2019 was an extremely important year for the semi industry in Israel. It started off with the purchase of Mellanox by NVIDIA for $6.8 billion in March and continued with the purchase of Habana Labs by Intel for $2 billion in December. These two deals alone constituted nearly 35% of the total exits in Israel that year and are the latest examples of an ongoing renaissance in the semi industry.
Globally, both the public and private markets are also bullish about the semi sector. The semiconductor SOXX (+202%) ETF has vastly outperformed the S&P 500 (+69%) since 2016. We’ve also seen an incredible rise in global funding activity in the industry over that same time period – from $1.2B in 2016, rising to $5.3B in 2018 and then tapering off to $2.1B in 2019.
Semiconductor & Info-tech ETFs outperforming S&P 500 over recent years.
Global funding activity in semi sees massive jump in 2016.
This renaissance is coming as the industry is faced with an exponential increase in the requirements of both compute capacity and compute complexity. The need for increased capacity comes from the staggering rise in the amount of data being computed, while the need for increased complexity comes from the progressively more sophisticated neural networks that are essential for the implementation of AI/ML.
Moore’s Law, which predicted a doubling of transistor capacity every 18 months, is starting to fail. This is actually leading to an even greater opportunity for startups. Here are the three main solutions startups can employ to close the gap between what the current industry is able to provide and the market needs: Problem-specific design (optimizations at the system level), architectural innovation (solving problems at the component level), and fundamental tech innovation (replacing semi tech as we know it).
Related: VIDEO – Semi Summit Tel Aviv 2020
Not all these solutions are equal. Multinational Corporations are excellent at problem-specific design, which requires incremental advances, but startups can shine in the more drastic innovations – in reimagining the architecture of a chip (different types of accelerators) or fundamentally reinventing semi technology (quantum computing, DNA storage).
But startups should also be aware of the challenges that exist in the semi space. Tectonic shifts are disrupting the industry, as giants are consolidating, and corporations are becoming increasingly adept at innovating from within, especially the hyperscalers.
This era has opened three main market opportunities for semi companies:
1) Next-Gen Datacenters
There has been massive consolidation in this segment. The hyperscalers are growing and creating huge hardware markets. But they’re not just growing as customers, they’re growing as designers too. Companies like Amazon and Google are setting up their own world-class design teams, leading them to compete with server OEMs and chip vendors.
In order to compete and thrive in this sector, we believe startups will need to have a more long-term outlook. Simply beating the current market solutions won’t cut it – they’ll need to reach 10x+ improvements just to gain market traction, which will lead to semi startups having higher risk profiles and requiring more funding.
2) Smart Machinery
There is a growing need for functional platforms that have to work independently in the field, usually by employing sophisticated, real-time algorithms that are very sensitive on latency. The biggest market in this area is automotive (as this industry transitions to autonomous), but fields like industrial robotics and drones are also up and coming.
Smart Machinery solutions, especially automotive, usually have a long time to market, requiring companies to be efficient and well funded. Additionally, there is a strong price pressure in these high-volume markets. In order to compete with the incumbents that have a strong capability to push down costs, we believe that startups should come with solutions that not only improve performance but also enable new functionalities that will be strategically valuable for the customers.
At least in the Israeli market, we see that Datacenters (AI chips) and Smart Machinery (automotive) have captured the bulk (71%) of the overall investments in semi companies in 2019.
3) Edge Computing
There are two trends playing out in this field: an exponential increase in sensors and the requirement for lower latency. Both of these will provide space for innovation in edge computing.
The implementation of 5G will accelerate innovations in this field, as companies try to reduce reliance on the cloud and the latency issues that entails. In a way, the cloud is getting closer to the end user, and although this trend is in its infancy, we believe that this category has a lot of potential to grow.
The VC Perspective on Semi Startups
There are two factors (above and beyond the usual) that we take into consideration when funding startups in the semi space – Funding and Team.
Funding – Semiconductor companies take a unique evolutionary path that affects the way they’re funded. In the early rounds (Seed and A), funding is 44% more capital intensive than other industries. In most cases, and unlike in other industries, semi companies create value for the A Round mainly by advancing the technology, not by gaining actual market traction. And the more entrepreneurs aim for ambitious technologies, the more significant this rule will be.
Semi funding is capital-intensive in Seed, A Rounds, but less so in later rounds.
In many cases, it takes more time than initially planned to develop technology, making it extremely important to choose a seed/A round investor – or a syndication of investors – with deep enough pockets to carry the company to later funding rounds. In planning for the next fundable stage, it’s also important to find a mid-term milestone that would reflect value creation and open an opportunity for additional fundraising (pre-tape out).
From Viola’s perspective, we want to fund semi companies early on. In fact, we try to be the first investor in the company, instead of waiting until after the technology has been proven (and valuations are already high) but go-to-market risks still exist.
Team – There are two key characteristics of a semi startup team that we pay special attention to: specific technological know-how in the area of innovation and a broad network in the semi space. The knowledge and experience of the team are crucial for startups to achieve their ambitious technological goals within a timeframe and cash burn rate that makes sense in terms of funding.
In semi, more than other industries, there is a war for human capital, and especially for the top-tier talent. Startups do not have the resources to compete with multinational corporations in terms of compensation, so prior relationships and charisma of the entrepreneurs are deciding factors in their ability to attract top talent. In addition, the relationships within the industry will allow companies to find the right customers, understand their needs, and let them shape the product.
Israel’s Semi Renaissance
Although the semi renaissance is certainly a global phenomenon, Israeli startups are enjoying a unique funding ecosystem. The country now ranks #3 in the world, behind only the US and China, in terms of investments in the semi domain in 2019, with 42% of the total semi investments of the US and 69% of the total investments of China. There are very few segments in which Israel is investing nearly as much as the two global superpowers.
Israel keeping pace with global superpowers in semi funding.
As I said before, 2019 was an exceptional year for Israel’s semi industry, with the Mellanox and Habana Labs deals. We don’t believe it was an anomaly, but a trend that will continue into the coming decade.
It’s the best time ever to start a semiconductor company.
At Viola, we’re always on the lookout for semiconductor startups. If you’re doing something interesting in the space, feel free to reach out at zvikao@viola.vc.