Most promising start-up in 2005: Mellanox

CEO Eyal Waldman wants a mega-IPO.

Mellanox Technologies has been chosen as the most promising start-up in Israel in 2005. Mellanox develops InfiniBand solutions that facilitate high-speed information transmission between clusters of computers, storage components, and servers.

It is not beyond the realm of possibility that by the time next years most promising start-up is chosen, Mellanox will already have held a Nasdaq IPO at a company value in the hundreds of millions of dollars. The company achieved profitability last quarter, and has estimated annual sales of over $40 million. CEO Eyal Waldman wants a mega-IPO. Hes waiting for Mellanox to grow further before he starts the process, one company investor told me.

Mellanox regularly appears on Red Herring magazines 100 Europe list of private companies in Europe and Israel that play a leading role in innovation and technology. Several weeks ago, the company also made sixth place on the list of the top ten private companies compiled by Byte and Switch, a widely respected storage website.

Mellanox began when Waldman, co-founder and former VP engineering at Galileo Technologies, left that company a year before it was sold to Marvell Technology Group (Nasdaq: MRVL) for $2.7 billion.

Waldman managed to raise the fantastic sum of $89 million before the bubble burst. That gave the company enough oxygen to last through the past few years. We managed to more than double our revenue in each of the last five years, and were seeing very nice growth in the markets now, he says.

Mellanoxs investors include several giant companies who decided to gamble on InfiniBand: Dell Computer (Nasdaq: DELL), IBM (NYSE: IBM), Sun Microsystems (Nasdaq: SUNW), and Intel (Nasdaq: INTC). US funds Sequoia Capital, USVP, Bessemer Venture Partners, and Raza Venture Management own most of Mellanoxs shares. Israeli funds Gemini Israel Funds, Jerusalem Global Ventures, and Walden Israel have also invested in the company, and there are other investors, too. Mellanoxs financing round in late 2001 was held at the huge company value of $246 million, after money.

Surviving the technology desert

The InfiniBand (infinite bandwidth) standard came into existence as an architecture for high-speed data communications between servers, storage components, and clusters of computers. InfiniBand competes with other data communications technologies, such as Fiber Channel. Its cost-performance ratio is better than that of another competitor, Ethernet.

The need for InfiniBand is growing, as grid computing becomes more and more popular. Grid computing costs much less than the older mainframe computers, but it requires a high level of connectivity between servers and storage parts.

Had this review been written three years ago, however, Mellanoxs chances of making the top ten would have been very poor. At that time, many companies that had previously supported the InfiniBand standard made an about-face, and Mellanoxs future looked rather shaky.

Intel, the worlds largest manufacturer of processors, decided against developing processors for InfiniBand. Microsoft (Nasdaq: MSFT) decided not to support it, as did Red Hat (Nasdaq: RHAT) and Novell (Nasdaq: NOVL). One after another, a long list of start-ups operating in the InfiniBand field, which had raised hundreds of millions of dollars from investors, collapsed.

We went through a tough three years, in which the technology industry almost stagnated. Many companies that began before we did failed to survive this technology desert. Developing new technology requires large resources and many years, Waldman says.

InfiniBand technology has recently recovered from its crisis, but it is still far from fulfilling the great hopes placed on it. Analysts estimate sales of InfiniBand-related processors and equipment at $75 million, which they expect to grow to $850 million by 2008. Those are big numbers, but thats not what investors had in mind when they poured huge sums into the field.

Mellanox does not have many competitors. Waldman says that his company has an 85% share of the InfiniBand silicon market. Mellanoxs main InfiniBand competitors include Fujitsu (TSE: 6702), Agilent Technologies (NYSE: A), and PathScale, a start-up. Its competitors also include Qlogic and Emulex in Fiber Channel and chip companies like Broadcom (Nasdaq: BRCM) and Marvell in Ethernet.

Important evidence of recovery in the field came in the acquisition by communications equipment giant Cisco Systems (Nasdaq: CSCO), a customer of Mellanox, of US start-up Topspin Communications early this year for $250 million. Topspin sells switches and cards for InfiniBand solutions in competition with Israeli company Voltaire, a promising start-up in its own right.

One thing is indisputable Mellanoxs chips lead the field. Although the company wont give its investors a phenomenal return, it will certainly dictate the pace in InfiniBand, and become one of the successful companies on the Israeli high-tech scene.

Considering acquisitions

"Globes": What makes Mellanox such a great promise?

Waldman: We have an amazing team in Israel, a successful architectonic solution, and high performance capability. We also have good relations with strategic customers, as reflected in our business and technological cooperation. The members of our board of directors, including Sequoia Capital partner Pierre Lamond, Raza Venture Management chairman and CEO Atiq Raza, Bessemer partner Rob Chandra, and USVP partner Irwin Federman are also making the company a success.

All the top-tier server manufacturers now make products with InfiniBand chips, and so do communications companies. We hope that storage producers will follow suit in 2006.

From a business standpoint, where will the company be in another year?

The company will go public at some point, but we intend to hold our IPO at the right time. I dont regard a flotation as a final goal, but as a milestone. We want to do it when we feel comfortable, because the aim of a flotation is to enable us to enter new markets in other words, to develop new technologies independently, or through acquisitions.

What acquisitions are you referring to?

Were looking at several companies, but I dont intend to disclose them.

What about in the other direction? Is there a chance that youll be bought?

A small chance. We think that the return on the investment will be higher in an IPO.

Do you think that one of these days, youll become a company worth billions of dollars?

Yes. In the future.

Where will you be a year from now with regard to technology?

Were currently developing the next generations that will provide faster data communications speeds, with less delay in data transmission between computers, and between storage components and servers.

What separates you from your competitors in the InfiniBand field?

Were the only ones with a solution that includes a host channel adapter and a switch. The other companies in the field have only part of the solution. Keep in mind that we compete with other technologies besides InfiniBand: Fiber Channel, which works mostly with storage, and Ethernet.

Youre also on the list of companies that have raised the most money. Couldnt you have succeeded with less money?

It was smart to raise a lot of money. It contributed to the company and our capabilities. Its hard for me to answer whether we could have done it with less money.

What was your biggest mistake?

Weve made many. We havent gotten storage companies EMC and Hitachi as customers yet, but theyre on the way, too.

What was your biggest success?

We have customers like IBM, Cisco, NetApp, Dell, Silver Storm, and Voltaire. Our ten largest customers account for 80% of our revenue.

Published by Globes [online] - - on September 15, 2005

Twitter Facebook Linkedin RSS Newsletters Israel Business Conference 2018