March 10, 2009
In its continuing quest to spread GPU goodness, today NVIDIA announced a new program designed to give a boost to some of the most innovative GPU computing startups. Called the GPU Ventures Program, it's an initiative that combines direct NVIDIA investment with marketing, sales and distribution support. The new program is mostly targeted at companies developing CUDA-based applications and middleware for HPC and visual computing and is designed to fertilize the software ecosystem around NVIDIA's GPU computing business.
The main thrust of the program is to make it easier for undercapitalized companies to get off the ground in this tough economic climate. Because of the recession, right now venture capitalists (VCs) are scrutinizing their investment portfolios more carefully, and they want to make sure that any new companies they sink money into have enough capital to ride out the current downturn. In general, investments by corporations like NVIDIA would tend to tail off when the economy is booming. That's because institutional investors are willing to take on more risk during economic expansions, so startups have a bigger window of opportunity to make a go of it. All that has changed.
"Especially in these kinds of economic times, where things are a bit depressed, we feel that many of the traditional VCs are pulling back." explains Jeff Herbst, NVIDIA's VP of Business Development. "So we need to give them a good reason to stay in the game."
By identifying the rising stars of GPU computing and putting up some of its own money and marketing/sales muscle, NVIDIA is giving these startup its seal of approval. Not only will NVIDIA directly invest in these companies, but by featuring these newcomers on NVIDIA's Web site and including them in industry tradeshows, the startups will get a lot more visibility. All of this is designed to attract the attention of VCs, investment bankers, and ecosystem partners -- not to mention customers.
Along the same lines, this fall NVIDIA is going to host its second Emerging Companies Summit, a venue for GPU computing startups to pitch their wares to investors and industry analysts. (Last year's event became the impetus to form the GPU Ventures Program.) Although not every company at the event will be a recipient of direct investment from NVIDIA, the participants are invited because of their contributions to the HPC and visual computing ecosystem.
In many ways, the GPU Ventures Program is just a way to formalize what NVIDIA has been doing for the past few years to grease the wheels of the development community. A prime example is Elemental Technologies, developers of the Badaboom Media Converter, whose software now appears for sale on NVIDIA's Web site. NVIDIA is also partnering with MotionDSP to help it launch its soon-to-be-released video enhancement software. Other recipients of NVIDIA's patronage include Keyhole Corporation, whose software evolved into Google Earth; Mental Images Inc., which was later acquired by NVIDIA; and Right Hemisphere, whose software is used to repurpose 3D CAD data for visual communication.
Closer to the HPC space is Acceleware, a company that offers middleware and GPU-accelerated clusters for technical computing applications. In January 2007, NVIDIA invested in the startup to the tune of nearly $3 million. Subsequently, additional investors, including Northern Securities Inc. and Investors Group Mergers & Acquisitions Fund, came in and invested an additional $10 million. NVIDIA also promoted Acceleware's software at tradeshows and analyst meetings, and the NVIDIA sales force pitched applications to customers.
According to Herbst, last year they put up between $4-6 million on just a handful of companies. Now that they've formalized the process, he expects the pace of investments to increase. No specific amount of money has been set aside; that will depend upon the opportunities out there and NVIDIA's enthusiasm for the individual companies. There will also be companies that NVIDIA doesn't invest in directly, but will be providing sales and marketing support. "The most strategic -- the most loved -- companies get investment from us, but we have developer support all the way up and down the line for the full range of companies," says Herbst.
When it does invest, the process is rather straightforward. Since NVIDIA has no limited partners, it doesn't require a separate investment arm. All the money just goes under Profit & Loss on the company's balance sheet. But ultimately NVIDIA is not looking to play the role of a big venture capitalist. The company is first and foremost a GPU manufacturer and it has its financial challenges to deal with. NVIDIA would rather have a real VC firm come in and set the terms of investment and then follow its lead. Part of the program's strategy is to partner with VCs and have them deal with the financial side, leaving the GPU maker to focus its strategic resources on the product side.
Ideally, NVIDIA would like to see the VCs set up a fund specifically designed to feed these GPU computing startups. With 52 companies already signed up to be in NVIDIA's GPU computing club, there should be plenty of demand for investor funds. From Herbst's perspective, "right now the pipeline is pretty full."
10/30/2013 | Cray, DDN, Mellanox, NetApp, ScaleMP, Supermicro, Xyratex | Creating data is easy… the challenge is getting it to the right place to make use of it. This paper discusses fresh solutions that can directly increase I/O efficiency, and the applications of these solutions to current, and new technology infrastructures.
10/01/2013 | IBM | A new trend is developing in the HPC space that is also affecting enterprise computing productivity with the arrival of “ultra-dense” hyper-scale servers.
Ken Claffey, SVP and General Manager at Xyratex, presents ClusterStor at the Vendor Showdown at ISC13 in Leipzig, Germany.
Join HPCwire Editor Nicole Hemsoth and Dr. David Bader from Georgia Tech as they take center stage on opening night at Atlanta's first Big Data Kick Off Week, filmed in front of a live audience. Nicole and David look at the evolution of HPC, today's big data challenges, discuss real world solutions, and reveal their predictions. Exactly what does the future holds for HPC?