Startups get big help from Big Blue

Armonk, N.Y.-based International Business Machines (NYSE: IBM) set up the IBM Venture Capital Group eight years ago in San Mateo, Calif., but don't be fooled by the name.

The group isn't handing out cash for equity, but it does have some valuable services to offer startups, including access to IBM's Rolodex, if the technology is a fit for Big Blue or, more specifically, for the needs of its big name customers.

Technologies that can help make green data centers, as well as hardware and software for other energy efficient applications, are of particular interest for IBM, as well the utilities, computer storage, and other companies that it works with.

IBM acts as a mentor, working with venture capital companies and their portfolio firms to get the technologies ready for market. While it's not a direct financial relationship, there is a financial incentive for IBM, which hopes that these startups will become valuable suppliers for IBM's customers, who will want to integrate the new technology into IBM systems.

IBM's Venture Capital Group has had over 1,400 venture-backed companies come through its operation over the last few years, and it works with over 120 venture capital firms around the world.

Drew Clark is the co-founder and director of strategy at IBM Venture Capital Group, and he spoke with the Cleantech Group about what IBM can offer startups, what kinds of technology the company is looking for, and where then next big cleantech boom is likely to emerge.

How does the Venture Capital Group work? Are you a consultant for startups and venture capitalists?

We're part of IBM's corporate staff, so we don't report to the CFO. We're not a financially driven or financially measured organization, but rather we're strategic.

Given that orientation, you might conclude that we're really mostly about driving top line growth to IBM, in terms of helping push strategic growth initiatives, rather than, for example, an equity arm, or something more conventional in terms of aVC limited fund. And that's certainly true.

We don't have our own fund. We decided back in 2000 that there was plenty of capital in the market, and indeed today there is plenty of capital. What there seemed to be a need for was a company like IBM to help in the rest of the life-cycle of a startup.

Beyond the initial capital infusion, who's going to help that company get pulled into enterprise solutions? Who's going to help that company mature and grow into a great ISV (independent software vendor) or a participant in the IT eco-system? And that kind of role falls to us.

Our expertise really is not around managing minority ownerships in small companies, our expertise is really around markets and customers and deep technology, that sort of thing.

It's really a partnership with the VCs, rather than kind of one of them or competing with them.

So I'm not sure you'd call us a consultant, I think we're more of a venture technologist, maybe, rather than a capitalist. We're out there really advising VCs, but not in the sense of a consultant. We're not getting paid for advice. It's very much what we call a give to get program, where we believe in giving our venture partners a great deal of value, in terms of helping their companies achieve great things, and at the same time we're getting back early access to some of the best new partners IBM could ask for.

Over 30 percent of our revenues come from our partner channels, so it's really critical that we have this pipeline full of great new partners. And this is one way of ensuring that we do. By working closely with VCs and their invested companies, we can determine our own fate, if you will, in terms of getting potentially great new partners line up and working with them and helping them succeed.

Because when they succeed, we succeed.

What companies have you worked with?

In cleantech, we think this maybe the most wide-open area for partnering with startups. In the area of, for example, smart grids and energy efficiency.

In the case of smart grids, for example, we're working with several significant startup players, including eMeter, Silver Spring Networks. SynapSense is another one.

These are companies that provide something that either fills a gap in our portfolio or helps us reach a market that we wouldn't have otherwise ordinarily been able to do, or at least as quickly as we can with this partnership.

So they're bringing us big-time value in terms of helping IBM drive revenue. What we give them in return, of course, is the same thing. We're giving them access to a revenue situation that they might not otherwise be able to get.

We can pull an automated metering company like eMeter into deals that, they may be able to get, but it may take them years to build the credibility and to essentially work the sales cycle with a large utility. And utilities can be tough customers, and are not exactly your prolific-at-IT-spending companies. They're very conservative and very careful about investments they make.

IBM, being a trusted partner of these utilities, is able to help provide a bridge, let's say, between these startups and the utility space.

In fact, we have an industry vertical at IBM called Energy and Utility. The utilities are our customers. PG&E and Edison, all those guys are our day to day customers. So we have earned their trust over the years, so it's much easier for us then to go in to a utility and say, 'Hey, we just partnered with Silver Spring, and boy have they got a great solution for moving information over an IP network. We really would like to talk to talk to you about adding this into your business process in terms of the way you access demand-response information.'

And all of a sudden we create business. We create business for IBM, we create business for the startup.

In energy efficiency, green data centers, obviously IBM's name is going with operating and building data centers. Our customers there anxious — anxious is the word — for IBM to help them get a better handle on the energy footprint in these data centers.

How much money is being spent on cooling, on energy to drive cooling systems, for example, and could we more efficiently cool these data centers?

Well, where do you start? Who's using the most energy? You have no idea, other than just kind of randomly pulling plugs and measuring watts per hour. Who really knows where the heat is and where cooling needs to be, and so forth, in an average data center.

IBM didn't make a device to monitor and discover that, but it turned out that startup SynapSense did. And SynapSense is up in Folsom (California). So we partnered with them to essentially build a solution that couples their wireless mesh network sensors with software that IBM wrote that can take the data and essentially analyze it and provide hot spot analysis.

So, again, there's a complementary kind of arrangement, where we can essentially surround this technology with IBM's process and business know-how and gets this tiny startup in front of some pretty major customers.

How broad is IBM's interest in clean technology?

Well, obviously, just about any aspect of a data center is going to be of interest to us, because it's interesting to our customers. You only have to look at what's a typical IBM customer and what do they care about and then you could probably get a pretty good read on what's important to us.

Everything from the obvious things, by obvious I mean the servers and storage devices and so forth that are in there, obviously we care a lot about that. But also power consumption, power usage, power management would all be important to us. Efficiency of management — in other words maybe a remote management is indicated versus on-site. Maybe we could do some automation to automate the deployment and management of these systems.

That brings in all kinds of potential startups in terms of energy management kinds of things. It brings in battery manufacturers, like A123 and others, who are making batteries that would power UPSs and other sources of moving intermittencies.

We're interested in companies doing DC power, for example, to help cut down on costs there. We're interested in companies doing air filtering and other environmental kinds of companies. We're interested in companies doing new kinds of cooling on the racks of these devices.

In fact, IBM has innovated there. We've got our own innovation called iDataPlex, which is a X86 configuration, but without raised floor or air conditioning. There's no HVAC needed. We have heat exchanges, like from a diesel engine, bolted on the back of these racks and it pulls the heat off and cools it through circulated water. So we're back to water-cooling again (see IBM cools down its supercomputers).

But basically that's an example of the kind of innovation we're looking for. Anything to help reduce that energy footprint, and-or help manage the energy consumed by the data center.

You talk about scope, it doesn't have to be a data center even, it could be a smart building. There's no reason that it has to be just an IT operation. Many of these same principles hold for other aspects of a customer's site. So we're looking more broadly than just the data center.

Our customers don't look at IBM as a point solution provider, they want us to manage the whole thing. So we're talking heterogeneous management — our devices, their devices. And we're also talking about smart software, which I haven't mentioned, but the idea is help get a better handle on why the servers are there in the first place on computing tasks.

If we can be smarter about how often we need to spool up servers, for example, when we bring them online, then we can cover a lot more ground, too. The idea is developing software, and IBM has a pretty good handle on some of this through our Tivoli brand. We have a pretty good idea how to manage workloads based on picking the most efficient cluster of resources to have spooled up at any given time. And I say cluster meaning a server, storage, other pieces that are needed to get a piece of work done.

We've created some concepts, and I can't talk too much more about it because it's not been announced yet. But in the area of cloud computing you could look at the inside of a data center here, and you could say that there's a lot of concepts involving highly scalable architectures that could be applied here to essentially create a much more efficient data center operation through some clever software that automatically provisions and automatically tailors workload based on an energy efficiency model of how it should run.

Lot more to come there.