Crash and Boom: Inside the 10GbE Adapter Market

By Scott Schweitzer

November 12, 2013

It may be hard to believe, but we’re coming up on ten years with 10GbE as an adapter option for servers and workstations.

In  2003 the first 10GbE network adapters based on a new breed of chips hit the market—and by 2006, the list had eventually grown to include nearly twenty (AdvancedIO, Broadcom, Chelsio, Intel, Emulex, Endace, Mellanox, Myricom, Napatech, NetEffect, Neterion, NetXen, QLogic, ServerEngines, SMC, Solarflare, Teak Technologies, and Tehuti Networks).

Designing & building a 10GbE ASIC is not a cheap undertaking. Even on a shoestring budget it could easily run $7-10M for that first working chip. Some of these companies never made it past that initial functional 10GbE controller chip. The above combined efforts represent nearly one quarter of a billion dollars to launch the 10GbE adapter market. To remain in this market long term most companies have had to follow Moore’s law crafting new chips every two years.  A significant number of the above companies couldn’t generate enough revenue or continue to secure substantial outside investment to sustain this level of development, and have since been acquired or faded away. So what can be learned from these startups?

With two rounds of investment totaling $47M, and four years, in the summer of 2008 NetEffect was the first one to crash. Originally launched from the remains of a failed Infiniband startup called Banderacom. This isn’t uncommon as silicon designs, like aluminum cans, are often recycled.  NetEffect believed their “secret sauce” was adapting Infiniband functionality, actually RDMA, and overlaying it on Ethernet. The actual feature set was branded iWARP, Internet Wide Area RDMA Protocol. NetEffect’s line of 10GbE adapters was launched in May 2007, and the market tested iWARP. There were several early high profile tire kickers, but one year later, after waiting for NetEffect to enter chapter 11, Intel stepped in, and picked up their intellectual property for $8M. Some might consider this strange, but Banderacom’s Infiniband roots were tied to Intel’s original VIA architecture. Intel has a real eye for bargain hunting, this expanded their 10GbE portfolio, and added 30 engineers. The intellectual property fueled by this $47M, and purchased by Intel for $8M still exists in Intel’s product line as several over priced “NetEffect Ethernet Server Cluster” adapters. Sadly, it appears that these products haven’t been touched since 2009. This may not be the last we see of iWARP.

In 2005 Teak Technologies appeared with a $3M series A, and received some early press, but by late 2008 they’d vanished.  Their Internet domain winked off-line, and the founder had updated his Linkedin profile indicating that he’d moved on. Very little is known regarding what happened to them, and we’re not sure whom if anyone ever actually used their silicon.

Tehuti Networks was launched in 2003 through an Israeli government sponsored incubator. Their focus has appeared to be purpose built systems utilizing blade, ATCA, etc… which implies they are likely producing Israeli defense products. In 2005 they required a series A round of $4.2M to keep moving forward.  In 2008 they took another round of funding, but the particulars were never published. Then in February of 2013 a press release appeared offering a one watt 10GbE NIC chip for $10 in quantities of 1,000. Tehuti is still small, under fifty employees, but only five show up in Linkedin which implies that they consider what they’re doing classified. This further supports the presumption that Tehuti is more focused on the Israeli embedded military computing market than they are on 10G Ethernet.

2009 Brought the 11th hour purchase of NetXen by QLogic for $21M. Founded in January 2002 NetXen launched their first 10GbE adapter in  March of 2006. To do this they needed a series A round in June of 2005 of nearly $15M. How much total capital they consumed over those seven years isn’t exactly clear. Compared to Intel’s acquisition of NetEffect this wasn’t a bargain, but it was still a good deal. Early on HP had bought heavily into NetXen’s story, and product line. They were HP’s default 10GbE configuration, which was a very enviable position. Unfortunately their power hungry controller, which required a fan on the heat sink, never gained serious traction in the market. Which explains why they were only in production for three years. Netxen taught us how valuable an early linkage with a major OEM like HP can be to raising your profile in the market. Compared to NetEffect this OEM connection contributed to NetXen having two additional years to reach profitability, unfortunately in the end that wasn’t enough.

Feeling the pressure from QLogic’s acquisition, and fearing the future threat posed by Fibre Channel over Ethernet (FCoE) Emulex expanded their portfolio into 10GbE in the spring of 2010 with the purchase of ServerEngines in a deal whose total cost was $81M.  ServerEngines had several early, and significant, 10GbE wins with IBM that Emulex had admired, this IBM OEM connection dramatically improved their final valuation. Furthermore, unlike NetXen, ServerEngines was moving forward, and had a staff of 170, mostly engineers. This acquisition quickly added 10GbE to Emulex’s line card. What Emulex found out later though was that ServerEngines had infringed on Broadcom’s patents.  Settling this legal issue cost Emulex an additional $58M, bringing the actual cost of ServerEngines to a staggering $139M. The ServerEngines team though still continues to produce, their latest Skyhawk (40GbE) silicon should be coming out soon. ServerEngines once again proved that a solid OEM relationship added significant value to their acquisition.

Neterion started in 2001 as S2io, and was likely one of the earliest, and clearly one of the hungriest for capital in this market. To produce their first chip S2io had to consume two rounds of funding totaling $27M, and in the Fall of 2003 they deliver their initial XFrame 10GbE server adapter for $4,990.  At that point the only other adapter was Intel’s priced at a staggering $7,995 higher. By June of 2004 S2io needed more capital so they closed a series C round for $42M, bringing the total investment to $69M. In January of 2005, just prior to bringing their second 10GbE controller chip to market, they renamed themselves Neterion. Neterion had also seen value in iWARP, and rolled this into their feature set. In March of 2005, nearly a year ahead of the rising tide of 10GbE adapters Neterion launched their second 10GbE product line. Neterion had partnered with SUN Microsystems for this product launch, and later that summer they continued down the Unix path and announced a partnership with IBM’s Power server group. The market perception was that Neterion’s boards were were not tightly integrated which significantly drove up their production costs and power consumption. Their initial products were full height, full length boards while at the same time competitors were offering similar feature sets in half height short cards. Neterion continued to burn capital, requiring additional outside funding, but alas in 2010 they were picked up by Exar for $10M.

Endace took a slightly different approach than most and leveraged its FPGA adapter products to also build appliances, which they then wrapped consulting services around. This broadened their product line, and increased their margin. More importantly though, while everyone else in the market was selling adapters then moving on, consulting enabled Endace to more tightly connect with their customers. This deeper connection educated Endace in a unique way providing them with more detailed customer requirements in the area of Ethernet monitoring & measurement. This enabled them to further tune their appliances to more closely align with the needs of the market.  In 2012 Emulex’s Software Solutions team had partnered with Endace. The partnership was designed to link Emulex’s OneCommand management software to Endace’s appliances so Emulex could expand out of Fibre Channel management into the much larger Ethernet environment. Eventually the partnership lead to the acquisition of Endace earlier this year for $130M.

Myricom is the most recent privately held 10GbE company to fall.  After nearly twenty years, most of it in the High Performance Computing (HPC) market, it was announced earlier this month that a customer of theirs, CSPi of Billerica Mass, had acquired them. In 2006 Myricom crossed over from HPC to Ethernet when they released a line of 10G adapters that were both HPC interconnect NICs and generic 10G Ethernet NICs. For the past 15 years CSPi has been the only significant volume consumer of Myricom’s silicon. These chips were used by CSPi’s MultiComputer division to fabricate purpose built imbedded clusters for defense department projects. According to Linkedin in March of this year Myricom’s former CEO, CTO, several Senior Software Architects, and a significant number of other PHds departed from Myricom and joined Google. Why this happened is still not public, but clearly it’s what lead to the eventual acquisition by CSPi.

So from the original eighteen we’re down to half leaving us with: AdvancedIO, Broadcom, Chelsio, Emulex, Intel, Mellanox, Napatech, QLogic, SMC & Solarflare. Six of the remaining are  publicly traded companies, and most are fairly sound. That leaves only four 10GbE startups: AdvancedIO, Chelsio, Napatech & Solarflare. Napatech, an FPGA based network adapter company, applied on October 24th to the Oslo exchange for an initial public offering. In their filing they stated that their 2012 revenue was $31M. This infusion of cash should fuel their recent push into the next new market, 40GbE adapters. Both Chelsio & Solarflare have announced 40GbE products this quarter, so they too will be moving on. Furthermore each has dominated the market in the initial profitable niche they carved out. Chelsio leads in10GbE iSCSI storage controllers. Early on Solarflare targeted High Frequency Trading (HFT), and as a result 90% of the world’s equities trades happen on server networks powered by Solarflare. Alas, AdvancedIO out of Canada appears to be the one left behind. They haven’t announced plans to move to 40GbE, and haven’t had much success expanding out of their initial Aerospace & Defense market.

So after ten years, and eight 10GbE startups having vanished what have we learned? Four things: partner with a tier one OEM when you have the engineering to support it, continue to innovate, diversify your product line through both hardware & software, and dominate one niche of the market before expanding out. As we saw with several of these startups you need to partner early with two or more top tier OEMs, preferably IBM & HP to help extend your market reach. Tier one OEMs are costly to deal with, demanding & process driven, but they’re global, and in many markets that startups will never reach on their own. These partnerships can extend a startup’s life by several years, and dramatically improve its value. To remain competitive startups must keep the chip crank turning, and continuing to innovate, you can bet their competition is. Several of the failed startups never made it past their first or second 10GbE controller chip. The remaining startups have all done four or more chip generations. Diversify beyond ethernet adapters. Endace caught Emulex’s eye because they had an appliance solution, Emulex was already in a partnership with another 10GbE adapter provider, and had their own ServerEngines 10GbE silicon. In the end Emulex acquired Endace for $130M because they had something substantially more valuable. Solarflare & Myricom both realized that adding high value software to their portfolios is another successful approach. Also Solarflare’s expansion of their ASIC based 10GbE controller line to also include FPGAs with very market specific software should further drive adoption in other very profitable niche markets. Finally, dominate your core market niche then expand out to dominate other niches within that market.

So where do we go from here? Simple, all the remaining players now have to ante up at the 40GbE table, and perhaps some new ones will sit down.

Subscribe to HPCwire's Weekly Update!

Be the most informed person in the room! Stay ahead of the tech trends with industry updates delivered to you every week!

MLPerf Inference 4.0 Results Showcase GenAI; Nvidia Still Dominates

March 28, 2024

There were no startling surprises in the latest MLPerf Inference benchmark (4.0) results released yesterday. Two new workloads — Llama 2 and Stable Diffusion XL — were added to the benchmark suite as MLPerf continues Read more…

Q&A with Nvidia’s Chief of DGX Systems on the DGX-GB200 Rack-scale System

March 27, 2024

Pictures of Nvidia's new flagship mega-server, the DGX GB200, on the GTC show floor got favorable reactions on social media for the sheer amount of computing power it brings to artificial intelligence.  Nvidia's DGX Read more…

Call for Participation in Workshop on Potential NSF CISE Quantum Initiative

March 26, 2024

Editor’s Note: Next month there will be a workshop to discuss what a quantum initiative led by NSF’s Computer, Information Science and Engineering (CISE) directorate could entail. The details are posted below in a Ca Read more…

Waseda U. Researchers Reports New Quantum Algorithm for Speeding Optimization

March 25, 2024

Optimization problems cover a wide range of applications and are often cited as good candidates for quantum computing. However, the execution time for constrained combinatorial optimization applications on quantum device Read more…

NVLink: Faster Interconnects and Switches to Help Relieve Data Bottlenecks

March 25, 2024

Nvidia’s new Blackwell architecture may have stolen the show this week at the GPU Technology Conference in San Jose, California. But an emerging bottleneck at the network layer threatens to make bigger and brawnier pro Read more…

Who is David Blackwell?

March 22, 2024

During GTC24, co-founder and president of NVIDIA Jensen Huang unveiled the Blackwell GPU. This GPU itself is heavily optimized for AI work, boasting 192GB of HBM3E memory as well as the the ability to train 1 trillion pa Read more…

MLPerf Inference 4.0 Results Showcase GenAI; Nvidia Still Dominates

March 28, 2024

There were no startling surprises in the latest MLPerf Inference benchmark (4.0) results released yesterday. Two new workloads — Llama 2 and Stable Diffusion Read more…

Q&A with Nvidia’s Chief of DGX Systems on the DGX-GB200 Rack-scale System

March 27, 2024

Pictures of Nvidia's new flagship mega-server, the DGX GB200, on the GTC show floor got favorable reactions on social media for the sheer amount of computing po Read more…

NVLink: Faster Interconnects and Switches to Help Relieve Data Bottlenecks

March 25, 2024

Nvidia’s new Blackwell architecture may have stolen the show this week at the GPU Technology Conference in San Jose, California. But an emerging bottleneck at Read more…

Who is David Blackwell?

March 22, 2024

During GTC24, co-founder and president of NVIDIA Jensen Huang unveiled the Blackwell GPU. This GPU itself is heavily optimized for AI work, boasting 192GB of HB Read more…

Nvidia Looks to Accelerate GenAI Adoption with NIM

March 19, 2024

Today at the GPU Technology Conference, Nvidia launched a new offering aimed at helping customers quickly deploy their generative AI applications in a secure, s Read more…

The Generative AI Future Is Now, Nvidia’s Huang Says

March 19, 2024

We are in the early days of a transformative shift in how business gets done thanks to the advent of generative AI, according to Nvidia CEO and cofounder Jensen Read more…

Nvidia’s New Blackwell GPU Can Train AI Models with Trillions of Parameters

March 18, 2024

Nvidia's latest and fastest GPU, codenamed Blackwell, is here and will underpin the company's AI plans this year. The chip offers performance improvements from Read more…

Nvidia Showcases Quantum Cloud, Expanding Quantum Portfolio at GTC24

March 18, 2024

Nvidia’s barrage of quantum news at GTC24 this week includes new products, signature collaborations, and a new Nvidia Quantum Cloud for quantum developers. Wh Read more…

Alibaba Shuts Down its Quantum Computing Effort

November 30, 2023

In case you missed it, China’s e-commerce giant Alibaba has shut down its quantum computing research effort. It’s not entirely clear what drove the change. Read more…

Nvidia H100: Are 550,000 GPUs Enough for This Year?

August 17, 2023

The GPU Squeeze continues to place a premium on Nvidia H100 GPUs. In a recent Financial Times article, Nvidia reports that it expects to ship 550,000 of its lat Read more…

Shutterstock 1285747942

AMD’s Horsepower-packed MI300X GPU Beats Nvidia’s Upcoming H200

December 7, 2023

AMD and Nvidia are locked in an AI performance battle – much like the gaming GPU performance clash the companies have waged for decades. AMD has claimed it Read more…

DoD Takes a Long View of Quantum Computing

December 19, 2023

Given the large sums tied to expensive weapon systems – think $100-million-plus per F-35 fighter – it’s easy to forget the U.S. Department of Defense is a Read more…

Synopsys Eats Ansys: Does HPC Get Indigestion?

February 8, 2024

Recently, it was announced that Synopsys is buying HPC tool developer Ansys. Started in Pittsburgh, Pa., in 1970 as Swanson Analysis Systems, Inc. (SASI) by John Swanson (and eventually renamed), Ansys serves the CAE (Computer Aided Engineering)/multiphysics engineering simulation market. Read more…

Choosing the Right GPU for LLM Inference and Training

December 11, 2023

Accelerating the training and inference processes of deep learning models is crucial for unleashing their true potential and NVIDIA GPUs have emerged as a game- Read more…

Intel’s Server and PC Chip Development Will Blur After 2025

January 15, 2024

Intel's dealing with much more than chip rivals breathing down its neck; it is simultaneously integrating a bevy of new technologies such as chiplets, artificia Read more…

Baidu Exits Quantum, Closely Following Alibaba’s Earlier Move

January 5, 2024

Reuters reported this week that Baidu, China’s giant e-commerce and services provider, is exiting the quantum computing development arena. Reuters reported � Read more…

Leading Solution Providers

Contributors

Comparing NVIDIA A100 and NVIDIA L40S: Which GPU is Ideal for AI and Graphics-Intensive Workloads?

October 30, 2023

With long lead times for the NVIDIA H100 and A100 GPUs, many organizations are looking at the new NVIDIA L40S GPU, which it’s a new GPU optimized for AI and g Read more…

Shutterstock 1179408610

Google Addresses the Mysteries of Its Hypercomputer 

December 28, 2023

When Google launched its Hypercomputer earlier this month (December 2023), the first reaction was, "Say what?" It turns out that the Hypercomputer is Google's t Read more…

AMD MI3000A

How AMD May Get Across the CUDA Moat

October 5, 2023

When discussing GenAI, the term "GPU" almost always enters the conversation and the topic often moves toward performance and access. Interestingly, the word "GPU" is assumed to mean "Nvidia" products. (As an aside, the popular Nvidia hardware used in GenAI are not technically... Read more…

Shutterstock 1606064203

Meta’s Zuckerberg Puts Its AI Future in the Hands of 600,000 GPUs

January 25, 2024

In under two minutes, Meta's CEO, Mark Zuckerberg, laid out the company's AI plans, which included a plan to build an artificial intelligence system with the eq Read more…

Google Introduces ‘Hypercomputer’ to Its AI Infrastructure

December 11, 2023

Google ran out of monikers to describe its new AI system released on December 7. Supercomputer perhaps wasn't an apt description, so it settled on Hypercomputer Read more…

China Is All In on a RISC-V Future

January 8, 2024

The state of RISC-V in China was discussed in a recent report released by the Jamestown Foundation, a Washington, D.C.-based think tank. The report, entitled "E Read more…

Intel Won’t Have a Xeon Max Chip with New Emerald Rapids CPU

December 14, 2023

As expected, Intel officially announced its 5th generation Xeon server chips codenamed Emerald Rapids at an event in New York City, where the focus was really o Read more…

IBM Quantum Summit: Two New QPUs, Upgraded Qiskit, 10-year Roadmap and More

December 4, 2023

IBM kicks off its annual Quantum Summit today and will announce a broad range of advances including its much-anticipated 1121-qubit Condor QPU, a smaller 133-qu Read more…

  • arrow
  • Click Here for More Headlines
  • arrow
HPCwire