The Selfies (2011-16)

The Virtual Clusters

What seems unique, both in history and today, about the San Francisco Bay Area (which most of the world simplistically calls “Silicon Valley”) is its ability to excel in so many different technologies, unlike the traditional industrial cluster that excels at only one; and the ability to create world leaders in relatively short time scales. Small businesses are welcome, protected, nurtured and sometimes catapulted to world celebrity not by government planning but by the very nature of the system.

Even after Silicon Valley became home to the world's most valuable corporations, innovation still came mainly from the small beginners. For example, in the mid-2010s Google was mostly famous for Maps (navigation), acquired from a small startup in 2004, Earth (aerial images), acquired from a small startup in 2004, Doubleclick (ads), acquired from a small startup in 2007, Youtube (video sharing), acquired from a small startup in 2006, Android (smartphone o.s.), acquired from a small startup in 2005, Robots, acquired from a small startup in 2013, DeepMind (A.I.), acquired from a small startup in 2014, and Nest (IoT), acquired from a small startup in 2014. Gmail was the only major success developed internally by Google after becoming a large corporation. Facebook's only major innovations since becoming a corporation had all been acquired from small startups: Instagram (acquired in 2012), Whatsapp (acquired in 2014) and Oculus (acquired in 2014). Apple had developed its hardware internally, but acquired the most popular software features: iTunes (acquired in 2000), Siri (acquired in 2010), the fingerprint scanner (acquired in 2012), the biometric identification Face ID (2013), Apple Music (2014), Animoji (2015), etc.

Silicon Valley is clearly not a “cluster” in the Michael Porter sense of the word: “A cluster is a geographically proximate group of interconnected companies and associated institutions in a particular field” (“Clusters and Competition”, 1998). Porter’s cluster relies on network relationships within a geographical community. Clusters foster sharing of knowledge, collaboration (or, better, complementarity) and competition. Example of clusters are Detroit for automotive manufacturing, the French wine producers, the fashion firms in northern Italy… Silicon Valley may have been a cluster (of semiconductor firms) in the 1960s but by 1998 it had diversified so much that it was not clear anymore what kind of specialty identified this cluster. Today the San Francisco Bay Area has success stories in too many different industries to be considered a cluster.

The globalized economy and the knowledge society have greatly altered the nature of collaboration and competition. Venture capital, graduate education, immigration and government funding have disrupted the natural formation of clusters. Clusters foster collaboration and competition, but globalization has made both ubiquitous. The knowledge society (universities, scientific journals, conferences, Internet, open source movement) has made technological know-how readily available. Access to information is free and often instantaneous.

Since talent mobility across borders has become the norm, venture capital can attract talents from anywhere; and, in turn, talent attracts venture capital. The geographical location is no longer a big advantage: access to capital and talent is. The Bay Area attracts capital and talent,

The Bay Area should more properly referred to as a location that can simulate any kind of cluster, that can simultaneously create multiple “virtual clusters” across the globe. Apple’s phones are manufactured elsewhere, Google’s engineers are spread in dozens of countries, etc. The Bay Area excels at simulating clusters, any kind of cluster.

A virtual cluster is far more complex than the traditional industrial cluster. A virtual cluster needs an unlimited supply of new ideas, of capital and of talents in order to churn out a large number of startups in all sorts of fields. The virtual cluster differs from the traditional cluster because experience from one field can betransferred to other fields; because talents from one field cn be merged with talents from other fields. It naturally achieves “convergence” of technologies. It naturally triggers the rapid adoption of new technologies and science.

The virtual cluster places emphasis on horizontal platforms rather than vertical applications; platforms (such as the microprocessor, the personal computer, Google, Facebook, the iPhone) that create whole new economies and change the world. The virtual cluster facilitates much faster adaptation and growth of successful firms. The virtual cluster inevitably nurtures variety of everything: multiple types of startups, multiple types of investors, multiple types of exits, multiple types of technologies, multiple types of company organization… multiple types of everything.

There was no “typical” startup in the Bay Area. Startup types varied across the whole spectrum: the hobbyist’s project (e.g. Apple, Ebay, Twitter, Paypal); the university spinoff (e.g. HP, SUN, Cisco, Vmware, most biotech firms); the corporate spinoff (Intel and most semiconductor firms including Nvidia, Oracle); the dormitory project (e.g. Google, Facebook); and the serial entrepreneur startup (Tesla, Uber). The only rare startup in the Bay Area was the government-funded startup. Investors ranged from venture capital firms (the “institutional” investors) to angels, from successful entrepreneurs to corporate funds (e.g., Intel) and universities (e.g., StartX). The possible “exits” too were multiple and all tolerated: first of all failures (the fate of the vast majority of startups), then firms that remained private forever (many small and medium-size firms), then those acquired by bigger companies (Instagram, Youtube and so on), and finally the ones that made it to an IPO (a tiny minority).

The other unique aspect of the Bay Area ecosystem in the 2000s was its ability to create monopolies in each field. The generation of Hewlett-Packard consisted of small and mid-size firms, not even remotely capable of becoming monopolies. The generation of Fairchild and Intel (the "silicon" generation) was mostly a cluster of semiconductor firms competing with each other (and, eventually, with the Japanese). Apple in the 1980s was, like Hewlett-Packard, a mid-size firm, and a specialty firm, and Oracle was fighting a bigger (German) competitor. But the post-dotcom generation of Google and Facebook created not only worldwide leaders but also monopolies: by 2019 Google's search engine had only nominal rivals in the world (expect in China where it was banned), Facebook's social network had only specialized rivals (again, except in China where it was banned), Airbnb had no significant competitor, Uber had only one significant competitor (also based here), Linkedin had no competitor, Tesla had only competitors in China, Nvidia had no significant competitor, Paypal had only one significant competitor (also based here), and Dropbox had only one significant competitor (also based here). Apple, Vmware and Oracle had become world leaders in their fields and quasi-monopolies, but not quite as good at eliminating competitors as the new breed of Bay Area-based world leaders. These younger firms had learned a way to establish an overwhelming dominance early by a brute-force approach that involved discouraging engineers and investors from starting rivals and buying out those who did and survived before they could become a threat.