Harmonic Inc (HLIT - $1.5bn market cap / EV) is going through two exciting transitions over the coming year (the “fundamental” thesis ….and the “event” thesis).  First, they are selling the weaker, profitless segment for ~20-30% of the market cap (“the event thesis”) which they can then use to repurchase shares and retire essentially all of their existing debt outstanding.  Second, and more importantly, the other business (“Broadband”) is transitioning from commodity-ish hardware to monopoly software/appliance (“the fundamental thesis”).  Lastly, 2024 numbers have been reset and we now have a reasonable bar/estimate revision path ahead of us. 

You may think “monopoly software” is flashy bluster (so did I), but they literally already won every customer with the scale to support a competitive offering.  It’s too late, go home…but fortunately it hasn’t really shown up in the numbers yet as this is a complicated, multi-year roll out.  The unclear timing around that roll-out and messy financials due to the segment that is actively being sold (in response to in-bound interest) creates the opportunity.  I believe HLIT has a path to $1.50-2.50 of FCF/sh in the next 2-3 years.  I will let the reader pick a multiple, but anything north of 5-7x unlevered FCF for monopoly software / appliances drives significant upside.  I will offer you 20x $2 for a 3-year 3.5x.   

Finally, I believe there’s significant downside protection as the company just redeemed their convert (reducing leverage to <.5x) and there are now two activists involved.  The company also recently materially upgraded their CFO, likely at the behest of one of the activists. 

Harmonic has long been an off the radar cable equipment company (think Arris/Commscope and try to hold your vomit for a second).  This situation led management, many years ago, to invest heavily in “virtualization” capabilities, to leapfrog their competitive position.  Shockingly, it worked.  HLIT’s Cable OS is now the backbone software for network virtualization for cable providers as they transition to DOCSIS 4.0. 

You can actually just pull up a Comcast ($200 bn market cap) investor event (RBC conference on 11/15/2022) and see they brought their CTO on for literally an hour not to talk about parks or linear media or EBITDA or FCF….but about the importance of their network modernization for which they are HLIT’s anchor customer.  Importantly, over the last year, despite HLIT’s stock price getting cut in half, the most important thing happened…..Charter, who was flirting with an alternate route, capitulated and picked HLIT too.  There is now no remaining customer available that is large enough for a competitor to survive, who would be 5 years behind anyways.  All the others have essentially given up.  

Most US households get broadband home internet via cable networks (IE Comcast, Charter, etc) built during the 80's with an esoteric design owing to their origin to deliver analog television signal.  Like many other telecom systems (enterprise data centers, storage, wireless comms), these networks are moving from an architecture ("CMTS") based on highly customized giant pieces of hardware in large central facilities (“appliances”, “head-ends”, or “CCAPs” in this case) to using simpler, more modern hardware, located close to the user (“the edge”) with the key functionality moved into the software. 

This change is called “virtualization” ("vCMTS") or “Distributed Access Architecture” (DAA).  To put some more technical words on this, the traditional CCAP or headend is a giant physical warehouse with racks of equipment whereas a Remote-Phy (the hardware in virtualization) is about the size of a carry-on suitcase and hung on a telephone pole near the user. 

Why bother with this architecture shift?  It drives not only lower operating costs for the network operator, but higher performance that they can use to retain their customers in the face of competition from fiber and fixed wireless. 

 

It also saves energy consumption (less pieces of giant hardware) and increases network reliability.  For example, in a traditional network a single point of failure could knock out 20-30k customers whereas in the new paradigm, a single node  could knock out 40 customers.  Additionally, it enables scalability – the ability to add capacity and functionality without rolling a truck out, digging into the ground, or adding additional hardware. The reduced hardware burden can result in 70-80% reductions in power consumption, supporting operators ESG goals as well. 

 

While this transition is long, slow, complicated and dates all the way back to 2016, we’re at the point where all the competitive alternatives has failed, leaving Harmonic as the dominant (read: ONLY) provider of the software, which is the critical glue enabling this transition.   I can’t emphasize enough the importance of virtualization to the operators.  Comcast, with a $200bn+ EV, dedicated an entire investor event (RBC TMT conference on 11/15/22) to their CTO talking about the benefits of virtualization.  Comcast is the anchor customer for Cable OS and represents >35% of HLIT’s LTM revenue. 

To understand how tiny Harmonic ended up in this position, we need to understand history.  Prior to 2016, Arris was dominating on the hardware side, Harmonic was nearly a generation behind, and they made the “life or death” decision to prioritize winning the next generation while Arris monetized the current generation.  Arris ultimately was acquired by Commscope, who is now choking on $10bn of debt related to that deal (nearly 8x leverage) and falling further and further behind.  Interestingly, the combination of being the only virtualization software provider and competing against incumbents who are either in financial distress (COMM/Arris) or de-prioritizing the segment (CSCO), is actually allowing HLIT to win a disproportionate share of the hardware sales as well. 

To date, cable networks representing ~100mm global subscribers (out of a total market ex China of ~180mm) have signed on with Harmonic.  The last major shoe to fall was Charter, who was pursuing a competitive solution that ultimately failed and signed on with Harmonic.  Without Charter available as an anchor customer, it’s not really possible for any competitive solution to have a large enough market opportunity/interim financial backing to succeed.  It’s important, however, to understand, that HLIT primarily gets paid not when the operator chooses to use Harmonic, but when they actually roll out their virtualized network.  A significant component of Harmonic’s profits are related to the hardware – this dates back to historical industry purchasing patterns and as much as the financial engineers in us want it to all be 90% margin, recurring SaaS revenue, I’d rather have a dominant position in a big market with a sub-optimal business model (license+hardware instead of pure SaaS), than be competing my face off against larger incumbents.  That said, the growing importance of software in the offering likely does increase stickiness.  It’s much easier to rip and replace a piece of hardware in a network if you find one 10% cheaper than to re-do all of the software. 

 

 

HLIT’s video segment (~$200mm revenue, ~breakeven EBITDA) is less core to the fundamental thesis, but an exciting kicker for the stock.  Historically, this business has been appliances that enable content programmers (IE HBO, CNN, ESPN, etc) to compress, transcode, and transmit them via satellite or fiber to the TV provider.  Streaming and the cloud have disrupted this process, but also created opportunities. 

 

As with the broadband business, HLIT’s management has been at the cutting edge, technology-wise (the CEO is indeed a phd engineer).  Their VOS product is a cloud-native video compression/transcoding, ad insertion, and packaging tool system.  This is not dissimilar to the transition the broadband business is making, but there is less economic asymmetry to it and there are actually competitors on the software side.  The feedback we get is that VOS offering stands out for live sports – powering many of the most viewed live streaming events in recent history. 

 

Historically, this business had significant customer overlap with the Broadband business.  However, today there is very little operational overlap as video and broadband infrastructures have veered apart.  In response to in-bound inquiries, HLIT announced they were going to run a full sale process in October of 2023. 

 

We believe the SaaS component will do ~$70mmm of revenue in 2024 (+~30% YoY after growing 47% in 2023) and the appliance business will do ~$135mm of revenue.  A .5x-1x multiple of revenue on the appliance business and 3-4x on the SaaS business gets you to $250-400mm or 20-30% of the market cap less some modest taxes.  We think those are likely conservative numbers and only assume ~$100mm of share repurchases with the proceeds.  

 

 

The company is not very publicly forthcoming about the economics of their various products because they’re in active negotiations with large buyers who all pay very different prices based on size and the company should not be bidding against itself. 

 

We estimate that for the ~60mm non-Comcast remaining homes that have already committed to deploy CableOS with Harmonic, the company will make ~$2k of software revenue and a 50% share of $2-5k of hardware per node.  One node is per ~100 homes.  That ratio will likely split to 50 over time to upgrade capacity, which will double this opportunity down the road.  However, assuming 100 homes per node in the beginning works out to ~$2-3k of profit per each of 600k remaining nodes…a $1.5bn+ profit opportunity over the coming 5-10 years (~$200mm per year on average).  For Comcast, they'll get ~$40mm of software revenue per year plus ~$1k+ of profits across 20k remaining nodes.  

 

That gets to ~$250mm of EBITDA for broadband is a hair above the ~$232mm the prior CFO estimated at an investor day in the past before they won Charter.  The new CFO has not confirmed or reiterated these targets and at a minimum they have been pushed out.  Beyond that $250mm, there’s another 80mm homes that will likely be built but haven’t been officially won yet – those would imply another $100-200mm of annual EBITDA.  You will also have upgrades / node splitting over time as bandwidth needs grow but that’s likely beyond the timeline of this investment.  There’s also a hybrid fiber opportunity as well, which would just be gravy.  

 

For simplicity’s sake, as long as you get to ~$250mm of EBITDA, you’re over $2 of FCF per share and in some cases dramatically higher (depending on what you do with the balance sheet, interim FCF, and the proceeds of the video sale).  I won’t give you a pencil-precise estimate here, but any “confidence interval” arrives in an extremely attractive territory relative to today’s stock price.  

 

Harmonic has two activist shareholders in Scopia and Ancora.  Scopia has been involved in several years and previously had board representation.  I believe these activists likely led to the CFO change, which most investors have talked to have described as a “day and night” upgrade.  I suspect the video sale process is likely also related to the activist involvement. 

 

Additionally, I believe that Cable OS has massive strategic value given its small but critical nature for many very large companies.  Given the small check size, lack of debt, and strategic value, we believe we're well protected.  We also suspect many strategics would prefer to own the company than allow it into others' hands.  There have been rumors in the past about Comcast acquiring it.  I’d also note that digital infrastructure PE firm Digital Bridge owns a stake in the company, though presumably that is a part of their long/short strategy.

Risks 

There are obviously many risks here.  The roll-outs of virtualization and DOCSIS 4.0 have taken a long time, are extremely complicated, and have been delayed on multiple occasions.  There are some very large customers who do have significant negotiating leverage with Harmonic.  Network architecture and technology are constantly evolving, so Harmonic could be leap-frogged in the next iteration just as they leapfrogged into this one (I think that is a decade+ away though).  The sale of the video business could fall through.