The top 8 companies from Y Combinator W17 Demo Day 2

0/8 Replay Gallery More Galleries

The top 8 companies from Y Combinator W17 Demo Day 2

3D printing factories, machine learning APIs, and electric planes made our list of the most high-potential startups out of the 51 that presented at Y Combinator’s Winter 2017 Demo Day 2. We asked the top Silicon Valley investors and conferred with the TechCrunch team to pick these 8 companies. But there were plenty more startups with big opportunities from this batch of backend engineering, hard science, and enterprise products.

You can click through/scroll to learn about each of our picks. Also, be sure to read our full coverage of the 52 companies from Day 1, our top 7 startups from Day 1, and all 51 startups from Day 2.

Additional reporting by Sarah Buhr, Ryan Lawler, and Connie Loizos


Voodoo Manufacturing – A robotic 3D printing factory

Voodoo wants to be the AWS for manufacturing. It’s building a robotic factory of 3D printers so clients can send a digital file and get a physical product in return with no molds, labor, startup cost, or minimum order size. That massively democratizes access to manufacturing, the same way AWS did for cloud computing. It’s already working with Nike, Microsoft, Verizon, and Intel, and is on track to make $330,000 this quarter at a 65% margin. Voodoo is looking to disrupt the $50 billion plastic injection molding market, and grow it by making manufacturing as flexible as spinning up servers.

Read more about Voodoo Manufacturing on TechCrunch


Wright Electric – Boeing for electric airplanes

Wright Electric wants to build the world’s first electric airplane. One of the main reasons airlines like Southwest can offer low fares is that they pre-purchase gas, but Wright sees an opportunity to make flights even cheaper by using electric planes instead. The company is targeting the 30 percent of all flights that are 300 miles or less, and partnering with EasyJet to start. As technology improves, it believes its planes will be able to go after the $26 billion short-haul flight market.

Read more about Wright Electric on TechCrunch


Lively – Modern healthcare savings account (HSA)

HSA’s are a triple advantage savings accounts and the “future” of savings says the founder and will hit a $435 billion market. Lively is a modern HSA that makes it easy to access what you have in the bank. It has already worked out permissions with the bank so you can have the money for what you need. Lively is also creating a healthcare marketplace and is a payments and banking platform.


Indigo Fair – Amazon for local retailers

Indigo Fair is a free, AI-powered platform where retailers can find new products for their store. It allows retailers to A/B test merchandise and they are able to return what doesn’t sell for free as well. The team built Square Cash and one of them headed Square Capital. Local retail in the U.S. is a massive market and “shop local” is a slogan that’s not going away. Most small business owners go to trade shows or buy through independent sales reps but so far no one has figured out how to get them shopping online until now. The company has 30 retailers in the past eight weeks since launch and says it has a high rate of return, potentially worth $450,000 in LTV and has on-boarded 6,000 makers.


NanoNets – A machine learning API

Machine learning will change the way business is done, but like databases, most companies don’t build their own from scratch. NanoNets’ API makes it easy for any business to employ machine learning. They just upload their data, wait 10 minutes for it to be analyzed, and add a few lines of code. They can then start seeing results of automatic data mining via ML, such as being able to identify the brand of a shoe in photos. NanoNets is able to recycle learnings from previous jobs to reduce the amount of data it needs to do future tasks. It’s already seeing 1 million API calls a week, and sells a $99/month subscription with up to 10,000 API calls. By working with multiple clients, it can improve its systems much faster than any single client building ML technology by themselves.


Ledger Investing – Helps insurance companies reduce risk

The concept of “insurance securitization” dates back more than 40 years, but it was other securitized products — think mortgage-backed securities — that ultimately captured the imagination of Wall Street. There’s a good reason for that, argues Ledger Investing, a year-old, Mountain View, Ca. startup that points the blame at the various and complex types of insurance that gets bought and sold every day. (Think more and less risky underwriting risk.) What’s changed? According to Ledger, at least, it has finally figured out how to create a business-to-business online marketplace where insurers and investors can confidently sell and buy different types of securities linked to various classes of insurance — and at scale. Certainly, the company’s CEO might instill confidence in those intrigued by this kind of financial product. Samir Shah, who joined the company last September, was formerly the head of Insurance Capital Markets at AIG.


RankScience – Software-automated SEO

“Just plug in RankScience, and search traffic goes up,” says CEO, Ryan Bednar. RankScience automatically optimizes websites for SEO through non-stop A/B testing, going beyond analytics to actually change your site. The catch is you have to route traffic through its CDN which adds a tiny bit of latency. Over the three-month period that RankScience was a participant in YC’s program, early clients saw their average search traffic jump by 68 percent. Without sharing how much the company charges for its results, he added that RankScience is currently seeing $80,000 in monthly recurring revenue. Google makes $50 billion off of search every year so there’s a big opportunity in replacing humans doing manual tests with computerized SEO fine-tuning.

Read more about RankScience in TechCrunch.


lvl5 — Computer vision maps for autonomous vehicles

You might have read recently that one of the biggest obstacles to building self-driving technologies is a shortage of special laser sensors like LiDAR that help cars figure out what’s around them. These sensors — which emit short pulses of laser light so that software in a vehicle can create a 3D image of its surroundings — can also be atrociously expensive, ranging from $8,000 to $80,000. Now, lvl5 thinks it has an even better, cheaper, and more plentiful solution: computer vision software that extracts visual landmarks like stop signs and landlines, then aggregates the data into a kind of 3D map of the world that enables cars to triangulate their locations down to within an inch. Basically, the company says it can achieve the same level of accuracy as LiDAR. But better, it alleges, its system combines its software with cheap cameras, opening up the possibility for the mass production of self-driving cars. Even more interesting here: lvl5 thinks that software will become a commodity, so it isn’t even charging for it right now. It sees the big money instead in its mapping data (and that ain’t free).