Established, venture capital firm known for pre-seed and seed-focused, Pear VC runs an accelerator with nearly a decade of experience and about 10 startups in each batch.

From early last year to 2020, the tiny yet powerful program spawned companies such as Viz.ai: its FDA-approved AI can diagnose strokes and was valued at $1.2 billion in 2022; relationship management company Affinity that raised an $80 million Series C at a $620 million valuation, according to PitchBook data; and Valar Labs, which uses AI to help doctors decide on cancer treatment. It closed a $22 million Series A in May.

This year, Pear has decided that the time has come to scale its accelerator in size and give the companies more than services, recruiting, and space inside the firm's new 30,000-square-foot San Francisco office. From here on out, the accelerator program, dubbed PearX, will run twice annually. Each round will include around 20 companies. The accelerator remains a long way from Y Combinator's, which accepts hundreds of startups every year.

And it's not just about size: The startups in each batch are usually not known until demo day-while YC packs more than 100 VC general partners into an in-person room in the fall, with some attending from flagship firms like Sequoia, Benchmark, and Index Ventures. While YC boasts of offering the standard deal to every company, funding for PearX startups by the firm can vary anywhere between $250,000 and $2 million based upon needs and stage.

This month's demo day was dominated by 20 companies-majority focusing on AI. Among them, here are five that stood out to us and the crowd in attendance with fresh approaches to complex business problems.

Neutrino AI 

What it does: identifies best infrastructure for multi-model AI applications

Why it matters: AI companies must be sure to get the right tool for the job. The search for the perfect LLM or small language models is a painful process because these models change and improve daily.

Nuetrino would make life easier for AI companies to find what could be the right mix of models and other systems to apply in their applications so the developers work faster, thus making a saving on the running of their products.

Quno AI

It does: Automates market research

Why it stood out: Brands spend hundreds of millions of dollars a year on market research. The process of questioning customers is not only boring but also quite a time killer. Quno AI agents can call the customer and gather qualitative and quantitative data. Results can then be analyzed in real-time. A bonus is that AI can quickly analyze results from these conversations.

ResiQuant

What it does: Creates catastrophe models for home insurance carriers

Why it matters: As natural disasters continue to worsen, property insurers are left searching for a way to know which houses would be most likely to incur major losses during a catastrophe. And that's mainly because access to data regarding homes' structures is not easily and cheaply available.

Based in Berkley, the startup is founded by two Ph.D.s in structural engineering. It has created models to estimate features in buildings and how they'll hold up during earthquakes, hurricanes, and fires. The company says it can help insurance carriers assess risk more accurately, which could lower homeowner insurance premiums to those who are deemed to be at lower risk.
Self Evaluate
What does it do? : Monitors real-world production and alerts operators of mistakes

Why it matters: Investigators say four critical bolts were missing, causing the doors of a Boeing 737 Max to blow out during flight in January. This is just one example of what can go wrong within quality assurance systems. Manufacturers of all types have similar needs-to detect defective products before they leave the factory.

It uses cameras and AI to address such problems through ensuring correct execution of works and identification of errors during production in real-time.

TeachShare

What does it do? It creates lesson plans tailored to the needs of each educator.

Why it matters: There's been adaptive software that makes lessons more challenging or easier depending on what a student knows for years. But educational companies still package much more in the form of one-size-fits-all model when it comes to curricula design. The end result means the teacher spends far too much time preparing lesson plans to fit their specific class. TeachShare wants to help teachers modify the content of daily teaching and assure it meets the standards.