A Contrarian View of Moonshots: Small Steps to a Giant Leap
Each new invention moves the adjacent possible forward and unlocks opportunities for future inventions and thus new market opportunities. Ramu Arunachalam

Every startup begins with an idea and a plan. The idea (for a new product or service) describes the “what”; the plan describes the “how”. We associate genius with singular acts of idea creation, eureka moments that bring to the surface new discoveries and insights. But great startups are often distinguished more by the genius in the “how” than in the “what”. Google didn’t come up with the idea of a search engine, but its genius lay in how it made search more relevant than anything before. Facebook didn’t invent social networking but it figured out how to build a vibrant social network of highly engaged users. In the world of technology the how is an engineering exercise to find and put together the right set of building blocks that make a product. It is a creative exercise for sure but one dominated by experimentation and tinkering. As Edison put it “Genius is 1% inspiration and 99% perspiration”. Of course, the “parts” required for the engineering miracle need to exist in the present. Sometimes the parts are so far out in the future, the idea remains an abstract distant dream. In the late fifteenth century, Leonardo Da Vinci came up with the idea of a helicopter (he described it as an “axial screw made of linen of which the pores are stopped up with starch ”). It took almost 450 years before we had the right technologies to actually build a helicopter. Charles Babbage conceived of a programmable computer in 1837 made of wood and metal gears, a design so complex and impractical that it was never actually built. It took another 100 years with the invention of vacuum tubes (the precursor to transistors) before we could build a computer.

Ideas are only constrained by the limits of human imagination—and if you believe human imagination is infinite then the set of possible new ideas is truly limitless. The act of translating an idea into a product, however, is constrained by the parts available in the present. The author Steven Johnson calls this constraint the adjacent possible which he describes as “a shadow future, hovering on the edges of the present state of things, a map of all the ways in which the present can reinvent itself”. We went from transistors to integrated circuits to mainframes; the iPhone was many inventions and decades away — well outside the adjacent possible in 1947 (the year transistors were invented). At any given moment, as Johnson puts it, “the world is capable of extraordinary change, but only certain changes can happen”.

One way to think about invention and technology entrepreneurship is as a continuous exploration and commercial exploitation of the adjacent possible. Each new invention moves the adjacent possible forward and unlocks opportunities for future inventions and thus new market opportunities.

Peter Thiel likes to start with the question “What valuable company is no one building?” You can approach this question bottoms-up by searching the adjacent possible for new sets of technologies and skills “hovering on the edges of the present” and linking them to product ideas. Alternatively, you could start top-down by looking at failed ideas from the past or fantastic creations from science-fiction and ask which one of these “crazy” ideas can be cobbled together with emerging technologies in the adjacent possible?

Whatever the approach you take, breakthrough product inventions are “moonshots” that pull together multiple technologies, all simultaneously maturing from the adjacent possible into the present. Take the first Apple iPhone as an example. It was a remarkable engineering and design feat that brought to the forefront a number of technologies from the adjacent possible. The core technologies that went into the iPhone (see below) existed well before 2007 (and Apple didn’t invent any of them) but the iPhone would have been impossible to build even a few years prior.

Enabling Technologies to the iPhone

  • Fast, power efficient mobile processors

  • Mature web standards and protocols

  • Touch screen LCD displays

  • Small form-factor digital camera technology

  • Battery technology

  • Wifi
  • GPS
  • Sensors (accelerometer, proximity, light)
  • 2G/3G/LTE network infrastructure

It took the right set of technologies all hitting price, performance or scale inflection points at the right time that made the iPhone possible.

Every moonshot comes down to two related bets. The first is a bet that the technology will work as promised outside the lab. The second is a bet that the technology will get better fast enough to be commercially viable. The first bet is easier to assess: either the technology works or it doesn’t. The second bet is much trickier to evaluate because it involves extrapolating the trajectory of a technology improving at an exponential rate (e.g. Moore’s law for computing). Human perception, however, is linear; our intuition leads us astray when analyzing exponential changes. In particular, early in the life of a new technology, we tend to overestimate the rate at which it improves (or gets adopted). But once the technology hits an inflection point, we underestimate the true impact of the technology.


When we’re in the zone of overestimation we run the risk of bringing products to market before they’re ready for broad consumption (Apple Newton for example). Many failed startups from the 2000 bubble were victims of overestimation (good ideas ahead of their time, overestimating things like broadband bandwidth growth, internet adoption, mobile penetration etc.). Conversely, when we’re in the zone of underestimation, we are late into important new markets (for example, IBM ceding the PC market to Microsoft and Intel).

It may be tempting to conclude that successful commercialization of breakthrough products comes down to pure timing and chance. Every moonshot can seem like buying a lottery ticket. But in fact the opposite is true. Smart companies realize they have much more control over their own destiny than is obvious. They don’t think of moonshots as all-or-nothing experiments but rather as multi-year, multi-product roadmaps — a series of incremental steps inching towards the final destination. Their business models are “multi-stage” rockets, each stage “firing” only when the technologies required to build the product can be found in the adjacent possible. You can see this in Apple’s roadmap going from the iPod (2001) to the iPhone (2007) to the iPad (2010). Apple’s roadmap was effectively a hedge against the risks of technology extrapolation. Predicting the right time to introduce an iPhone in 2001 was a risky proposition. The same prediction in 2004 with the benefit of all the experiential learning from the iPod was a much safer bet.

When VMware invented PC virtualization technology in 1998, it was too slow to run datacenter applications so VMware instead focused on the test & dev market where the performance requirements were relatively modest. As its technology got better and hardware got faster, VMware entered the datacenter market in 2001. It took another six years before virtualization technology paved the way to cloud computing. If VMware had instead chosen to push cloud computing early in its existence, it would have failed like so many others did in the early 2000s (Exodus, Loudcloud to name a few).

Once you recognize this pattern of multi-stage businesses you see it everywhere. Netflix started as an internet DVD business that pivoted into streaming once broadband networks got faster.

Amazon is perhaps the most extreme example. It started as an internet bookstore that then became an ecommerce mega-store. It now has a cloud computing business (AWS) that is growing so fast the some have referred to it as a cloud computing company that also happens to sell books. This is not to suggest that Jeff Bezos in 1996 thought of Amazon becoming a cloud computing company someday. The point here is that smart companies are always mining the adjacent possible, looking to connect the dots between new technologies, their core competencies and moonshot product ideas. In Amazon’s case, the process of running a large internet scale business gave them a front row seat on emerging cloud technologies. It was less extrapolation and more repackaging of an existing internal cloud for public consumption.

There are many lessons for startups here. Being early to market can be just as catastrophic as being late. Extrapolating technology trajectories is exceedingly hard. Winning companies are based on multi-stage business plans that are robust to these extrapolation risks. You have to think big, but progress incrementally. It may not be the most popular thing to say in Silicon Valley, but slow and steady wins the race.