Parabricks finds a niche to target its computing power

11/5/2018

Parabricks LLC, a 2015 spinoff from the University of Michigan that signed an exclusive licensing agreement with the school last year, is a case study in why entrepreneurs need to be flexible when trying to figure out their business model and who their customers are, or will be. It took the company’s founders more than a year to realize they needed to be a genomics company. Once they did, equity funding and paying customers soon followed.

As did large federal funding. In October, Parabricks was awarded a National Science Foundation Small Business Innovation Research Phase II grant of $748,000, which came with a matching grant of $125,000 from the Michigan Emerging Technologies Fund. That followed an NSF SBIR Phase I grant of $225,000 in 2017, which had a matching state grant of $25,000.

The name of the company is a reflection of the founders’ original intention. While Ph.D. students in computer science and engineering, Mehrzad Samadi and Ankit Sethia developed software that allowed computer chips called graphics processing units to be used for a variety of applications, able to replace the much more common chips called central processing units.

GPUs are much more robust and faster than the CPUs used in most computer functions. Traditionally, the chips have been used to manipulate the huge amount of data and pixels used in video games and smartphones.

Samadi and Sethia figured out how to use GPUs for processes that had nothing to do with graphics. They could, they thought, spin off a company that would create software tools that would in turn allow other developers to build software applications of their own using GPUs instead of CPUs. They thought of their software tools as parallel-processing bricks that could be used to build a powerful application — hence the company name. They got their professor, Scott Mahlke, to serve as the company’s technical adviser.

Samadi is now the CEO, Sethia is the chief technology officer and Mahlke is the chief science officer.

But doing market research, they realized people wanted to buy software and not the tools to build their own.

So, time for their first pivot. Instead of selling tools for others to build with, they would target the auto industry. When autonomous vehicles are on the street, making decisions now made by humans, they will require enormous amounts of computing power. What better way to manage that volume of data than GPUs?

But, again, what made sense to Samadi and Sethia didn’t make market sense. Auto companies had invested billions and were well down their own paths toward autonomous vehicles. They weren’t going to pay attention, or write checks, to a couple of UM grads just because they’d done a licensing deal with the school’s tech-transfer office.

What other market needed copious amounts of fast computing power? Aha! Genomics. Genomics is widely expected to lead a personalized medicine revolution. But for it to be a fast revolution, and one that is affordable to individuals in need of a specific cure for their specific cancer, it will require ultrafast data crunching — the kind of thing GPUs are great at.

Patti Glaza, the vice president of Invest Detroit, a nonprofit that invests in early stage companies in Michigan, is an adviser on the advanced transportation hub of the Michigan Translational Research and Commercialization program, which was founded at UM in 2012 and later became a statewide program with funding from the Michigan Economic Development Corp. She first encountered Samadi and Sethia there. Ultimately, Parabricks got an MTRAC grant of $100,000.

“I’ve been watching them a long time. They were able to realize there were better markets with a stronger business case. They were extremely coachable,” said Glaza. As a result of their pivot, Invest Detroit became an investor, as did Invest Michigan, another nonprofit, and UM’s Monroe-Brown Seed Fund.

The company was also able to enlist Dave Gregorka, an experienced and well-regarded local entrepreneur and venture capitalist, as its president.

The Monroe-Brown Seed Fund is the largest investor, with a first investment tranche of $50,000 and a second of $150,000. Invest Detroit invested $65,000 and Detroit-based Invest Michigan invested $50,000. Ann Arbor Spark, an economic development nonprofit, awarded a grant of $50,000.

In all, the company has raised $315,000 in equity funding in a seed round it is targeting at up to $1.5 million.

Perhaps more important than those investments is a partnership with Santa Clara, Calif.-based Nvidia, which marketed the first GPU chip in 1999. The partnership began when Parabricks attended the company’s GPU Technology Conference in San Jose, Calif., in May 2017. Nvidia executives were impressed by Parabricks’ technology, eventually naming it as a preferred partner and introducing it to three of its genomics customers, in Thailand, Japan and Singapore, who are now paying customers. Samadi said nondisclosure agreements prohibit him from naming them.

Monroe-Brown was created more than two years ago with a $3 million endowment from the Monroe-Brown Foundation to commercialize faculty research. Hirak Parikh, the program manager at Monroe-Brown, says the fund has invested about $1.6 billion in nine companies, with have in turn raised $12 million in other equity funding.

“I’m really excited that Parabricks is focusing on genomics, which is going to be a really big and we want to get there early. Eventually, everyone’s genome will be sequenced,” Parikh said. “Parabricks speeds up the genomic-sequencing process 48 times, and when you tell people that, they just don’t believe you. The founders are great coders and they have an extremely good partnership with Nvidia, the lead manufacturer of GPUs. And Dave Gregorka brings a ton of connections.”

Samadi said that what takes 48 hours of computing power to sequence can be done in less than an hour with their technology.

Parikh said the fund will likely invest more in the company as needed. “I have high hopes for them. They are going to be a breakout company and get acquired,” he said.

“Through all our discussions, they were very focused on the market, who potential customers are and how to reach them. That’s important for a university spinoff,” said Charlie Moret, Invest Michigan’s chairman and CEO, who invested in Parabricks last year. He said the partnership with Nvidia can’t be overrated. “It’s a go-to-market partner that accelerates your path to market. For an early stage company, Parabricks is hitting all the milestones you like to see.”

Gregorka, the Ann Arbor-based venture partner for the Chicago-based venture-capital firm of Baird Capital, is also a mentor-in-residence at UM’s tech transfer office. The key for him was the pivot to genomics.

“My reaction was they had a very strong value proposition in a space that was growing very rapidly. Genomics is just exploding,” he said. Parabricks offered, he said, huge improvement in efficiency at a dramatic reduction in cost. “And the relationship with Nvidia was huge. That helped convince me.”

Parabricks, which has five full-time employees and two part time, is based in the A2 Startup Garage, an incubator facility that shares space in the downtown Ann Arbor headquarters of Menlo Innovations, a software development firm. Rich Sheridan, Menlo’s co-founder and CEO, was impressed by Parabricks when it was one of 36 semifinalists in the 2016 Accelerate Michigan Innovation competition in Detroit and offered the company one year of free space in his incubator. It has since become a paying tenant.

“There are two big trends in the technology world, speed and data. Data is exploding, particularly in the area of genomics. When data explodes at the rates we are seeing, speed becomes critical and this is where Parabricks comes in, turning computing processes that currently take days to ones that can be done in minutes and hours,” he said. “Using GPUs to their fullest extent is the future of computing, and the founders of Parabricks are among the brightest minds in this field of computing. Quite frankly, it is one of the joys of being in Ann Arbor, nestled up against one of the most prolific research universities in the world, that we get to hang out with guys like Mehrzad and Ankit.”

“I’ve known Mehrzad and Ankit since before they launched the company. They are a couple of the most humble and knowledge-hungry young entrepreneurs I’ve met in a long time,” said Adrian Fortino, managing director of the Ann Arbor office of the Houston-based Mercury Fund. “We spend a lot of time on cloud-biology investment themes, and they have built a truly impressive platform to accelerate analysis of genomic sequencing data.”