggar of YC. “This is where multimodal AI can create opportunities,” he said.
His examples include voice assistants who can coach someone in an AR/VR environment that can simulate real-time tasks and tasks, while the vision model provides live feedback like a human trainer.
Video generation as a core technology
YC said the video has evolved into “new basic building blocks in software.” Unlock new types of apps, developer platforms, and real-time experiences.
“The video generation model is really getting better,” writes YC's David Reeve.
“In no time, you'll be able to produce near perfect footage of anything on the spot, for the marginal cost of getting closer to zero,” he said. “When this happens, many new ideas are possible.”
The video generated by AI is “unquestionably changing media and entertainment,” and reconstructing the way games and simulations are shopped and built, he added.
First 10 people, $100 billion company
The first 10 people, a $100 billion company? YC believes that is not only possible, but also imminent.
Accelerator said it was possible to build a multi-billion dollar company, even a small, ministry team, and even a solo founder.
The new AI tools “make it easier for ambitious founders to expand with far fewer people,” writes YC's Aaron Epstein.
“With a large and efficient team, they don't suffer from politics, excessive meetings and a lack of focus to stop huge businesses,” Epstein writes. “They can focus on winning with better speed and execution.”
The best startups of the future will “optimize one metric: revenue per employee.”
YC's focus comes after at least 12 AI startups cross the $1 billion valuation threshold with small teams, Business Insider reported in May.
“We'll see 10 companies being valued at $1 billion right away.” Openai CEO Sam Altman I said this in February 2024. “In a chat with my tech CEO friend of my small group chat, there's this betting pool. In the first year, there's one billion dollar company. [it] It will happen. ”
Infrastructure for multi-agent systems
A future where AI agents not only work alone, but also work together.
AI agents have evolved from “single-threaded loops” to “multi-agent systems,” writes YC's Pete Koomen. These systems can break down large tasks, run them in parallel, and tackle complex jobs faster than a single model
But they are difficult to build and can introduce new problems that need to be solved with “higher levels of abstraction,” Coomen said.
YC is looking for builders who “feel this pain in this production” and wants to create tools that make it easier to build and maintain multi-agent systems.
AI-Native Enterprise Software
Finally, YC said that with AI at its core, it is seeing a great opportunity to reinvent enterprise tools from scratch.
YC said the next generation of tech giants will be built around AI-Native software, just like how Salesforce and ServiceNow rode Cloud Wave 25 years ago.
“Today's incumbents struggle to rebuild their products around this new technology, giving today's startups the time they need to beat,” said Andrew Miklas of YC.
These startups have AI “becomes deeply thoughtfully embedded throughout, helping employees work faster and more accurately,” he said. “Think about your cursor for sales, HR, accounting.”
