Interview by Declan Denehan, Managing Director at BNY Mellon

Arvind Sodhani is a legend in the corporate venture capital world.  As Executive Vice President of Intel Corp. and President of Intel Capital, Arvind leads a team that typically invests between $300-$400 million annually and has deployed over $11 billion in capital since 1991.  He has received many accolades, including headlining the Global Corporate Venturing Power List for the last two years.

At the recent Corporate Venture Summit in Huntington Beach, California hosted by Intel Capital, we had the opportunity to host a Fireside Chat with Arvind. The chat was conducted by Declan Denehan, Managing Director at BNY Mellon.  In the conversation, Sodhani shared many of the lessons gained from building one of the largest and most respected CVC units. Here are just a few of the highlights of conversation, in which he provided such valuable insights as:

1] To build a sustainable CVC, your investment thesis must be designed to move the needle for your company. The thesis must be defined and must extend beyond direct investment returns, and your metrics should capture all value delivered.

2] Success as a CVC arm depends on building the right team. Typically that means bringing in investment professionals as well as individuals who effectively work internally. Teaming the two groups can reap the highest business unit engagement.

3] Ensure you’ve got commitments to stay through the inevitable cycles and have enough runway to realize strategic value [minimum of 4-6 years]

To learn more about what works in Intel Capital, read the  edited transcript of the discussion with Arvind and Declan click here (link to be inserted). It was an hour-long discussion, but well worth your time.

The following is an edited transcript of the November 6th 2014 discussion between Arvind Sodhani, Executive Vice President at Intel Capital, and Declan Denehan, Managing Director at BNY Mellon.

Declan: Intel Capital is one of the largest, most active and respected corporate venture capital arms in the world, but I’m sure it has been a journey. Could you share with us how, why and when the ventures effort started?

Arvind: In the ‘80’s, when Andy Grove was CEO, the PC and the semiconductor industry were developing. We were losing some folks from Intel and others were starting up companies that were very relevant to us in terms of moving the needle forward for our processors and memory.

It turned out that we could not afford to do all the R&D or development in-house. I was Treasurer at that time and went to Andy Grove and said, “Look, why don’t we take small equity positions in these companies? That way we will have a relationship with them, and we can stay current on the technology that is developing.”

3dfx was the first one we invested in. I think I did the next half a dozen at random, in no organized fashion. The companies came to us. In most cases, in fact, they approached us. That’s how Intel Capital got started.

Later in the early ‘90s, it got more organized, and Les Vadász took over. He did a great job taking it from a random to a much more organized approach. We renamed it from Corporate Business Development to Intel Capital. Les ran it for 10-15 years until 2002. I was also involved at that time as Treasurer because we did a lot of the execution of deals and portfolio work. I took over in 2005 when I moved from Intel’s Treasury Group to Intel Capital.

Even though we had changed our name to Intel Capital, we were still largely operating as a business development effort. In 2005, it moved into being a real venture organization. There is a huge difference between a business development effort and a true venture effort.

Declan: Maybe you could talk a little bit how the investment thesis evolved.

Arvind: Over the years, we went through a number of iterations of “why are we investing here?” Every CEO will ask you that, and it’s not always obvious. In other words, sometimes, a company will enter a commercial relationship or buy a processor or some other component from us. But that doesn’t move the needle for a company like Intel. It hasn’t for the last 15 plus years. Intel’s business units are billion-dollar businesses. A small startup company doing $1 million dollars in revenue is not going to move the needle.

What really made the difference and was valuable to us was helping a company be disruptive, or innovate in a whole new segment, help create a new industry that in turn will need a lot of our products. Well, that turned out to be a very powerful thesis.

That worked really well for us when we got into China in the mid-nineties. The market had barely opened up. There were no Chinese companies working on any applications that would make any sense for Chinese PC users.   So we went in and invested in a bunch of companies, portals, insurance companies, etc., all kinds of applications that could use our products. That’s been true systematically in a lot of different areas over the years.

The Internet is another example of how, in the ‘90s, we invested in a lot of companies that created all these applications that we see today. That was the thesis that got us here. When we invest in a company, we’re not looking for a single relationship that will move the needle for us, or that they will buy parts or components from us; we want to help jumpstart complete new segments or industries.

A more recent example is our drone investment. We scratched our heads for quite a while asking, “Where does this fit into our thesis, why should we invest in it?” We finally figured out that a drone would take a lot of video. In fact it takes so much video that the drone runs out of storage before it runs out of battery power, which is unusual. Usually, things run out of battery power before they run out of anything else.

For example, insurance companies would use them to take video of a flooded area, fly over it with a drone, figure out what houses are underwater and write the check right away, because the property is a write-off. Owners get their check a lot faster than waiting until after the flood has receded and adjusters are able to go out.

Another example is agriculture. Among the services farmers are being offered are drones that have the ability to fly over large farms and, through video and running analytics in the cloud, can help farmers improve in their seeding, their watering and the way they run their farms.

These are brand new applications. They generate tons of data that goes into the cloud, into the data center, and then a lot of analytics get done on that data. Doing analytics on video images is quite complicated and compute intensive.   If we can help the drone industry take off, it’s a big win for us because it’s a new industry, a new application, a new utilization of our storage, a new utilization of the network and of our servers in the data centers. That is the kind of thinking behind our thesis.

Declan: Fascinating. What specific segments are you most active in right now?

Arvind: We are investing in a technology ecosystem continuum that ranges from Internet of Things applications like smart devices, wearables, laptops and desktops — essentially appliances on one end of the spectrum. Then data centers, cloud, cloud infrastructure and data analytics on the other end of the spectrum. And everything in between.

This includes silicon, services, computing, consumer Internet, media, education, etc. It’s very broad and are all areas that are relevant to Intel.

Declan: I understand you currently invest between $300 million and $400 million a year.   Obviously, if you’re covering such a broad landscape, you’ve got to have a very efficient machine identifying opportunities and building deal flow. Could you tell us a little bit about how your organization operates?

Arvind: We typically invest between $300 million-$400 million. Some years we invest a little bit more and that, of course, does not include large deals like the one we did this year with Cloudera.   We’re organized by sector and by geographies.

As you can imagine today, data center, cloud infrastructure, networks and data analytics are among the focus areas. Wearables and IoT are other areas we started last year and we are pretty excited about them. Security and silicon, those are consistent investments that we do.

We could easily invest $500 million in a year or more, since we have over 150 people in our organization, and about half are investment directors.

On deal flow, we’ve gotten to a place where we see virtually every deal. We get spectacular deal flow because of our brand and because we are engaged in everything and virtually every company out there.

Also, our value proposition is quite unique. For example, with our Global Summit, it’s helping our portfolio companies find their customers. In two days, we had over 2,800 meetings between 200-plus portfolio companies and over 300 industry executives and VCs.

Another benefit of the Global Summit is that our portfolio companies get the kind of publicity they will never see on their own. They’ll get carried in usually 200-plus articles. It’s all part of our value proposition. That is very powerful.

Declan: What interaction do you have with the core business units when going through decisions about how and where you’re going to allocate your capital? Also, how much input do the core business units have into any investment decisions you make?

Arvind: Our process is based on the priorities and objectives of the CEO. We therefore also honor the priorities and objectives of the business units. We work very closely with the business units, and our sector organization mirrors the Intel organization. For example, our data center sector is mirrored with the data center group at Intel.

We have another category of deals called “eyes and ears” where the business units are not necessarily as engaged, because we are looking way out in the future to see if there is an opportunity to do something that might benefit Intel and intersect with Intel business units down the road. Eyes and ears are probably 15-20% of our total investments.

Declan: A lot of corporations house their venture units in their traditional corporate development area.   Talk a little bit about decisions you made moving out of corporate development and separating Intel Capital.

Arvind: Separate doesn’t mean that we’re not Intel; we’re very much an integral part of Intel. There is a very disciplined business process in Intel Capital. We want to make sure every deal fits the requirement that it has both a strategic objective and a financial outcome.

We do not want to compromise either of those two things. It must be relevant to us: We’re not going to invest in real estate, food processing or any other company that happens to be very attractive financially but will have absolutely no relevance to us. We do not want to compromise that discipline just for a financial return. In the business development role, the financial return aspect overwhelms all others.   What we have discovered over the years is that companies that do not thrive financially will have limited strategic impact because they’re not going to survive. That’s the huge difference between business development and corporate venture. Our focus is to sustain this effort over a very long period of time.

We are investing a lot of money. $300 to $400 million dollars a year is a lot of capital to deploy. And we are very fortunate: Over the last six years, we have been cash flow positive.

Declan: Something around 100 deals a year?

Arvind: Closer to 50 to 60 deals a year, plus a lot of follow-ons. We also have a very organized business development effort, with twenty people dedicated in one form or another to helping our portfolio companies, introducing them to customers in an organized effort. That’s probably bigger than most CVC organizations.

Declan: Could you elaborate a little bit on the actual decision making process for investing? How much autonomy is given?

Arvind: We have a three-step process that starts with a deal concept meeting, where you have to bring your deal in. It’s made up of myself, legal, treasury and the sector MD [managing director]. We look at the concept, understand why it is relevant to us, what’s the strategic rationale? All the usual questions, and the IM [investment manager] and MD will go back and work the deal further.

The second step is a meeting between myself, the committee and the [startup’s] CEO. We always want to meet the management.

Obviously the IM and the MD doing the deal will be working with the company in a very intensive fashion. Then, finally, when the deal is properly cooked, developed, validated and structured, we’ll come back for another meeting where we would approve it. I say most likely approve it because it has already gone through a number of “looks” and a technical review. It’s a very organized process.

I dare say that three looks are probably at the bottom end of the range of the times that I want to look at the company. We want to have several checkpoints to understand if things have evolved or if we’re not seeing business unit commitment.

The other thing we do is we ask business units to grade the deal. If a deal doesn’t get a meaningful grade, we’re going to have reservations, because the business unit is not there. Then we have to ask ourselves, is this relevant?

Declan: What’s the overall staff size? And maybe tell us little bit about the makeup, the tenure and how much mobility there is between Intel Capital and Intel corporate?

Arvind: We have more than 60 Investment Managers and Managing Directors who are interacting directly with the deal flow. We have about 20 in portfolio business development; approximately 20 lawyers who help put the deals together, and about 12 portfolio managers. We also have treasury and finance folks who work with us. The total head count is around 120.

Declan: What financial metrics and what strategic metrics are you using?

Arvind: We look at sector IRR, geographical IRR, cash flow; we look at our portfolio every which way. Our interaction with the business unit tends to be more qualitative. We sit down with them and they grade us on each deal. Later on, the grade can be lower or higher because it didn’t turn out the way we thought it would, or the business unit has changed its mind and decided to take a different tack.

In addition to grading, our sector MDs sit in the business unit staff meetings, which is the highest level meeting of that group and part of the ongoing dialogue.

Declan: There’s been a lot of press recently about valuation levels, the role of corporate venture and if CVCs are paying too much. What are your thoughts on the overall industry right now?

Arvind: First of all, we began in the 90s, which was when a lot of corporate VC’s got started. Then came the 2000 downturn, and they abandoned it. We had 20+ corporate venture groups bailing out three years later. We don’t want to see that, because it’s not good for the industry.

It’s important that corporate VC’s get their boards and their CEO’s mentally ready to accept the reality that unless you’re incredibly lucky, the losses will come before the gains will come, and the losses will come before you’ll be able to prove the strategic value.

Corporate VC’s are doing this to add strategic value to their businesses. Unless you’re incredibly lucky and you hit a home run right off the bat, which is not going to happen for most players, the reality is that the losses will come first.

The strategic outcome will not be that fast or that evident. In order to sustain, you have to prepare the organization to recognize this is a long-term effort. It requires hiring the right people. It requires understanding that you’re competing with other VC’s who have done this all their lives, who are motivated by financial return and who probably know this industry better than you do.

I would implore every corporate VC to make sure the decision makers understand that in order to sustain and do this long term, the losses will probably come earlier, and that’s not the end of the world. You will get the gains, but not necessarily first, unless you’re incredibly lucky.

Declan: In terms of companies that are contemplating starting a corporate venture function, when meeting with the CEO or the board, how much runway should they seek?

Arvind: By the fourth year, you will probably begin to see successes. Four to six years would be the time frame. The other thing that’s critically important is to find the right people. It takes a while, and finding the right people that can make successful investments is difficult, as you are competing with a lot of other companies.

Declan: We are coming up on time. Perhaps a few questions from the audience

Participant: You’re very sophisticated about measurement and metrics. How do you measure business development?

Arvind: One of the ways we measure business development is through business development events like the Intel Capital Global Summit. It’s a very important event for us, for our portfolio companies and Intel’s customers. We host the summit with the help of sponsors, and charge a small fee to attend. Over the years, the Summit has grown into a valuable networking event where the ecosystem comes together to do business. We help organize 2800 targeted meetings.

We also measure the results of the Intel Capital Technology Days (“ITDs”). Our team typically arranges about 75 ITDs in a year. We look through the feedback we get from the ITDs and from the host companies we do ITDs with. So a combination of numerical as well as qualitative metrics.

Participant: Regarding relationships with VCs, there’s been a big change in the dynamics in the venture world over the last 5 years or so. So I’m also curious, with the proliferation of microcap funds, angels, etc., your position on investing alongside that group of players rather than the institutional VCs.

Arvind: We have developed relationships over the years with the angels, the accelerators etc. and those can be a good source of deal flow, but those would be very early stage, so it is more difficult for them to have an impact on your business. One of the things you’ll have to figure out is in what stage you want to play. We are stage agnostic, with the ability to invest in anything from seed all the way through post IPO. We prefer not to do too much hand holding, which is usually a requirement at the seed stage of the company. We prefer the A through D rounds. All our value propositions are kicking in when the company is $30, $40, $50 million in revenue and we can help with the technology road maps and help the company to scale.

With respect to interaction with traditional financial VCs, we compete with them, we win against them, we lose to them. We work very closely with a number of VCs so we see a range of expertise and behaviors.

Declan: Tell us a little about some of the real challenges of going into places that don’t have developed infrastructure or of essentially leading the development of the venture capital industry in new markets.

Arvind: We’re incredibly lucky in this country to have this highly developed ecosystem. We went to Turkey a few years ago, we went to China in 1998 and we went to India in the same time frame. What we found was there was no developed ecosystem at all. And by developing the system, I don’t only mean the fact that there is no legal framework. The lawyers don’t know how to write a venture investing contract. In many places you can’t get a preferred instrument because that’s not legally possible. So those are infrastructural problems, environmental problems.

But there is a bigger challenge, and that is the role model of taking the risk to start a company. For an engineer who has had some experience working at a technical job and wants to establish a company, the role models aren’t there, so getting them to start companies is unusual. Then even when the founders do start a company, finding people who will join and become employees is even harder, because they are not going to leave safer jobs. Ecosystem challenges are huge in countries that don’t have a history or an ecosystem in place. Developing that can be very tough, but it can be developed. China is a classic example of how it’s developed fantastically. It’s amazing how many startup companies we see in China and the ability they have to attract talent.

Participant: The time frame for investing rounds seems to be running shorter and shorter. Can you discuss how long it typically takes you to work a deal?

Arvind: The time it takes for us to get a deal done stretches from days to months. Sometimes we can get a deal completely approved, ready to sign, in 2 to 3 weeks. That’s probably driven by the ability of the portfolio company to move. We can move fast at times: bring 12 lawyers and treasury guys. We can bring an army to the table to get a deal done. The problem tends to be on the other side of the aisle.

Typically I would say that it’s probably a 2 to 3 month process from when our IM makes contact, or gets to know a company, to when we would have the approval. It would also be subject to other investors: Another thing we bring to the table is our syndicate program. We have now approximately 40 syndicate partners. We show our deals to our syndicate partners. And some of them consistently co-invest with us so we can help the company raise their round faster. But typically the investment process takes between 2 to 3 months.

Participant: I was just wondering if you could share in a little bit more detail the interaction your MDs have with the business units, and how you transfer that to the founders.

Arvind: We have programs like the Intel Capital Technology Days and Intel Capital Global Summit to help our portfolio companies meet their customers. We also bring our portfolio companies in to meet our business units on a regular cadence and see if there is a possibility of having a business relationship. That interaction can vary between business unit and business unit.

Some business units are going to be more engaged than other business units. We also have our own IT organization that can be very helpful to companies figuring out where they’re going to be able to get traction. There are a lot of different ways in which we work with our portfolio companies to help them.

We are all about value add. Our brand gets us through the door; once we are in through the door, our calling card is our previous history, the people that are going to sit on the portfolio company board and how they are going to help you — our business units, our IT organization, the technology road map, our ability to introduce you to relevant decision makers from global 2000 companies worldwide. That’s how we get into the deals. That’s all part of the process, and we really do deliver.

One of the things from our Intel Capital Global Summit that just finished yesterday is that we get feedback from everybody that participated, and we rigorously go through that and say, “Okay. How can we improve that?” All these meetings took place, did people get value? When we do the ITDs, (Intel Capital Technology Days) we follow up and we see what business did the portfolio company develop participating in in those ITDs? We host these in various parts of the world, and 20-30 portfolio company executives travel there and so on. For the host, they are bringing 20-30 of their decision makers for a whole day. So, all metrics have to be in place. It can be expensive, but extremely valuable.

Declan: Arvind, we’re just about out of time, but I do know that we have a couple of folks in the room today that are helping their companies try and decide whether or not to actually start a CVC unit or not. Any last words of advice?

Austin: A corporate VC organization can be a very valuable organization or part of your company. It can be a catalyst.

However, it cannot be done in a short period of time. It needs time, and you need to survive over a 4-6 year period, because if you get shut down you won’t deliver any value and you’ve not created any returns. In order to prove that this is a viable effort, you need runway. Without that runway it just gets hard. It also takes time to find the right people. All those things have to be done right, because it’s not easy and you’re going to be competing with established VCs, and other corporate VCs and you’ll be competing with us as well. So, join our syndicate program, contact us about attending the Intel Capital Global Summit and let’s work together to build great companies.

Declan: Thank you so much, Arvind, for your hospitality and for sharing your insights.