Myricom, An HPC Story and Lessons Learned from the Fall of an Industry Darling
For nearly a decade and a half, Myricom, a hot tech startup at the turn of the Millenium, measured success not by profitability, but by the percentage of the Top500 running its flagship Myrinet interconnect. Sure, we had to be profitable, but the CEO/founder was laser-focused on the penetration of his network into the fabric of supercomputing. Myrinet-2G dominated the list for several years, but by SuperComputing in 2009 (SC09) in Portland, it had become apparent that Myricom had lost the number one Interconnect position, possibly forever, to Infiniband.
2008 brought the Mortgage crisis, which ran well into 2009, and impacted nearly every sector of the economy, including supercomputers. So as we rolled into SC09 in November for our big trade-show event of the year, and Myricom found its booth located directly across the main aisle from Voltaire, one of our two primary competitors. This was shaping up to be a battle not seen since the Hatfields fought the McCoys. Voltaire and Mellanox were Myricom’s sworn enemies; they peddled Infiniband, an inferior design based on bus technology, but repositioned as a point to point network while we invented Myrinet specifically for HPC fifteen years earlier. Enventually Mellanox had gobbled up Voltaire, but at this point in 2009 they were still separate companies, and Voltaire had positioned their newest and largest HPC switch along the pathway separating the booths and faced it directly at Myricom. A sign on Voltaire’s switch claimed that it was the highest density switch available. The CEO of Myricom took issue with Voltaire’s claim and promptly removed, and discarded their sign on opening night. We arrived at the show the following morning and found that Myricom had a new sign facing Voltaire’s booth, calling out its switch’s superiority. As the show progressed, additional Voltaire’s signs appeared throughout the event, only to vanish within hours. In retrospect, Myricom was in denial, a particularly dangerous and too frequently seen symptom of a technology company racing headlong toward the edge of a cliff.
During Myricom’s first ten years, 1994–2004, it grew to a staff to just over forty people, about half of those being PhDs, and often they saw annual revenues north of $20M. Myricom had successfully navigated several product generations, and by the summer of 2005 had more than 27% of Top500 Super Computers wired with Myrinet, second only to free Gigabit Ethernet. They’d become the undisputed king of High-Performance Computing (HPC) Interconnects. At that point Infiniband was 5% or less, similar to Quadrics success. Unfortunately, despite its auspicious beginnings, the following years showed that Myricom was a startup that refused to grow. The CEO capped the full time headcount at 49 people so he wouldn’t be subject to California law requiring him to provide his employees with a health plan. Its failure to expand beyond forty-nine full-time employees hampered its growth and stunted its corporate maturity. Eight years after reaching the zenith of its success, its c-suite had enabled Google to swoop in and strip out much of its future value. What remained were ten employees, none of which were PhDs, a seven-year-old 10GbE controller chip, and a bunch of Myrinet-2G technology.
I joined Myricom in August in 2005, to improve their relationship with IBM and facilitate the expansion into a 10GbE NIC business. In 2012 we were doing nearly $40M annually, and the vast majority of that was 10GbE NICs. By May of 2013 I was selling off everything that wasn’t nailed down. I’d managed to sell the remaining Myrinet-2G stock to Myricom’s only chip-down customer, CSPi, for several million dollars, creating much needed operating capital for the ten remaining employees. Nobody was buying Myrinet-2G anymore for HPC, and CSPi wanted our remaining stock, masks, tools, and test equipment to satisfy some long-tale Department of Defense (DoD) contracts they held. On completing that sale, I suggested we to approach them to buy what remained of the company. At that time it was their Myri-10G HPC business which was in decline, and our rapidly aging 10GbE NIC business. While I started the ball rolling on the acquisition, and had an agreement with the acting CEO to be paid on its successful completion, but I wasn’t compelled to remain. So I left for one of Myricom’s 10GbE competitors, Solarflare Communications, in August of 2013. In November, the deal with CSPi closed, and Myricom was sold for less than one percent of the value they had sought exactly one year previous from Emulex. Stripped of critical skills and resources, they no longer had chip designers, only two hardware engineers, a few programmers and one sales person. The markets for its remaining products were shackled with legal restrictions imposed by Google.
Why did this promising technology company fail, and what lessons can we learn from their failure?
Most successful companies operate with a plan. During the first sixteen years of its existence, Myricom was led, and its corporate culture was defined by a single person whose vision was to build the fastest & most efficient HPC network possible. He hired the best and brightest and created an impressive suite of products for the HPC market that dominated for several years. While Myricom was growing and a technological leader, it thrived without a formal business plan, I asked several times from 2005 to 2010 to see the plan and was told by my VP of Sales there wasn’t one, as far as he knew. However, when the HPC market began to turn toward Infiniband, without a formal business plan or a competitive strategy we lost ground. Further, this executive ascribed to the belief that ‘great products sell themselves’ and that a sales team was nothing more than a necessary evil… marketing was little more than a task to be contracted out to a part-timer to compose datasheets and issue press releases. Myricom had hired its first sales person after six revenue producing years, in 2000. Then sales people two through five were all hired in 2005 when there was a notion of also branching out into the generic 10GbE NIC market. In those early 10GbE days of 2005–2008 HPC deals always got priority over the 10GbE business.
CEO’s need to focus on selling the business to customers and investors. It is typical of nearly every successful startup that the CEO is also its most effective salesperson, often spending a considerable amount of their time selling the company to existing and potential investors and customers. CEOs also often define the corporate culture and stoke the fire within the employees of the company. The CEO’s of startups usually seem to thrive on evangelizing their company to anyone who’ll listen. However, this was not the case with Myricom; once the decision was made to stop actively selling HPC switches and move into the broader 10GbE market, their fortunes diminished noticeably. The already minimal number of customer calls involving the CEO with HPC customers had dropped off substantially as the company shifted markets. Myricom’s CEO was a veritable celebrity in the HPC community, yet in the wider (and larger) enterprise Ethernet market, he was a virtually unknown.
CEOs need to listen more, and talk less. Myricom’s CEO knew more about networking than its customers, and this know-it-all attitude was a pervasive thread throughout our corporate culture, permeating all of its business relationships and partnerships. In my first five years with the company, I handled east coast sales, and had only taken the CEO on a handful of customer calls. He lectured each and every one. Now it should be noted that our CEO was a former professor at CalTech and MIT before that, so he was always more comfortable lecturing than listening. As we moved into 10GbE, I was instructed not to bring him on those calls as he had nothing to tell them. In 2008 and 2009, the US Navy and a three-letter US government intelligence agency were two of Myricom’s top five accounts. They’d never spoken with, let alone met with, the CEO. As mentioned earlier, Myricom seemingly operated without a business plan, and sales were made wherever the sales team was able to find traction. The sales guys talked with each other almost daily to share tips and tricks that worked.There often was no established set of sales milestones, and MBOs were written after the quarter had ended to justify the results; this was a rudderless ship under full sail.
CEO’s should be excellent at delegating, and expert as not getting hung up in the details of day to day operation. For the first fifteen years, Myricom’s CEO was also its webmaster; he wrote nearly every page of Myricom’s website in raw HTML. Now it should be noted this started in 1994 when pretty much everyone was writing in HTML, so back then, it wasn’t uncommon. He was an engineer at heart, and the company was run by engineers for engineers.
CEO’s need to know how to work with the board, and listen to those closest around them. In the summer of 2010, the CEO lobbied the board for new outside investment to accelerate the completion of Myricom’s latest Ethernet Controller processor, the Q20. Outside investment often necessitates diluting the value of the initial investor’s shares (who comprised the majority of the board) in exchange for the longer-term benefit of that capital. The new Q20 chip was designed as a dual-core 10GbE high-performance packet processor that would scale to 40GbE and eventually 100GbE interfaces focused on lowering both TCP & UDP latency. By the summer of 2010, the Q20 was on schedule for completion in December, and had this product made it to market in the spring of 2011 as originally planned, Myricom would likely still be a major force in the high-performance Ethernet market. Unknown to the CEO, though, his second in command had lobbied the board to take a different approach: issue a dividend to shareholders and deliver the Q20 chip a bit later. It was never fabricated while at Myricom nearly three years later. The board believed, and followed this short-sighted approach, compounded their error by jettisoning the founder and CEO, and installing the second in command as the new chief executive.
With enough time a poor direction can often be changed, and a company turned around. Even though the departing CEO initiated Myricom’s death spiral, the new Chief had more than adequate time and the resources necessary to pull it out. Alas, it didn’t happen; she superficially updated the company logo, added several ineffectual staff, which she later fired, and replaced the VP of Sales who’d been with the business for a decade at this time. While the sales team waited for the promised Q20 and kept their key customers engaged, she secretly gathered up the completed Q20 chip design and eleven other of the company’s key PhDs. These twelve formed the “Google twelve,” who then jumped ship in March via a technology transfer deal. Incredibly, the board approved this agreement, which garnered far less than the amount the CEO had sought six months earlier in an acquisition from a more suitable business partner, Emulex.
Why did Myricom’s second CEO fail so fast, lack of experience, and loss of respect? If one were to attempt to explain the plummeting fortunes of Myricom during the term of the second CEO, two characteristics suggest themselves: a lack of experience, and the loss of respect and credibility by rank and file employees. Unlike Myricom’s founding CEO, who had professional experience working at several other companies before Myricom, this CEO’s resume only included Myricom. While Myricom’s founding CEO earned respect through technical leadership, his successor who was the prior VP of Operations, displayed no such talent. The initial “honeymoon period” any chief executive enjoys is usually short-lived. At Myricom, it lasted for perhaps two months, then reality hit home as product schedules continued to slip, and key milestones came and went, unrealized.
CEOs have a fiduciary responsibility to the board and share holders, but that isn’t always respected. The second CEO earned her MBA at UCLA while serving her initial year as the company’s chief executive. When the second CEO was named, the employees assumed that it was intended to save the business and grow revenue. So they remained with the company and prepared to fight and win back lost market share and leadership. The new CEO formed a partnership with Emulex in the spring of 2012, but it never produced any substantial sales for either party. Regardless, Emulex had serious acquisition discussions with Myricom throughout 2012, but by the fall, Emulex was offering only half what Myricom sought for the company. Roughly equal to Myricom’s annual revenue at that point. At the same time Google had crafted a deal to acquire the Q20 and some of Myricom’s prized PhDs. At the end of 2012, Emulex had announced its acquisition of Myricom’s competitor Endace, which put an end to any credible hope of acquisition for itself.
It should be noted that Myricom never took any Venture Capital, only Angel Investor money, and only very early on. During the company’s existence, those Angels were paid dividends several times, and I was told they had received at least a 5X return on their initial investments over the years leading up to 2010. Many of the Angels were also board members, why they had not accepted Emulex’s offer in the fall of 2012 is still a mystery.
So how did the Story end?
Myricom started 2013 with 40 something employees, and by the end of March, with the departure of the “Google Twelve,” instantly, others resigned. In May, a layoff shed several more employees, and over the summer, a few more, including myself, left voluntarily. Early fall saw all non-essential people asked to leave. Finally, during the first week of November CSPi (of Billerica, MA), an existing Myricom chip-down customer, acquired the brand and rights to continue manufacturing, selling, and supporting current products, but under some very restrictive terms spelled out by Google. CSPi also hired most of the few remaining employees from Myricom.
How about the people involved, how did they make out?
If you were one of the “Google Twelve,” you cashed out and did pretty, pretty good. Every one of them earned at least seven figures, or so I’ve been told.
If you were a Myricom Angel, while you did well over the years with your investment in Myricom, taking several dividens, your final payday was a small fraction of what it would have been under an Emulex deal. The decision to reject the Emulex deal came at nearly the same time as Google was expressing their interest, and some might say it was too coincidental.
If you were a rank and file employee, the people who handled the heavy lifting, closed the deals, built the brand, and brought new products to market, you saw nothing beyond your normal payday. The moral of the story is simple, expect that in a small start-up your stock will NEVER be worth anything, and get what you can in your normal paycheck.
