Shoot first. Ask questions later.
By now, we all realize that the market did not fancy what it heard from Infinera, and it turned and started shooting. I was grazed, but I’m pretty sure my wound will heal; my portfolio is diversified. Time to start asking questions. I think the most important questions will be those surrounding expenses and why so few pennies are falling to the bottom line. Many of my prior write-ups have focused more on the company’s “story”, which I still find intriguing. But I think it is time to focus a bit on the business model. 1Q16 results seem fine. GAAP earnings were light, but in line with expectations. 2Q16 guidance seems disappointing. Management seems to be laying the blame on one large unnamed Transmode customer whose purchases have dried up of late. Should we take this at face value, or is it a sign of deeper problems? 2Q16 guidance is for a 23% year-over-year (YoY) revenue increase. That doesn’t sound bad on the surface but, in reality, it is mid-single-digit organic growth if we exclude the expected bump in revenue from Transmode. Is Infinera’s business slowing? Is the Transmode acquisition failing? There’s a lot to explore here, and this will be a long post. But let’s start with the headline news…
Earnings Report Highlights
The earnings press release can be found here: http://investors.infinera.com/new-releases/press-release-det… Seeking Alpha’s transcript of the conference call can be found here: http://seekingalpha.com/article/3968845-infinera-infn-thomas… (Thanks, Seeking Alpha.) Unless stated otherwise, all italicized quotations are from the Seeking Alpha transcript.
1Q16 Revenue: $245 million This is 31% higher than 1Q15. Based on Transmode’s size prior to the acquisition, it is probably fair to assume that half this growth is organic and half due to acquisition. Management’s guidance had been $240-250 million, so they hit their number cleanly. Wall Street expected $246.1 million. So they missed expectations slightly.
1Q16 Gross Margin: 47.5% GAAP and 50.2% non-GAAP Management had guided for non-GAAP gross margins of 47.5-49.5%, so they beat their guidance. This result is above their long-stated intermediate goal of 50% non-GAAP gross margins and they were quick to point out that this level of gross margin is unsustainable at present (although it might be soon), and “blamed” the high number on quarterly seasonality and product mix. Generally, the year’s first quarter is a time when market participants are still finalizing yearly budgets and revenues are slower as only urgent needs gain immediate approval. For Infinera, projects that increase network footprint are lower-margin but offer longer-term growth (and are often deemed less urgent by customers still finalizing budgets). Projects that increase bandwidth within existing footprint, on the other hand, are often higher-margin and might be urgent enough to sail through customers’ approval processes regardless of budgetary finality, due to immediate need. In the first quarter, Infinera’s seasonality skews towards the latter type of project, and hence, product mix.
1Q16 Operating Margin: 6.1% GAAP and 12.3% non-GAAP Guidance had been for non-GAAP operating margin of 10-12%, so management beat the top end of guidance. The gap (if you will) between GAAP and non-GAAP numbers has increased, as non-GAAP numbers ignore (rightly or wrongly) certain acquisition costs.
1Q16 Net Income: GAAP $12.0 million ($0.08 per diluted share) and non-GAAP $28.0 million ($0.19 per diluted share) Management’s guidance for non-GAAP earnings per share was $0.15-0.19. Wall Street’s expectations were $0.17 for non-GAAP earnings.
1Q16 Cash Flow From Operations (CFFO): $10.0 million This is the second-worst quarterly CFFO showing in twelve quarters. CapEx was $10.8 million, so free cash flow (FCF) was negative, only the second time that’s happened in the aforementioned time span.
Revenue: $250-260 million (vs. $207.3 million in 2Q15)
Non-GAAP Gross Margin: 47-49% (vs. 47.4% in 2Q15)
Non-GAAP Operating Expenses: $93-97 million (vs. $80 million in 2Q15)
Non-GAAP operating margin: 10-12% (vs. 13.0% in 2Q15)
Non-GAAP earnings per share: $0.15-$0.19 (vs. $0.18 in 2Q15)
Wall Street had previously expected 2Q16 revenues of $272 million and $0.22 in non-GAAP earnings. This is a big negative surprise.
INFN earnings day share price: $12.01 -22.90% (vs. S&P 500 -0.92%) The light second quarter guidance was probably the biggest driver of the sell-off. Negative FCF may have been a source of angst as well. I have to admit I was expecting better results, and I added to my INFN position in the week before earnings. I’m as disappointed as any of you.
The count of Cloud Xpress customers increased from 20 to 25 during the quarter.
There were three 10% customers for the quarter: a cable operator; a wholesale and enterprise carrier; and an Internet Content Provider (ICP). According to the annual report on SEC form 10-K, for the full year 2015, two 10% customers accounted for 17% and 13% of revenues (no other details were disclosed).
Chief Executive Officer (CEO) Tom Fallon noted that the conversion from 10Gb to 100Gb is mostly complete in the long haul market. However, he was quick to add that increasing bandwidth consumption presented an ongoing revenue opportunity for Infinera. In fact, he described Q1 as the “best bookings quarter we’ve had in a couple of years.” This may drive margins higher for Infinera in this market because, as I mentioned before, adding bandwidth yields higher margin than does new deployments.
Data Center Interconnect
There were two key points that CEO Fallon wanted to make about this market. First, he indicated that this was Infinera’s best quarter yet for DCI revenue, as well as a record bookings quarter.
Second, he wanted to clarify his view of the PAM-4 news. This is the technology that Microsoft (working with Inphi) highlighted at the OFC conference as an alternative to the kind of technology that Cloud Xpress (CX) provides. He sees it as a complementary, rather than competitive, technology. It is only applicable in short-reach situations and, due to its “pluggable” nature, entails higher operational complexity than products such as the CX require. Jokingly, he dismissed it as “a flash in the PAM”. During the Q&A portion of the conference call, Infinera President and co-founder Dave Welch confirmed that Infinera had no plans to develop PAM-4-based products. I see that as a good thing, because it is my impression that (1) the market is relatively small, (2) the equipment is very commodity-oriented and low-margin, and (3) the equipment is not amenable to any of Infinera’s strengths. To expand on the third point, the benefits of features like Instant Bandwidth and Software-Defined Networking would be impossible to gain (or would at least be limited), as network changes require manual manipulation of the connections between the transport equipment and the switches. In the short-term it may be a cost-saving solution for Microsoft, but it doesn’t seem well-aligned with where the networking industry is headed (that is, more automated control).
Metro Core, Aggregation, and Access
Infinera describes its experience for the quarter in metro – its newest market – as “mixed”. Citing positives, CEO Fallon noted the recent win with Telia (https://www.infinera.com/telia-carrier-low-latency-us-networ…) and other wins in mobile front-haul (e.g., https://www.infinera.com/platinum-core-solutions-first-in-ma…). Front-haul, as I understand it, refers to the equipment between cell towers (or other aggregators of wireless signals, like small cells) and the centralized equipment that directly feeds the long-haul network. Front-haul is getting increasing attention due to burgeoning demand for data from smartphones, as well as an increasing number of wireless devices generating data as the Internet of Things gains traction. The new architecture is called C-RAN or Cloud – Radio Access Network. This architecture will become increasingly critical when 5G networks are rolled out.
On the negative side of the ledger, the revenue synergies expected from the Transmode acquisition are taking longer to materialize than some anticipated. I don’t believe Infinera is surprised, since they’ve experienced customers’ product qualification cycles before. I think it is Wall Street that is itching for the cross-selling opportunities to come to fruition, boosting near-term revenue estimates. Further, it is certainly not helping that an important Transmode customer has not purchased any equipment in a couple of quarters. CEO Fallon doesn’t believe that Infinera has lost the customer to competitors. He is also aware of a deal that this customer is looking to close. If the customer can obtain that business, it is very likely that they would need additional equipment. However, given this customer’s recent history with Infinera and uncertainty surrounding when their deal will close, Infinera did not assume any business from that customer when crafting its guidance. I’ll return to the issue of Transmode’s customer relationships later.
Another important point in the Metro space is that Infinera currently has very little market share, especially domestically. Winning market share means unseating incumbents. This roll-out is more likely to resemble the DTN roll-out, which also faced entrenched competition. With DTN-X, Infinera already had significant DTN market share, making the transition easier for Infinera customers. With Cloud Xpress, Infinera’s box is first-in-class and best-in-class (at least so far). The market is new enough that, where there are incumbents, the box Infinera would be looking to displace is not purpose-built. Front haul, while a new market in a sense, will probably have a competitive profile more like Metro than DCI. While there may not be existing infrastructure to rip out, the front haul providers probably all have existing relationships with one or more metro equipment providers.
How is Infinera Managing its Operating Expenses?
Since Infinera’s price decline after earnings, I’ve noted some concern on the boards about Infinera’s spending habits and whether operating expenses have run amok. I don’t think they have. I think they’ve generally been pretty consistent as a percentage of revenues, and I’ll present that case to you. The only way that operating margins are going to improve (assuming gross margin stays constant) is for operating expenses to increase at a slower rate than earnings (or they could decrease, but that’s probably asking too much at a company that’s growing). We’ve already been told that Research & Development expenses will grow apace with revenues. That means declining rates of growth must come from Sales & Marketing and from General & Administrative. Please note that I’m using GAAP numbers, so stock-based compensation is already factored in. Let’s explore what we’re actually seeing. Since the Transmode deal had an impact on Infinera’s cost structure, I’m going to look at three periods: 2014, 2015 pre-Transmode, and post-Transmode.
• In 2014, Sales & Marketing averaged 11.8% of revenue; range: 10.9%-12.5%.
• In early 2015, Sales & Marketing averaged 10.8%; range: 10.4%-11.3%. The direction is favorable…
• Post-Transmode, Sales & Marketing averaged 12.7%, registering 13.1% in 4Q15 and 12.3% in 1Q16.
The averaged numbers all fall within a couple of percent of each other. They were improving prior to the Transmode acquisition, backslid immediately after the acquisition, and seem to be improving since then, although we don’t have many data points. In a good year, 4Q is probably a seasonal high point, as some sales people generally get paid higher commissions when they’re above quota. I have little insight into Infinera’s compensation structure, but this seems logical to me.
• In 2014, General & Administrative averaged 7.3% of revenue; range: 6.4%-8.6%
• In early 2015, General & Administrative averaged 7.4%; range: 6.8-7.9%
• Post-Transmode, General & Administrative averaged 6.5%, registering 5.9% in 4Q15 and 7.1% in 1Q16
The trend here seems favorable, especially if one notes that 1Q tends to be the highest quarter in any given year, due to year-end bonuses.
Why did Infinera only report $0.08 GAAP diluted earnings in 1Q16 then, despite 50% gross margins? Fair question. I think it’s a combination of factors. Between 1Q14 and 4Q15, R&D spending averaged 20.2% of revenue with a range of 19.2% to 21%. In 1Q16, R&D was 22% of revenues. I think I know why the spending was so high this particular quarter, and I’ll touch on that in the next section. So that took about 2 percentage points out of operating margin. The post-Transmode Sales & Marketing costs are inflated compared to pre-Transmode, so that probably took another 1 percentage point out of operating margin. Finally, to get to the question of “Why so few pennies?”. Please recall that the purchase of Transmode was effected with both cash and stock. Between June ’15 (pre-acquisition) and December ’15 (the first quarter where the newly-issued shares were fully counted), the share count rose by about 10 million shares, or nearly 8%. In the EPS calculation, an increasing denominator doesn’t help when increased expenses also impact the numerator.
Infinera Has Been Talking a Lot about “Investing”. What is it Investing In?
Last quarter, I noted that Infinera management used the terms “invest” and “investment” eleven times in their prepared remarks (this count specifically excludes customer investments and only includes investments that Infinera is making). The corresponding count this quarter was nine. Infinera is clearly not hiding the fact that it is investing, but – other than a vague “its future” – what exactly is Infinera investing in? This section will include both some hard numbers and some hypotheses.
People – After the Transmode acquisition closed, company headcount jumped from 1598 to 1978. Two quarters later, it is 150 people higher still. It is clear that Infinera is investing in people. One aspect of this is that stock-based compensation is high. If you think that most of Infinera’s stock-based compensation goes to top executives, that hasn’t been true since 2009. Fewer than one dollar in six of stock-based compensation has gone to top executives in the 2010-2015 timeframe. Infinera resides in Silicon Valley. To attract top engineering and software talent, stock-based compensation is an important form of currency. With Software-Defined Networking becoming an increasingly-important part of the product set, Infinera competes for talent in the software space too.
Inventory – This one is intriguing. I think that four distinct reasons factor in to Infinera’s inventory build-up. First is the Transmode acquisition. Finished goods inventory bumped up 17% between the second and third quarter of 2015 – a huge increase by Infinera’s standards, but in line with each company’s revenue streams. The second factor is “time as a weapon”. CEO Fallon indicated that there would be a deliberate build-up in inventory to support speedy deployments at a time when supplier lead times were lengthening. The third factor is the increased number of products that Infinera is supporting. I went into excruciating detail about that a few quarters ago, and I won’t repeat that here except to say that the original DTN-X box has too much capacity for some customers, so smaller versions were created. Finally, most intriguing of all, is, I think, due the next generation products that have been announced but are not yet available for sale. I infer this because the current quarter’s inventory build-up is in “Raw Materials” and “Work in Process”, rather than “Finished Goods”. The build-up of inventory was one of the biggest drags on CFFO and FCF this quarter.
Other Random Musings
Transmode’s Large Customers
Here’s an interesting quote about Transmode: “We believe that it was mostly just due to two things: one, the economic challenge in Europe. As you know, Europe, in general, has not been spending in DWDM growth over the last couple of years. Second of all, it was a couple of fairly significant customers who had slowing buying patterns, based upon their own internal budgets and network needs. We have obviously done a reasonable amount of diligence with those customers, and believe that it was a temporary piece of time; there was not a loss of market share. There was not a loss of position within those customers. And we believe that those customers’ spending patterns should go back to more historic norms.”
CEO Fallon spoke those words more than a year ago – on the day the Transmode deal was first announced to the world. Given the issue with this quarter’s 2Q guidance regarding a large order from a Transmode customer that may not occur or may be delayed, it sounds as if the situation hasn’t improved as much as Infinera expected it would, eh? http://investors.infinera.com/financials/sec-filings/sec-fil…
Market reaction to this quarter (or at least the guidance) was awesome and fearsome. I think I still hear gunshots reverberating! I think we’ll all need to tamp down our expectations for near-term Metro market penetration. While vertical integration has its advantages, the overhead means that Infinera needs to achieve a relatively high level of sales – quarter-in and quarter-out – just to break even. A constantly increasing share count (even if it’s not going to top execs) means more dollars of profit are needed to add pennies to the bottom line. While Transmode may have brought a great team and great products to Infinera, it now seems reasonable to question their customer relationships.
Although I remain convinced of Infinera’s long-term growth prospects, at some point during Thursday’s trading, I think a ricochet shattered my rose-colored glasses.
Thanks and best wishes,
TMFDatabaseBob (long: INFN)
Peace on Earth