on Apr 19 2015
We build devices that cool down electronics, using a technology called “Kinetic Cooling.” Electronics produce a lot of heat when they are being used. In order for them not to melt or malfunction, they need to be cooled down. For the past 50 years cooling has been done by using a block of metal (heat sink) to conduct heat away from the chips and using a fan to blow air across the metal. The whole setup hums loudly and takes up a lot of space.
We asked ourselves, “why do we need two parts?” A key insight is that plastic fans embedded into a heat sink waste space. In theory, couldn't we combine the two and have the assembly become half the size? That's what we've done. Kinetic Cooling reclaims the space and provides an alternative heat path for heat flow. Heat takes the path of least resistance. When there are two paths, there is less overall resistance.
We've combined the fan and the heat sink into one assembly. When you do that, everything can shrink by up to a factor of two, while providing acoustic and efficiency gains. Some of our hardware is a hundred times quieter, and the efficiency goes up, meaning you can remove more heat than you previously could. This enables manufacturers to add more processing power than they could before Kinetic Cooling.
If you look at the evolution of Intel processor speeds over the last two or three decades, you'll see that processor speeds have doubled roughly every 18 months (Moore's law). This exponential performance increase has increased the heat output, but has not been matched by cooling performance increases. There are billions of transistors in very small packages generating a lot of heat, but there is no good way to get rid of the heat. Semiconductor technology has enabled better processors to get smaller and smaller - think about cell phones, clients, laptops, etc. - but cooling solutions have gone in the opposite direction. This is a problem and many are struggling for solutions. The fan-cooler protects the electronics that have benefited from Moore’s Law, but has not benefited itself. Fan-coolers have become bigger, louder, and less efficient.
In the electronics industry, size matters. You want something that is smaller, fits in your pocket, but also something that's quieter. You don't want it to cost you an arm and a leg. We are allowing manufacturers to decrease size while still improving performance. Manufacturers have to resort to all kinds of active power management to keep heat related issues at bay, but sacrifice performance in return…resulting in slower speeds, poor connectivity and lack of reliability.
Consumers are doing more and more things that require more processing power and bandwidth. Since a lot of the heavy lifting is shifting to the cloud, the thermal challenges in data center and enterprise servers are burgeoning. The chip-makers always want to market the fastest chips, and data centers want to pack in as many processing units as they can into a rack to be as efficient as possible. Additionally, significant challenges are beginning to show up in the Telecom space, particularly in network stations and wireless towers. If we want more performance, if we want movies to not buffer for a long time, if we want to do new computationally-intensive tasks, whether it's gaming or streaming from somewhere very far from us to our laptops, or whether you're an engineer or scientist who's actually doing locally intensive tasks in a professional context, then we need to solve the heat problem.
Historically, thermal management has always been an afterthought when it comes to system design. Traditionally, organizations set aside very few dollars to improve on the fundamental technology. Only near the turn of the millennium did DARPA identify thermal management was a problem. DARPA looks out a decade in advance of when we may hit certain technological frontiers. And, they identified thermal management as being on a crash-course.
Thus, there haven’t been significant investments in this space, and as a result the space has seen very little innovation. Additionally, the big thermal management players have competed in relatively low-margin businesses with the OEMs exerting a lot of pressure to push down prices. They have been squeezing suppliers for every penny that they can in order to extract as much value as they can. Cooling OEMs & suppliers have been wholly focused on reducing costs for the existing technology and haven't had the R&D expertise in house to come up with next-generation solutions. A typical cooling OEM spends less than 1% of its revenues in advanced technology.
However, over the last 10-15 years, there have been increasing investments in the thermal space for new technologies, as well as increased M&A activity where the bigger cooling OEMs have tried to acquire new competencies. The time appears to be ripe for a technology like Kinetic Cooling to make a significant impact in the thermal management space.
Absolutely. The Amazons, Facebooks and Googles of the world also have issues with their existing cooling solutions. The typical server rack for servers has an array of boxes that are 1.75 inches tall. You have to squeeze a bunch of electronics, and fans to cool them in there. The specifications limit how tall the fans can be, how much heat they can remove, and ultimately, how much processing power a rack can support. The constraints make rooms super cold to mitigate the heat, but they still have the fundamental issue of making the cooling solution inside the actual IT equipment - that is the servers - more efficient.
Some companies are trying to change the rack architecture (e.g. ,Facebook’s Open Compute Platform), but they have lowered the density of the rack to address thermal challenges. This presumes you have the luxury of vast amounts of physical real-estate. The typical data center operator does not have this luxury. Our goal with Kinetic Cooling is to allow data centers to maintain and increase rack density to improve computing performance while maintaining efficiency.
Noise is also a huge issue for these data centers. If you're a data center operator, and need to replace or maintain equipment that goes down, you often have to go into these rooms with thousands or tens-of-thousands of boxes, each with about 12 tiny fans per box. If you have 42 boxes in a rack, and tons of racks, it gets really loud, sometimes sounding like a jet engine at a few meters away.
Regulators impose noise limits on how loud these rooms can get. It becomes a health hazard, which leads to lawsuits and loss productivity. In the few years, the two bodies that regulate the back office and data center environments, NEBS and ASHRAE, have declared reductions in noise levels in data centers by ~8 decibels, which equates anywhere from 4-8X reduction. It's very hard to imagine how today's suppliers and OEMs can use existing solutions to meet those restrictions. There is no way of doing it.
If you are a data center operator, and you know Cisco’s or Dell’s equipment will not meet regulation, guess what? You’re going to go to the next competitor whose equipment can. Otherwise, you’ve got a major liability in your hands. Kinetic Cooling can potentially ameliorate the stringent constraints for many OEMs.
We expect to be cost competitive with traditional forced-air convection systems. We are committed to delivering innovation that seeks multiple ways to be cost efficient such as leveraging existing high-volume manufacturing processes.
There's no point in being an exotic, premium solution that will exclude certain segments of the market from benefiting from the technology. That means every laptop user and gaming console user, but also the data center operators who are extremely cost-sensitive. That said, we want to make sure that we get paid for the additional value we bring to the table.
Generally, we are within 20% of the alternatives when it comes to cost, and in some scenarios, when we include the system-level cooling benefits we could deliver, we could reduce the total cost of ownership (TCO). It varies by the type of system. If you look at a typical 1U server system and you replace the fans now with Kinetic Cooling, you'll actually save money on the upfront capital equipment and over the life of a box, the expense of operating that unit.
In terms of performance, we have shown up to 50% better thermal efficiency, 10-15 dBA lower acoustics and up to 2 times smaller. We think Kinetic Cooling could be more robust than conventional fan solutions, due to an essentially sealed motor configuration, but that claim can strictly be made only after running a series of reliability tests.
More efficient cooling to increase processing speed is what customers want out of a cooling system, but they also need the system to be reliable, price competitive, and generally fit in with their current design process. You cannot have downtime at a mission-critical server farm. Reliability matters a whole lot. We made the physical design of our system similar to the incumbent technology so that it's easy to adopt Kinetic Cooling into future product platforms. Price matters, reliability matters, lifetime matters, and performance certainly matters.
One of the partners we've teamed up with is a large Taiwanese company called Cooler Master. They are one of the most well-recognized brands in the electronics cooling industry on the gaming side. They have an active market and very strong retail presence. They also serve every name brand in the electronics OEM space. That has been our approach to be cost-competitive and have supply stability.
Retail products need to have 30,000-50,000 hours of lifetime backed by the manufacturer. A unit going into a data center likely needs a few hundred-thousand hours. To prove this, you need to sell a bunch of units, and collect data on how reliable they are in the field, while also getting accelerated test data in the lab coupled with HALT/HAST testing.
There are qualification processes that each OEM has, which can run anywhere from 6 to 18 months. For any supplier to engage in that investment, they’d better be really sure that they’re going to pass at the end. That’s why it’s important for us to generate revenues in less demanding market segments on the path to winning the large OEM business.
Cooler Master is one of the most recognized brands in the retail market. They have products at the major big-box retailers, such as Micro Center, Best Buy, and Fry's, and most of the Enterprise users that build their own systems know who they are.
The partnership is the perfect marriage of our strengths.
Cooler Master has two separate entities. One side of the company does business with OEMs, and the other side works with retailers & distributors. The best strategy for us is to go into retail first and use that as a field test. We can then take what we learn and the revenues from there to fuel design of OEM products.
We're infusing Kinetic Cooling technology into a family of Cooler Master’s existing products. We showcased some impressive hardware at the 2015 Consumer Electronics show in Las Vegas and received tons of attention around the joint effort, and interest from distributors that are eager to try Kinetic Cooling. We received overwhelming positive coverage from over 15 media outlets such as Tom’s Hardware, IEEE Spectrum, and PC World. We even made best of CES round-up by TechGage. We earned over 10 million media impressions. Based on that feedback we are working on a development effort with Cooler Master on a low-profile high-performance gaming unit. We're trying to get that unit out late-2015 to early-2016. We expect to sell anywhere from 75,000-125,000 units that first year. That should be a pretty successful product launch.
It’s important to recognize just how little innovation there has been in this space. The current cooling products are commodities, and for manufacturers to secure their future, they need to provide new solutions for their customers' platforms. Cooling solution manufacturers know that if they only stick with the current technology they’re not going to make money. We estimate that we can make conventional fan technology obsolete in certain spaces within the next 5-10 years. That gets us excited, and it gets a lot of our potential customers excited.
Right now we are looking at both licensing and reselling models, so we will see revenues through royalty payments and potentially some initial low volume high, margin sales. For some of the higher volume opportunities, leveraging a partner that is already selling in high volume in existing spaces is a good way for us to get started. This is really the only way that makes sense for us to get started since we need to have hundreds of thousands of units in the market as soon as possible to start proving out reliability while establishing our credibility. Alongside that, if we can develop a lower volume, niche customer-base, we can start developing our sales force to slowly expand into other high volume spaces.
The part that we focus on standardizing is the rotating element of the assembly, the Kinetic Cooling Engine. The engine takes the most time and effort, since it's the most complicated piece. Once we standardize the first three engines we can build several different form factor products around those engines to achieve different performance levels. These different products can then be proliferated into multiple market verticals. The initial three engines will form the core of our first 10-15 products. Currently, from a product roadmap standpoint, the PC gaming solution is likely the first to be released, followed by a 1U/2U form factor server solution. Additionally, we are talking to industrial customers, where there may be opportunities for some niche high margin products. We expect to have a clearly defined product roadmap by 3Q-2015.
The cell phone market is a stretch for Kinetic Cooling in the short term. The smallest profile that we can target right now is the Microsoft Surface Pro. We have some other stuff in the works for those very thin applications, but it is based on a different physical phenomenon than Kinetic Cooling. In fact, we haven't even discussed it publicly yet. It is still in early stages of development and not ready to be integrated in our first set of products. We can say that we are looking to make thumbnail size cooling elements that would be sub-mm thick. These would be perfect fits for a myriad of wearable and mobile hardware.
The overall thermal management market is a 10 billion dollar space. Our focus is currently on computing, which would be microprocessors, graphics processors FPGAs, ASICs, etc. that could go into several market verticals like consumer devices and servers. That’s about 3 billion dollars, a third of the overall thermal management space. Following that we have industrial and telecom, which we can go after once we have established reliability. This market is about 1 billion dollars. Within computing, the initial market we’re targeting is the high-performance computing segment, which is roughly a half-billion dollar segment. That includes gamers, and the enterprise users who need to get the most out of a workstation for productivity. Once we have become established in these spaces, we can introduce new technology for mobiles and wearables that allow us to address a sizeable piece of the remaining of 6 billion dollar space.
Microsoft hired us to develop more silent, more efficient cooling systems for the Xbox. Now, it is odd for system OEMs to invest directly into components, but Xbox has had a really tough time finding the right cooling solution for their system, and are no longer leaving it to the suppliers to solve the problem on their own. They experienced a huge issue in one of their Xbox 360 consoles a few years ago. Consumers want smaller consoles, and the cooling system is what is preventing that. That’s why they’ve brought us in to help them achieve that using Kinetic Cooling technology. We are also in discussions with several players across the ecosystem of the electronics industry from chip-makers, like Intel and AMD, all the way to contract manufacturers, like Foxconn and Delta.
Intel occupies a pretty significant market share in the semiconductor industry. They set a lot of the standards and every big OEM that uses an Intel processor says, “Hey, Intel, you made this chip. Who do you recommend we go to for a cooler?” So, we’re working with Intel. We say, “Hey, Intel, we see you’ve been doing this one way for so long, and you should consider integrating these two parts (fan plus heat sink) moving forward. And, then everything gets smaller; we don’t need to have these mechanical hang-up reasons that are very restrictive for the system builder. Everything is quieter. And hey, by the way, we can be cost-competitive.” We’re working with workstation groups, laptop groups and currently a PC gaming unit at Intel.
We’re the only company that’s deployed Kinetic Cooling technology where we’re getting heat onto a rotating reference frame. Other companies have tried this in the past, but it wasn’t nearly commercial. This is a first.
Our best partners are also our biggest competitors. Cooler Master makes fan and heat sink assemblies. Strictly speaking, they are a competitor. But they, like others, also know that if they don't invest in innovation they’ll be rendered obsolete. For us, they bring things to the table that we need. They have the manufacturing wherewithal and expertise; we don’t have to build that. We’ve structured our dealings such that we are not locked into an exclusive relationship with any of our partners. If it makes sense for us to own more pieces of the value chain, such as manufacturing capabilities, in the future, we will do that.
At this point, it’s reliability, and only because we have not build enough units and put them through the paces. We’ve done enough to convince Cooler Master to start looking at putting 100k units into the field. They have been in the space long enough to identify where potential failure modes are, and say “Hey, this is something we can overcome, we do this all the time.” Cooler Master can validate that our biggest risk is not as big of a factor as some would have think.
From a fundamental technology standpoint, there are no sub-components in Kinetic Coolers which represent an “unknown risk.” A lot of the sub-component technology that goes into Kinetic Cooling is well known from a manufacturing standpoint. The risk is in how the overall system behaves when these components come together and that is a matter of getting in a test lab and going through the testing.
The three founders met at MIT. Our technology lead, Lino, earned a PhD at MIT in chemical engineering. Lino and I met in 2003. He and I founded a cleantech company in the automotive space in 2008, which was a rough time for startups selling to auto manufacturers. We learned a ton and were exposed to the thermal management space. In 2010, when we met Steve, our operations lead, it was the perfect storm. We won the MIT Clean Energy Prize in 2011 and got to work right away.
Today, we’re a team of seven. Our VP of Business Development is a seasoned entrepreneur with a thermal background. He started a thermal management company, much like CoolChip, that was acquired last summer by one of the tier-one suppliers. When he became a free agent, it was the perfect time in the life of CoolChip. We have a Director of Marketing who is a program manager and marketing powerhouse. She’s got a brand management background from Proctor & Gamble, Unilever, and most recently the Playstation division of Sony. She’s very good with the retail consumer space. She understands consumer electronics, she understands the electronic industry, and she’s a phenomenally talented lady. We’ve got two other engineers that work with us at the Boston office: a senior mechanical engineer and another engineer with a lot of background in gaming and in industrial design. That’s the full-time team.
We are one of Sandia’s Labs industry partners. We have licensed something from them called the Sandia cooler and then built our own IP around it to turn it into something that was commercialize-able. We have thirteen, and counting, patented & patent-pending filings currently, and our portfolio is growing rapidly year-on-year.
Already have an account? Login
Don't have an account? Signup