This has been an amazing year for Inphi (IPHI), as the market has woken up to not only the company’s strength in physical layer technologies for the data center and telecom markets but also its ability to execute on those technical capabilities. Looking into 2020, Inphi has the chance to leverage 400G ZR, 200G/400G PAM4, and its M200 coherent DSP into even larger addressable markets, keeping the company on a trajectory to a served addressable market of $2 billion in 2022 against a likely 2019 revenue figure around $365 million.
Inphi’s qualities are certainly no longer overlooked, with the shares up over 130%. While I realize that growth/momentum investors will not be discouraged by valuation, it’s harder and harder to work the numbers to support such a robust valuation.
Moving From Strength To Strength
Whereas investors looked at most semiconductor earnings reports in the third quarter with a hopeful eye that the businesses were bottoming (despite guide-downs for the fourth quarter), there was no such concern with Inphi as the company posted another beat-and-raise, beating expectations by almost 5% at the top line and guiding fourth-quarter revenue about 5% higher than the prior sell-side average.
Inphi is going into 2020 with a head of steam and a strong portfolio of new drivers.
Back in November, the company debuted its Canopus 400G ZR 7nm DSP – the first to market with a complete solution that includes DSP, coherent driver, coherent amplifier, and silicon photonics. Although this will probably be more of a 2021 revenue opportunity, there will be commercial sales in 2020 and Inphi is beating Acacia (ACIA) to the punch.
I’ll be very curious to see the reception Canopus gets. Some industry participants, including Ciena (CIEN), which has business at risk to ZR, believe the opportunity has been overstated and that real-world use-cases will be smaller than expected. Then again, Ciena itself is planning to enter the market in 2020 and given the spiraling costs of data center equipment, the idea of replacing a rack unit with a pluggable and reducing costs for 400G interconnect seems at least superficially appealing. Longer-term, I expect Canopus to have relevance in markets outside of the data center interconnect, and I’ll be interested to see how Ciena responds.
The intriguing opportunity is the ramp of 200G/400G PAM4 solutions, with Amazon (AMZN) expected to ramp up with Inphi fairly early in 2020 (at 400G) and later Facebook (FB) and Alibaba, while Google has been ramping with 200G. Although there have been concerns that Amazon and Microsoft (MSFT) would be delaying 400G upgrades until later in 2020 (or even 2021), Inphi management has had multiple opportunities in the last month or two to echo that warning and they haven’t, with the company guiding for a ramp of “1.5” PAM4 customers in 2019 to “3.5” in 2020. Broadcom (AVGO) remains around six months behind Inphi, and although MaxLinear (MXL) is getting some business with Amazon, it doesn’t seem to be much of a threat to Inphi – management said at a recent conference that they’re “not seeing competition”. Management believes they’ll have about 50% share of the PAM4 market in 2020 (an addressable opportunity around $320 million).
Last and not least is expected ramps of the M200 coherent DSP for metro applications with customers like Nokia (NOK) and ZTE. This isn’t a huge market (an addressable market size of maybe $200 million to $250 million), but it’s big enough to be worthwhile to Inphi.
Using M&A To Shore Up Capabilities
Back in mid-November, Inphi made an interesting acquisition that I believe will prove to be significant (but probably invisible) down the line. The company agreed to pay $216 million for most of eSilicon (with Synopsys (SNPS) buying the rest). eSilicon has thus far been a custom silicon designer, using its own IP in SerDes, advanced packaging (including FinFET), and DSP and combining it with customer IP to design custom ASICs.
There’s a lot to unpack here, but I think the gist of it is this – Inphi realized it needed to upgrade its 5nm/7nm design and advanced packaging capabilities to keep its product roadmap on track. I know Inphi management has commented that customer demand for 2.5D packaging has exceeded their capabilities, and I likewise know that many larger companies have struggled with the challenges and demands of design at the 5nm/7nm level. On top of that, I think Inphi can make use of eSilicon’s capabilities in multi-chip modules (or MDM) and leverage some technology in the relatively short term to extend its PAM4 opportunities in the data center to the <7m segment.
Not only will eSilicon cost Inphi some cash, but it will also suppress gross margins – quite possibly on a long-term basis. I believe Inphi can recapture most, if not all, of that through more efficient R&D spending and improved long-term operating leverage (more revenue). At worst, I think this is an “it takes money to make money” deal that helps secure Inphi’s ability to go toe-to-toe with the likes of Broadcom beyond this current/upcoming generation of physical layer products.
With strong momentum exiting the third quarter and an attractive roadmap for 2020/2021, I have no problem boosting my modeling expectations, and I think Inphi is looking at at least two years of 20%+ growth, and maybe three (and maybe more than 25% growth). I likewise don’t believe the long-term margin consequences from eSilicon are all that bad. With that, I think Inphi is now looking at long-term revenue growth annualizing in the mid-to-high teens, with future adjusted FCF margins in the 20%s.
The problem is the valuation, as the Street is amply rewarding Inphi for its highly differentiated growth opportunity and product ramps over the next two to three years. None of my normal approaches work in terms of valuation, though again I’ll note that investors routinely pay up for differentiated growth coming out of a sector cyclical downturn.
The Bottom Line
If you want to continue riding the momentum at Inphi, I wish you luck – that’s just not my investing style, though I know from experience that it’s a lot of fun to find a stock like this in the early days and see where the market’s enthusiasm will carry it. I will sound a note or two of caution on the risk of data center spending delays and competitive market entries, and likewise, it’s worth noting that if Inphi stumbles, the stock will likely have an outsized reaction. While this isn’t a name for me at this valuation, I think Inphi is set to do very well in 2020 and it’s a name I’m following in the hopes of taking advantage of a future stumble.
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.