Technology Stocks : Mellanox Technologies, Ltd.
MLNX 83.40+0.8%5:20 PM EDTNews

 Public ReplyPrvt ReplyMark as Last ReadFilePrevious 10Next 10PreviousNext  
From: Jim Mullens12/21/2017 10:26:35 AM
1 Recommendation

Recommended By
The Ox

  Read Replies (1) of 940
Mellanox Knocked Around, But Definitely Not Knocked Out

Snip From Comments-

Regarding R&D spending, note Mellanox's disclosure on 12/7 that it killed the NPS line from EZchip and another "low ROIC" line to be disclosed in January.


Mellanox Knocked Around, But Definitely Not Knocked Out

PRO Pick

Dec. 19, 2017 8:27 AM ET
by: Stephen Simpson, CFA


Mellanox has seen a significant slowdown in its core InfiniBand business, as Intel has made inroads and important end-markets have slowed.

Ethernet is a meaningful growth opportunity for Mellanox, but it carries lower gross margins and multiple big-name competitors.

Upside in the shares today is largely tied to a potential M&A transaction, but a stronger re-acceleration of revenue growth certainly wouldn't hurt.

It’s not uncommon for elevated volatility and controversy to surround growth stocks, but Mellanox ( MLNX) seems to get more than its share. It certainly doesn’t help that the company competes with heavy-hitters like Intel ( INTC) and Broadcom ( AVGO), nor has it helped that the company’s InfiniBand revenue (once the prime attraction of the story) has fallen off significantly. Add in elevated spending concerns, and it has been a bumpy ride for shareholders.

That bumpy ride has also been relatively productive for shareholders recently. Between a well-known activist shareholder taking a stake and management ratcheting down operating expenses, the shares have shot up this year and finally started outperforming the sector again.

I liked Mellanox back in May of this year, but as things sit today, I think most of the remaining value lies in the extent to which Mellanox attracts solid M&A attention and/or provides credible visibility to renewed InfiniBand growth. The stock price already assumes mid-teens growth in adjusted free cash flow, but a buyout bid would likely start in the high-$60's, if not the $70's.

InfiniBand Erosion Is A Real Issue

Mellanox is a more diversified company than it seems to get credit for (more on that in a moment…), but the fact remains that InfiniBand, the company’s high-speed, low-latency interconnect technology for high-performance computing, is an important part of the business.

It’s an important part that has not been performing especially well. Revenue has been lumpy, but moving down, as the last four quarters have seen 15% yoy growth (Q4’16), 14% yoy contraction (Q1’17), 3% yoy contraction (Q2’17), and 17% yoy contraction (Q3’17). That’s not the trend you want to see from a “growth” stock, and especially when that’s the higher-margin part of the business.

100G InfiniBand continues to grow as a part of the business, and management continues to maintain that Intel has had to buy its business for its rival standard Omni-Path. While that’s probably largely true, the majority of the Top 500 supercomputing sites run InfiniBand and InfiniBand’s new wins are well ahead of Intel’s Omni-Path, it is still having an impact. What’s more, softer end-market conditions in areas like storage (long an important market for Mellanox/InfiniBand) have done their damage as well, such that it looks like there will be little-to-no growth for InfiniBand in 2018 after a down year in 2017 (I’m assuming/estimating).

Longer term, the outlook is not so dire. 200G InfiniBand is on the horizon, and 400G will eventually follow that. Intel’s Omni-Path is “good enough” and good enough will be good enough for many application, but I believe InfiniBand will remain the go-to for high-performance applications. Specifically, ongoing cloud deployments, high-performance computing, and AI in particular should support robust long-term demand for this platform … but it won’t all come at once.

The Rise Of Ethernet Is A Mixed Blessing

As InfiniBand has been weak, Mellanox’s Ethernet business has picked up some of the slack. Ethernet (specifically RoCE or RDMA over Converged Ethernet) has been closing the gap with InfiniBand as a connectivity option; it’s not as fast, but it has gotten better and it is familiar to many enterprises. Importantly, while this is a competitive space (with Intel, Broadcom, Cavium ( CAVM) and others in place), Mellanox did really well with the move to 40G and should do well with 25/50/100G (I’ve seen estimates of Mellanox’s 25G range from 45% to 65%). Moreover, to the extent that customers want a strong competitor to balance Broadcom, that should help Mellanox’s efforts.

Management believes Ethernet could grow 100% or more next year, including ongoing growth from Spectrum switches (about 12% of the segment today). That’s great, but it comes at a cost – Mellanox’s Ethernet products have lower margins, and the increasing mix of Ethernet has been pushing GMs down, and will likely push them below 69% next year.

An Activist Has (Arguably) Rattled The Cage

Perhaps the most significant recent development with Mellanox has been the announcement that Starboard has taken a nearly 11% stake in the company and does not intend to be a quiet spectator. Not only has Starboard taken Mellanox to task for its elevated expenses (including very high R&D spending), but it also castigated the company for refusing to enter into M&A discussions with Marvell ( MRVL) when that company expressed interest (Marvell later made a bid for Cavium that was accepted).

Since Starboard came on board, there has already been some change. When management gave guidance for 2018 earlier this month, it did so with much more detail that it previously had offered. What’s more, while operating expenses had been trending in a better direction this year, management’s guidance for 2018 suggests a more significant improvement in expense control – possibly a year-over-year decline in expenses despite double-digit revenue growth.

Does Mellanox spend too much on R&D? We could debate that all day, but they will spend close to 40% of revenue (non-GAAP) this year, and that is quite a lot. The “too much” question is tricky, though, because Mellanox has in the past been able to parlay that R&D into industry-leading performance and innovation that has turned into revenue and market leadership. Even so, I certainly don’t doubt that they spend money on “speculative” projects (which I actually think is a good thing) and that a larger acquirer could come in and slash a lot of that spending.

The Opportunity

So, now what? I do believe the long-term opportunity in AI is real (the company is the connectivity partner for NVIDIA’s ( NVDA) DGX AI/deep learning platform) and I believe InfiniBand can and will still grow from here. I likewise believe there is real growth potential in the Ethernet business and in new offerings like BlueField (a high-end system on a chip similar to Cavium’s LiquidIO in some respects).

I expect mid-to-high single-digit long-term revenue growth from Mellanox, with a few more years of double-digit revenue growth left to go. I also expect margin leverage, and I’m glad that management is pulling some of that forward. I think 20%-plus non-GAAP operating margins are attainable in 2019, and I likewise believe that adjusted FCF margins can reach 20% in 2020/2021 and improve further, supporting long-term adjusted free cash flow growth in the mid-teens.

Discounted back, though, those free cash flows don’t offer much upside to today’s price. Turning to margins, revenue growth, and EV/revenue, I believe the double-digit revenue growth and improving margins support a forward revenue multiple now of 3.5x, or a mid-$60’s fair value. I also believe that a buyout price in the high-$60’s to low-$70’s is plausible, with the difference being the accretion that a buyer can extract from COGS, SG&A, and R&D that the company won’t be able to achieve on its own.

The Bottom Line

On its own merits, Mellanox looks like a hold to me. While I do believe that InfiniBand can grow again, it will take some time and the current price already does reflect ongoing above-average growth. Although I'm fairly confident that management could sell out in the $70s and the buyer would still see accretion, buying solely on the assumption of such a transaction is risky. All told, then, I think the move in Mellanox fairly reflects the present state of the business and its prospects.

Thank you for reading this Seeking Alpha PRO article. PRO members received early access to this article and get exclusive access to Seeking Alpha's best ideas. Sign up or learn more about PRO here.

Disclosure: I am/we are long AVGO.

I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.

>>> Comments >>>>>>>>>>>>>>>>>>>>

Bob Wheeler (TLG)

Regarding R&D spending, note Mellanox's disclosure on 12/7 that it killed the NPS line from EZchip and another "low ROIC" line to be disclosed in January.

20 Dec 2017, 11:41 AM

One that knows

To emphasize your points even further:

One should notice that Intel artificially withhold PCIe Gen4 for 7 years.

This is coming to an end with IBM releasing Power Gen9 servers that bring PCIE Gen4 and first DOE clusters are starting to pop up with great numbers

One should also notice that Intel is under pressure from other CPUs makers such as AMD, Cavium and Qualcomm which are finally bringing in very competitive server grade chips.

And, of course, their 10nm is delayed and delayed yet again not to the end of 2018.

Intel is going to pay the price for spending many resources on weak Qlogic stuff architecture.

All in all, Intel already started to shift its focus, during the latest SC the Omni Path was less present than ever before.
Report TOU ViolationShare This Post
 Public ReplyPrvt ReplyMark as Last ReadFilePrevious 10Next 10PreviousNext  

Copyright © 1995-2018 Knight Sac Media. All rights reserved.Stock quotes are delayed at least 15 minutes - See Terms of Use.