10x or Bust
Developing step-change solutions is the only way to build successful and highly profitable frontier tech companies.
In his 2017 book, Zero to One, Peter Thiel consistently reinforces the need to build products that are 10x better than currently existing solutions. This is one of the oldest concepts in venture capital, but is often forgotten in overheated markets where investors are prone to chase fads (e.g. solar start-ups in the late 2000’s or scooter start-ups in the late 2010s).
In Frontier Tech, this is an especially important concept when building new companies or investing in new technologies. In many cases, Frontier Tech companies are building products for markets where the competition has been established for years or decades. Switching costs are high, and industry learning curves constantly advance. When coupled with the fact that it could take several years to bring a Frontier Tech product to market, it becomes obvious that companies must start off with a massive, step-change advantage over competition. Otherwise, they’re bound to leap frogged by the time their products get to market.
This isn’t just a question of success or failure though: the stronger the value proposition when products get to market, the bigger the value capture for a start-up. Naturally, step-change solutions provide more value to customers, and will therefore command higher premiums and stronger margins. When built correctly correctly, Frontier Technology companies can command margins that approach those of SaaS companies.
To better demonstrate the need for step-change innovation, we’ll take a look at two examples: a first-hand account of the solar bust in the late 2010s, and a simple mathematical example using Moore’s Law.
The Solar Bust
In the mid to late 2000s (before the US experienced a glut of oil and gas brought on by the fracking revolution) the belief that the world was rapidly running out of fossil fuels led to a boom in clean energy investment. One of the largest recipients of these investment dollars was the US solar industry, where hundreds of solar start-ups sprung up overnight all promising to make solar electricity as cheap as coal.
However, a lot of the “novel” solar technologies being developed were incremental improvements compared to the existing dominant solar technology: silicon. Additionally, they were chasing a cost metric ($1/Watt solar panel cost) that was naturally eclipsed by silicon solar technology coming down the cost curve more quickly than anyone expected. In the end, it was not a fundamental technology innovation that broke this cost barrier — it was simply scale (most of which came from Chinese manufacturing capacity). In essence, venture investors were pouring billions of dollars into hundreds of incremental solutions, nearly all of which failed.
This series of events represented an industry-wide ignorance of one of the oldest axioms in venture capital: the need for 10x or step-change innovation. It also led to a bust in the cleantech industry that resulted in a black eye so massive, it’s taken a decade for capital to return to climate and sustainability.
Moore’s Law — The Most Famous Learning Curve
When building new products, many frontier tech entrepreneurs fail to consider industry learning curves for their competition. They compare their own future, aspirational metrics, like cost or efficiency, to metrics from today’s products. However, today’s products are constantly getting better due to a natural phenomenon called a learning curve.
The most famous learning curve is Moore’s Law, and it states that computer chips get 2x faster and cheaper every 18 months. (Note: While there is evidence that Moore’s Law is coming to an end or at least meaningfully slowing down, we assume it still holds for the purposes of this example.) Because development timelines for new semiconductor technologies (such as machine learnings inference/training chips) can be 2 to 5 years, understanding not only where the industry is today, but also where it will be in the future, is critical.
A simple calculation shows that in 4.5 years, existing chip technology will naturally be 8 times faster/cheaper than today, simply due to inherent industry improvements. If a start-up is building a chip that promises to be 5x better than today’s technology, there’s a good chance it will have no advantage, or even be behind, by the time the product gets to market.
On the other hand, if a start-up is building a chip that will be 100x or 1000x better than today’s technology, that start-up will still have a meaningful advantage once it gets to market, even if it misses on its performance goals. This is just one example of a learning curve, and it’s important for entrepreneurs to understand the specific learning curves of their industries.
Summary
At Bison Ventures we strongly believe that backing 10x, step-change innovations is the only way to build successful Frontier Tech companies. A step-change improvement not only increases the chances of success for a Frontier Tech company, but it also increases the likelihood of meaningful value capture and the high margins that result in a very profitable company. While this may seem obvious, it’s an axiom that many have forgotten at various points in time. At Bison Ventures, this is engrained in our investment process due to the learnings we’ve seen firsthand, both good and bad.
