By Juan-Antonio Carballo
Global semiconductor venture capital investment remains healthy at around a billion dollars a year. EDA is approximately at an order of magnitude below, which is even healthier in relative terms. Historical data shows that VC-backed semiconductor ventures are still successful, or at least not significantly less successful than other types of ventures. Indeed, about a quarter of silicon startups have been acquired, and about an eighth of them have issued a successful public offering, percentages that are quite close to venture-backed companies in general.
But something doesn’t quite add up. Basic technology and economic trends in silicon systems seem to make it impossible to gain a reasonable return-on-investment for an average venture investor, as multiple incompatible goals stack up, including numerous cost components and steep risk factors.
Consider the example of a system-on-chip venture, which will on average swallow around $50 million in capital outlays, almost double the five-year average for all venture-backed technology companies. A typical investor will require an expected return after investment of at least 5X, and preferably 10X. Yearly sales of such company, using generous multiples, would need to be over $50 million to achieve such a valuation upon exit, either by a sale or IPO.
Consumer markets, with their huge volumes, have been favorites of investors, but their unit pricing is often very competitive — $20, $10, even under $5 for certain geographies and markets. This places the required yearly chip volume in the tens of millions, a huge challenge for a company that was built from scratch in an acutely competitive and dynamic market.
Yet the venture capital community can be creative — and manage not to abandon such a critical piece of the quarter-trillion-dollar semiconductor market. Emerging venture capital trends, including globalization-driven collaboration and application-driven specialization, are bringing new life to silicon ventures. This emerging class of IC ventures seems to be focused on efficiency. Indeed, despite claims of skyrocketing tape-out costs and shrinking market windows, historical financials show that venture-backed semiconductor startups have not dramatically changed the number and size of investment draw-downs over this decade, on a per-company and per-stage basis. And fortunately, acquisitions and public offerings are happening at a reasonable pace.
How is this possible? Consider the impact of globalization on venture capital and start-ups. Emerging global players have surged. Between two and three billion dollars of venture capital — and growing — are being now routinely invested in emerging Asian regions, reaching over a quarter of all new investments, and quickly extending into areas as diverse as Russia, the Middle East, and Western Canada. Globalization is especially attractive for new players, since emerging geographies have numerous “white spaces” such as mobile TV not yet dominated by large players, are friendlier to high-growth open standards, and cover a broad set of diversified growth sectors.
Perhaps more importantly, today’s typical venture-backed semiconductor player is global since inception, from an operational, organizational, and financial standpoint. An interesting development has happened with “bi-coastal” start-up organizations — with two offices on two coasts — becoming “global,” as venture-backed companies quickly reconfigure themselves to place every function at its optimal geographical location, indeed beyond the traditional US/Asia or EU/Asia split.
Additionally, an intriguing specialization trend is developing, where special-purpose, often multi-core, architectures that leverage the concept of “accelerators” achieve very difficult requirements using repeatable and proven silicon platforms. Combined with a high level of openness and standardization, the seemingly-impossible capital outlays appear controllable to a point. Unfortunately, clouds remain in areas such as software, a yet-unsolved productivity problem.
A parallel specialization trend in venture investment is the funding of highly differentiated, high growth sectors covered by analog and mixed-signal components. While still very labor intensive, these markets can enable huge volumes, can be defended with a strong intellectual property and talent positions, and do not easily require enormous verification and software teams.
What does one of these new “global” silicon ventures look like? Typical characteristics include small “lite” teams that radically outsource much of the design, special-purpose architectures largely based on existing platforms, and very aggressive leverage of new design technology and IP, including system-level design methodologies. These ventures will partially compensate for the constant decline in venture-backed silicon starts and potentially mitigate the oft-mentioned death of the silicon startup.
— Juan-Antonio Carballo is a general partner at Argon Venture Partners, Redwood City, Calif.