
Perplexity is pulling small businesses into the agentic AI era, and it is doing it through a federal backdoor most AI labs ignore. The company has launched the Main Street AI Accelerator in collaboration with the U.S. Small Business Administration, committing $25 million in credits for its Perplexity Computer agent platform. The program offers $250 each to roughly 100,000 eligible companies that have taken SBA-backed financing, timed to America's 250th anniversary.
The eligibility list is narrow on purpose. Credits go to recipients of SBA 7(a) loans, 504 loans, and microloans, three of the most widely used federal small business financing programs. Applications are not yet open, and Perplexity is collecting interest signups on the program site.
What $250 actually buys
To understand the offer, it helps to know how Perplexity Computer is priced. Computer launched on February 25, 2026 at $200 per month as part of the Perplexity Max subscription tier, and Max subscribers receive 10,000 credits per month, with each task burning credits based on how many sub-agents spawn, which models handle each step, and how many iterations the orchestrator runs. A $250 grant therefore lands somewhere between a free month of Max and a healthy top-up on credits for a heavier user, enough for a small business to actually try the product without committing budget.
That matters because Computer is not a chatbot. It is a multi-agent system that can reason, research, write code, build products, delegate tasks and run workflows without human input at each step. Under the hood it coordinates 19 different AI models, using Opus 4.6 as its central reasoning engine, Gemini for deep research, Grok for fast lightweight tasks, and ChatGPT 5.2 for long-context recall. For a 10-person operation with no IT team, that is the closest thing to hiring a junior analyst.
Don't miss what's next in AI
Join 300,000+ engineers and researchers who get the signal, not the noise.
- Full access to in-depth AI research breakdowns
- Be the first to know what's trending before it hits mainstream
- Daily curated papers, repos, and industry moves

