Advanced Computing in the Age of AI | Monday, January 30, 2023

Is Entrepreneurship Flagging in America? Not in HPC 


HPC has always been a fountain of innovations that eventually become mainstream. These innovations have been both technological as well as in business models. HPC’s long history of fostering entrepreneurship and creating new markets is well established. What we see now is a growing opportunity for HPC practitioners to lead the new wave of products, applications, and services.

The IT industry is being transformed by cloud, social media, big data, Internet of Things, and artificial intelligence. All of these trends require sophisticated computational models and highly scalable systems. HPC capabilities and skills are critical to the companies that want to establish leadership positions in these sectors.

However, there has been increasing commentary around a slowing in the rate of business formation in the US during the past 15 years.

In a paper late last year from the National Bureau of Economic Research titled Where Has All The Skewness Gone? The Decline In High-Growth (Young) Firms In The U.S. the authors state:

“The pace of business dynamism and entrepreneurship in the U.S. has declined over recent decades. We show that the character of that decline changed around 2000. Since 2000 the decline in dynamism and entrepreneurship has been accompanied by a decline in high-growth young firms.”

They mapped growth rates and measured the dispersion and skew, which are interpreted as measures of dynamism. Dynamism has been decreasing in both public and private firms since 2000. They determined that:

“Prior research has shown that the sustained contribution of business startups to job creation stems from a relatively small fraction of high-growth young firms…We show that the shape of the firm employment growth distribution changes substantially in the post-2000 period…The overall decline reflects a sharp drop in the 90th percentile of the growth rate distribution accounted for by the declining share of young firms and the declining propensity for young firms to be high-growth firms.”

In a word, there have been fewer high-growth firms especially after 2000 (in comparison to the 80s and 90s). The startup rate has been in secular decline.

One also has to examine the main two types of new businesses: Mainstream (e.g., family businesses in well-established sectors) vs. Transformational (e.g., R&D intensive high-tech with a big idea). The latter category is primarily responsible for job growth and productivity increases. However the authors have found that even Services, Information and High Tech sectors are becoming less dynamic, based on the decline of growth rates.

“We don’t yet have a good explanation for the decline in dispersion and skewness we observe in the high tech sector…Something has happened to the incentives or the ability to be a high-growth firm in the high tech sector.”

Kauffman Foundation findings are similar. In the case of startups by young Americans:

“Kauffman’s report found that 35% of young Americans aged 20-34 launched startups in 1996; in 2013, that percentage dropped to 23%.” –

Some possible reasons for the slowdown are:

  • Slower growth economy
  • 2001 shock
  • Great Recession shock
  • Aging workforce, younger people more likely to do startups
  • Gig economy – lessened growth via outsourcing to contractors
  • Regulation sprawl (larger firms can cope better with this)
  • Company growth increasingly driven by automation and international expansion
  • Millennials more risk-averse than Baby Boomers (living in the basement and doing a startup vs. just living in the basement). Counterculture movement convinced Boomers they could change the world.
  • Student loan debt levels high for Millennials

But as it relates to HPC, we believe another reason is a

  • lack of targeted education and community support specifically focused on entrepreneurship.

This is the gap that a grass roots effort called StartupHPC aims to address.

In earlier days, HPC startups generally meant new system hardware companies. Such opportunities remain. The slowdown and shift in Moore’s law, combined with new applications in data intensive computing, and accelerated computing have created new needs for system-level innovation.

Today, there are also many opportunities that are HPC-inspired but not confined to addressing the HPC market itself. For example, deep learning and computational analytics are capabilities that are mixed with emerging consumer and industrial products and services.

The StartupHPC effort is focused on community support and practical entrepreneurship education and advice for HPC practitioners. This includes advice on emerging markets, sources of funding, interactions with venture capital firms, building a team, general management skills, go-to-market strategies, etc.

Now in its third year, the next StartupHPC Summit will be on Monday November 14 at the Grande America in Salt Lake City where the SC16 conference is held this year.

We believe such efforts are necessary to foster and accelerate entrepreneurship among HPC practitioners and to help HPC lead in the emerging digital economy.

StartupHPC is a grass roots community for STEM- and HPC-inspired entrepreneurs, venture capital firms, academia, government and support organizations.

About the Author

stephen-perrenod-headshot-400xStephen Perrenod is an energetic blogger on IT and science issues, and a published author of two well-received books on cosmology and extragalactic astronomy.

He possesses expert knowledge in key market sectors for enterprise computing including cloud computing, big data, IoT, deep learning, and HPC. At OrionX he works with Fortune 500 companies, SMEs and startups to enhance their business strategies in enterprise computing.

He has held a variety of senior sales and marketing positions at Cray, SGI, and Sun Microsystems, as well as for two startups. In these roles he has been responsible for business segments in the Americas, Europe and Asia-Pacific. At Sun Microsystems he doubled the HPC business for the Asia-Pacific region within a 3 year period.

He holds a bachelor’s degree in physics from MIT and a masters and Ph.D. in astronomy from Harvard University.

Add a Comment