But companies go through the motions because a: that’s what they’ve always done and b: they want to stay legal. That model may not be dead yet, but its days are numbered, at least according to some observers. What’s contributing to its decline is the increasing use of public cloud infrastructure like Amazon Web Services which together with Software-as-a-Service offerings pretty much obliterates the need for most on-site server and software upgrades.

Trying-before-buying model gains steam

And, more enterprise customers — not just startups anymore — have glommed onto the try-before-you-buy model that lets them download software, use it as they see fit, and when the time comes to deploy across company or to get support, all they have to do is pick up the phone.

Sunil Dhaliwal, who founded Amplify Partners, a VC firm that backs infrastructure startups, sees a massive transition underway in how companies buy IT. “This threatens the big infrastructure and enterprise IT companies to their core. It’s even bigger than changes in the technology itself,” he told me recently. (GigaOM’s Stacey Higginbotham wrote about Amplify here.)

The real deal, he said, is that enterprise customers are “very tired of getting the short end of the stick in their sales experience.”

Enterprise customers arise

Paul Santinelli, partner at North Bridge Venture Partners, agreed. Many companies just don’t have to install software for many core functions any more. Instead they try out things like Box for document sharing and storage or Okta for identity management. “You download it, try it and then buy it without ever meeting the sales guy,” Santinelli said.

And, there’s a generational shift among IT buyers. Younger people are happy to download and try things and let business units make their own choices.

That’s not the kind of sale companies like EMC, Oracle, IBM and even VMware — all with big well-paid sales organizations — want to hear about.

Granted the transition will take time and the legacy vendors are not stupid — they see it happening but it’s hard for them to react fast. “This is not about them not getting it. EMC gets it. The problem is the classic Clayton Christensen Innovator’s Dilemma stuff — they have to make their quarters and you don’t do that by cutting your direct sales people,” he said. But companies like EMC, which Dhaliwal called the “best-ever factory for turning BC football players into highly-compensated sales guys,” will have to adapt eventually.

Newer generations of IT buyers won’t want to relinquish the freedom of downloading specialized software from young, nimble vendors instead of locking into one or two huge vendors for a wide array of applications.

Exception that makes the rule

One thing that may mitigate against faster change is the current regulatory and compliance climate. Companies in the financial services industry, for example, must show that all their technology is up to date and fully supported. That explains why companies still pony up for Red Hat Enterprise Linux as opposed to CentOS, even though many see no substantive differences between the two.

And vendors play right into that fear of being out of compliance. A VP with a large New York-based bank told me vendors like IBM and Oracle “give you full access to all their software to use or not but then come in with an audit or the threat of an audit to make sure you pay for every bit of it and try to lock you into an enterprise license agreement,” he said, indicating he is not at all pleased with that situation.

But, regulations aside, change will come, Santinelli and Dhaliwal agreed. Enterprise buyers are so fed up with the old model that they’re willing to take risks.

“One reason that OpenStack has gotten so much traction for something that’s not cooked is because it’s an alternative to [VMware] vCloud Director and companies don’t want to see vCloud as the new Microsoft CAL-style license lock-in,” he said.

Microsoft is famous for using its client access licenses, which are often initially cheap or free for new products, to get companies using those products and then jack up the license prices. That’s just the sort of enterprise sales technique that companies resist.

Financial services companies are locked into Oracle  – and its sales people — for now, Santinelli said. The top IT guy “will keep buying Oracle from a rep in a suit but many of the people who work for that guy are already running applications like Hadoop or Couchbase on a server under their desk or in a VM in the public cloud. Those people will likely replace that IT guy in 5 or 7 years. Then they’ll be buying software, compute and storage just like you buy electricity — on a monthly usage-based rate. They won’t need to own the power plant.”

Feature photo courtesy of Shutterstock user Peshkova

Related research and analysis from GigaOM Pro:
Subscriber content. Sign up for a free trial.

• A near-term outlook for big data
• Infrastructure Winners and Losers of 2009
• Report: The Future of Data Center Storage

As for the fund, his goal is to invest between $50,000 and $1.5 million as part of a syndicate of other venture firms in seed or Series A deals. He plans to do about eight to 10 deals a year, which is a lot for such a small shop, but Dhaliwal plans to bring in a team of experts from the industry to help mentor startups. He declined to give names, but he says they are operational and architecture experts, who will take equity stakes from the firm — not from the individual company — in exchange for their help.

It’s an interesting model for the VC world, where there is an evident split between the giant, big-name funds and smaller, niche players. The NVCA recently released its fourth quarter 2012 fund-raising data that showed the top five venture capital funds accounted for 55 percent of total fundraising this quarter, noting that niche, smaller funds and giant funds made up the bulk of the new fund raises. Amplify is clearly on the niche end.

So far current Amplify Partners portfolio companies include AppNeta, Continuuity, Datadog, Fastly, Wibidata, as well as multiple companies still in stealth mode.

So what else is Dhaliwal planning to invest in? The big picture is in companies that are part of the upcoming shift in computing, networking and storage infrastructure thanks to scaled out architectures, the growth of data, and real-time infrastructure. Most of our readers are well aware of the massive shift occurring today in technology infrastructure. Between scale out computing (hello, Facebook and the Open Compute Project), disruptions in networking (shout out to Nicira!) and changes in how startups and enterprises source their computing (what’s up AWS and devops!), we’re seeing the kind of shift that hits computing once a generation.

Dhaliwal calls it Infrastructure 2.0 and he says that the trillion-dollar IT industry is under siege, which means the giants are spending big on what I think of as “save-me deals,” such as Cisco spending $1.2 billion for Meraki, or VMware shelling out $1.26 billion for Nicira.

“For the first time there are companies with trillions of dollars of market cap and they are absolutely under threat in a way they haven’t been before,” Dhaliwal said in an interview. “These are the types of changes that tell me a lot of those companies will not survive with a trillion dollars in market cap. A couple hundred billion of that amount will get transferred to challengers and startups.”

It’s not just the technical shift that Dhaliwal finds promising. He’s also attuned to the cultural shift — maybe call it the “devopsification” of IT — as the buyers of big IT change, and even the older buyers look for less of the multi-million sales pitch from a guy in a suit and more of the expertise and problem-solving that comes in part from the open-source world.

“There are all these technical changes and those merge with the cultural … and that’s pretty antithetical to how infrastructure 1.0 has run,” Dhaliwal said. “Very few of those companies have done anything to adapt to the change in who the early adopters are and how to sell to them.”

Dhaliwal says that when he started out investing in infrastructure in 1998 he used to visit the CIOs of big Wall Street banks, because if he understood their needs he was confident that 18 to 24 months later those would be the needs of most IT buyers. Now the leading indicator customers are the Facebooks and web guys. That doesn’t mean the Wall Street folks aren’t important. Dhaliwal went to them to ask if they saw the same infrastructure shifts brewing when he was thinking about raising his fund.

They not only agreed with him, they handed him checks. With that base of limited partners investing — as well as a few other sources of capital — Amplify and Dhaliwal were on their way.

Related research and analysis from GigaOM Pro:
Subscriber content. Sign up for a free trial.

• The fourth quarter of 2012 in cloud
• How direct-access solutions can speed up cloud adoption
• The role of converged infrastructure in the data center

Such services are generally referred to as cloud computing, and the game-changing potential of those services has venture firms sitting up and taking notice. Indeed, after spending the past few years pouring money into Facebook applications and me-too social networks, venture firms are starting to invest in infrastructure again, with both hardware and software plays tied to the cloud.

“Clearly there is a renewed interest and investment in infrastructure,” says Bernard Dallé, a partner with Index Ventures. “Twenty-four months ago it was all about the consumer Internet and still a lot of money is going after that, but it has been rebalanced. Now firms see the value that EqualLogic and virtualization has generated, and it’s time to invest.”

So far this year, companies providing cloud services or building services on top of the cloud have raked in more than $70 million. That’s nothing compared to the $14.9 billion that VCs invested during the same six-month period overall, but interest is picking up.

Just this week, 10gen raised a $1.5 million first round from Union Square Ventures to create a platform for programmers to build products on top of the cloud. Appirio, a company trying to help enterprise customers link their data among clouds offered by Google, Salesforce and Amazon raised $5.6 million from Sequoia. And earlier this month, EngineYard raised $15 million from New Enterprise Associates, Amazon and Benchmark Capital to build out a development platform for programs built using Ruby on Rails.

These companies join a growing ecosystem of startups trying to create utility and business models built on top of thousands of servers. If you peel back the fog surrounding cloud computing there are several layers of services. It all starts with a virtualized server running a hypervisor. Between the hypervisor and the operating system — such as Linux or Windows — sits a class of service providers, among them Elastra and Enomaly. They provide tools and services that help an IT manger build out, monitor and manage the virtual hardware inside the cloud.

On top of those are development platforms that can be tailored to a specific cloud, such as Amazon’s, or tied to a programming language, such as Ruby on Rails. Startups here include Bungee Labs, EngineYard and Coghead. Once developers have their programs built on the cloud, they need to monitor and tweak them using tools from the likes of RightScale and Hyperic. With the exception of Enomaly and Coghead, all of these startups have scored venture funding this year.

Coghead raised $8 million last year, and Reuven Cohen, co-founder and CEO of Enomaly, says he’s fielding about 40 calls a week from venture firms that want to invest in his profitable, boot-strapped company. “I’m not in any desperate need to raise funding to keep the business afloat, but we could grow substantially faster,” Cohen says. “We are open to it and from what I can see, there won’t be a better time for valuations.”

Valuations might be on the rise, but VCs are still focused on capital efficiency. For example Sunil Dhaliwal, a general partner with Battery Ventures, says he’s not interested in investing in any cloud provider that wants to build out thousands of servers as there are plenty of companies — among them Google, Amazon and Rackspace — doing that already. In fact, he’s approaching the entire cloud space with caution.

“We’re so early and there’s still plenty of money to be lost — and I underline lost here — and plenty to be made,” Dhaliwal says. “I think there is going to be a lot of trial and error. People are defining and building solutions without people knowing how they really want to consume them.”

He believes the big opportunities will lie with firms that can help corporate customers connect their existing IT networks to the cloud, as well as with those that provide computing as a service to small and medium businesses that don’t want to manage their own IT networks. One way or another, the move to computing delivered as a service is a huge change in the way businesses and even consumers will consume information technology. As far as venture dollars floating amongst the clouds, this is only the beginning.

This was originally published on BusinessWeek.com.

Related research and analysis from GigaOM Pro:
Subscriber content. Sign up for a free trial.

• Infrastructure Q2: Big data and PaaS gain more momentum
• Cloud computing 2013: how to navigate without a map
• Infrastructure Q4: Big data gets bigger and SaaS startups shine