I was recently asked: “If you had a spare dollar to invest in 2003, would you invest it in communications semiconductors?” On the surface, that seemed easy to answer, but of course things are seldom what they seem. Depending on whether you are an optimist, a pessimist or a schizophrenic, you can safely land on either (or both) sides of the fence, and it will still take years to figure out who was right in the end.
It didn’t take me too long to come up with a laundry list of reasons to avoid this market like the plague. Here’s a quick summary:
The Pessimist’s View
It takes approximately $100 million to start a manufacturing process technology company, and even $15 million to $30 million for a fabless company. Leaving some room for management, by the time you get to ship samples to generally distressed customers who may buy someday, you are sitting at a $20 million to $50 million post-money valuation. You have to sell a lot of sand to get a payoff on this bet. By comparison, enterprise software startups can go a long way with that first $15 million.
Developing a communications chip is inherently risk. I keep looking for the company that invented the paperclip: It didn’t take long to engineer and they must have made a bundle. Unfortunately, most breakthroughs in communications semis are just the opposite. They’re really tough, complex challenges fraught with development risk. As you move from .18 microns to .13 and downward, accelerating NRE and mask costs raise the cost of making a mistake. Re-spin an ASIC once or twice and the bills really start to add up.
Customers are broke-and risk averse. It’s no secret that most well-established equipment companies are becoming more unwilling to buy from startups. For example, Cisco announced in 2002 that it would actively seek to consolidate its vendors and push them hard on pricing.
In addition, startups face their own forecasting nightmares. During my last due diligence check, the best reference said: “My 90-day forecast is subject to plus or minus 100% variation, and the only thing I will commit to is a Friday order for Monday delivery.”
Since equipment suppliers’ revenues are falling at an alarming rate, their R&D budgets are being reduced, fewer new designs are being started and product lifecycles are being drawn out, it’s no wonder that these trends favor incumbent vendors. So a startup company better have some compensating elements (read: really cool, breakthrough technology and plenty of business development contacts) to offset this bias in today’s market.
Market-savvy Series A investors are harder to come by, even though valuations have become more attractive. The traditional business model of raising just enough money in Series A to develop a prototype is coming under attack.
Increasingly, Series B investors are demanding evidence of design wins and revenue traction before funding. That means that Series A guys either have to get enough money around the table to fund the company further (hard to do, in this parsimonious VC environment) or simply pass and wait to feed on the carcasses of later-stage companies.
Now, if no one is willing to step up for Series A rounds, there won’t be any later-stage opportunities. Either way, it’s not a pretty picture.
VCs are under pressure from their investors. They are not immune to the jittery nerves of their limited partners. Increasingly, venture capitalists are being held accountable for their investments in ways that would have seemed unimaginable just a few years ago.
Not surprisingly, it’s becoming harder to justify investments in areas such as communications, especially when the macro environment is roiling with problems (remember Global Crossing and WorldCom?).
The Optimist’s View
It is easy to follow the pack, but on the other hand sometimes a contrarian perspective can be beneficial. After all, above-average returns can be earned by betting on the dark horses. The more I pondered this market, the more convinced I became that 2003 may actually be a unique time in history for startup communications semiconductor companies. Here’s why:
* It’s still a big market. Despite much consolidation and down-sizing, this is still an $18 billion market, according to a September report by Credit Suisse First Boston. While some segments are floundering, many others are flourishing, such as wireless LANs, gigabit Ethernet, and MEMS (micro-electro-mechanical systems). For example, Piper Jaffray in October estimated that gigabit Ethernet would grow 56% in 2003, and the wireless LAN market would grow more than 70% to $577 million this year. Someone’s got to earn that revenue. Why shouldn’t it be startups?
* There’s an opening. Public companies are being forced to cut back on their R&D and focus on their core competencies. Broadcom recently announced it would cut back its research and development spending by 25% and abandon all projects that could not produce short-term (within 12 months) revenue. Retrenching by the big guys opens up more room for startups to fill the innovation gap.
* Rent is cheap and fab capacity is available. Today, in Silicon Valley alone, there is more than 100 million square feet of available manufacturing capacity. Worldwide, as of this writing, fab capacity is running at approximately 65%-in other words, there is literally more than enough room for innovation. And the combined Darwinian forces of economic recession, reduced VC investment and a firmly-shut IPO window have reduced the competition’s numbers. That’s yet another attractive sign for newbies in this space.
* Great talent abounds around the globe. It’s a truism that communications is global, especially with regard to engineering talent. There’s a tremendous amount of innovation coming out of Israel, for example. In fact, a recent Partech investment in Redux, which provides an innovative chip for last-mile access, supports this trend. While the company’s key target markets are Asia and North America, it will retain its engineering base in Tel Aviv. And these days, thanks to a hard-hit California recession, it’s even affordable to find talented and reasonably priced engineers in Silicon Valley.
* Innovation hasn’t taken the year off. Established silicon suppliers are still doing many of the obvious things of collapsing existing functions into fewer components. These same trends are also enabling a rich set of new functionality and algorithms that may radically change the basic building blocks of communications systems. For example, while everyone has been talking about system-on-a-chip (SoC) for years, we’re now actually seeing products shipping with this technology. And, as mentioned above, analog signal processing is still hard to do. The subsets of the market that value these skills (wireless, copper broadband, etc.) are still growing rapidly.
* Material and packaging advances also continue to create new opportunities. We are seeing new applications for MEMs technology creating value around design and manufacturing skills. That’s why in mid-2002, Partech invested in MEMGen, a company that has a breakthrough manufacturing process for MEMS, called eFab, which should accelerate the adoption of MEMS technology in a number of areas.
* Security is top of mind. As customers demand more from their network infrastructure, system equipment providers like Partech-backed Ingrian Networks have tapped into this trend. Ingrian provides ultra-secure encryption services at the Web server tier and/or within the application infrastructure level. It and others are riding a wave of new silicon that is optimized to provide services at Layer 4 and above.
Traditional CPUs cannot keep up with line data rates at 1G and above. NPUs, while good at some functions, also do not meet the specialized needs of these security devices. Open up these boxes and you will see a number of new chips that help to protect the volumes of packets flowing in and out of an enterprise.
By now I’ve revealed my true colors: I’m a schizophrenic. Most days, I think I am nuts and want to be a lemming. But there is still just enough contrarian in me to continue the search for true breakthroughs, while the optimist is looking for another paperclip invention. Maybe that just makes me a compassionate conservative.
Tim Wilson is a General Partner at Partech International, which he joined in 2001. Previously, he spent more than 20 years in the communications industry, most recently as chief marketing officer at Digital Island (acquired by Cable & Wireless). Prior to Digital Island, Wilson was general manager of Lucent Technologies’ $2 billion Business Communications Systems unit (now Avaya). He sits on the board of Teknovus Inc.