SAN FRANCISCO--Startups emerging from Harvard University and University of
Pennsylvania invariably flee Massachusetts and Pennsylvania to grow their businesses, a
disturbing trend for communities and governments nearby that want to
build tech-savvy economies in the coming decades.
That's one disturbing nugget from a new 60-page
study (free registration) from the
venture-capital research firm CB Insights.
To be sure, it's not all doom and gloom. Harvard and Penn are two of
six major universities--the others are Stanford, U.C. Berkeley, MIT
and NYU--that the researcher examined to determine how many startups emerge from universities, how much venture funding they attract, the type of technology they advance and where those startups ultimately locate.
Startups from Stanford, Berkeley stay in California, partly due to the Silicon Valley's appeal.
The study found that among those schools, $12.6 billion was invested
in 559 deals involving alumni between 2007-2011. For context, that's
a fraction of the $135 trillion in venture capital invested overall
during the same five years.
Stanford is the big man on this campus, taking in $4.1 billion
in 203 deals since 2007.
Stanford and Berkeley startups almost without fail locate in
Berkeley and MIT churn out the most technologically diverse
startups; NYU and Penn, by contrast, nurture startups heavily
focused on software and healthcare.
Harvard and NYU are enjoying the fastest growth rates (31 and
71 percent CAGR, respectively), while growth rates at Penn (-10)
and Berkeley (-8) have slid since 2007, according to the report.
It is unfortunate that so much trust is put into the numbers from this report. These CB Insights numbers seem as accurate as the Startup Genome project, which is not very accurate. We have tried to contact CB Insights numerous times to understand how they got them because they simply don't make sense. We have gotten no call back unfortunately. As soon as we saw them, we did a quick back of the envelop calculation of what we knew off sub-set of MIT companies and found them to be off for MIT by at least 2.5x. This calls into great question the data. We want to promote entrepreneurship and are much less wrapped up in the ratings headlines game than we are about wanting to learn something to improve our service to out students. For this to be possible from these reports, the data has to be credible which it is not from what we can tell. It is also true that we measure output and not inputs at MIT as they are much more important so the proper measurement is economic impact (companies, jobs, revenue, etc.). This is not to say that Stanford is not wonderful - it is and we all work together to promote entrepreneurship for our students. It is a common cause that unites us. All I will say is that in this thirst for "rankings", we are unfortunately seeing a lot of unrigorous data analysis.
David Patterson, known for his pioneering research that led to RAID, clusters and more, is part of a team at UC Berkeley that recently made its RISC-V processor architecture an open source hardware offering. We talk with Patterson and one of his colleagues behind the effort about the opportunities they see, what new kinds of designs they hope to enable and what it means for today’s commercial processor giants such as Intel, ARM and Imagination Technologies.