In the wake of venture capital firms investing more in 2010 than they had in the previous two years, and some significant funding for hardware startups in the last 12 months or so (Tilera, Calxeda, SeaMicro, various fiber companies), I wondered if underneath the social media frenzy there’s a hardware renaissance happening. I’m not suggesting a return to the bubble years when telecom and networking startups managed to go through billions before flaming out, but a more measured investment tied to a few trends — namely the need for more power efficiency computing, connectivity everywhere and better, faster networks.
Today, two stories started me pondering this question: A survey of upcoming trends in Data Center Knowledge and a story in Light Reading about networking conglomerate NTT willing to evaluate startup’s products despite a dearth of startups in the space. From Data Center Knowledge:
Grove says some of the biggest advances will come from startups whose products are still months and even years from production. “The most significant development in the data center industry in 2010 was the influx of several young companies offering up a vision of energy-efficient semi-conductors,” she said. “This will turn the industry on its head when their claims of 50% to 80% increases in power efficiency are substantiated in pilot studies – but we won’t see the results of their pilots yet in 2011.”
The article also tackles the need for advances in networking, including the Open Flow project at Stanford, which is pretty cool. You can learn more about that from our Structure 2010 presentation here. Other data center experts explain that open source networking and more intelligence at the networking layer inside data centers is becoming crucial. Outside of data centers, communication between building and people is also seeing a need for innovation. NTT America Inc.’s Michael Wheeler, is looking for new vendors to work with who want to build something interesting rather than push an existing platform, but such startups are harder and harder to find. He told Light Reading:
“They aren’t very common any more -– [IP router vendor] Procket may be the last one I remember, then Cisco bought them,” recalls Wheeler. “We were talking to them early on, and they had some pretty interesting things to talk about.”
Unfortunately for NTT and those looking for new data center technologies, venture firms have been mostly retreating from the capital-intensive silicon and networking spaces. In the most recent 2010 MoneyTree report, investment in networking equipment hit a low of $665.8 million, which was the least amount invested since 1996. In 2009 and 2010 investment in chip firms was below 1 billion, which hasn’t happened since 1999, but investment was up in semiconductor startups from 2009.
So if this hardware renaissance is going to occur it’s likely to come from a relatively small cadre of venture-backed deals currently building products, or perhaps we’ll see a return of VC into the sector. It may be wishful thinking, but hardware is the underpinning of our broadband and cloud computing revolution, so it stands to reason that innovation will have to happen there to keep technology moving forward.
Related content from GigaOM Pro (subscription req’d):
- Supercomputers and the Search for the Exascale Grail
- Pushing Processors Past Moore’s Law
- Nvidia Arms Itself: Who Has Most to Lose