|Shares Out. (in M):||456||P/E||12.3||11.2|
|Market Cap (in $M):||112,622||P/FCF||0||0|
|Net Debt (in $M):||6,344||EBIT||0||0|
Let me start out with what this idea is not before even getting into the thesis itself.
Admittedly, I tend to search for (and prefer) ideas where I see a significant mis-pricing and thus an opportunity for outsized returns based on a fairly visible “monetization path” within a fairly visible timeframe (what some may call “contrarian” calls). Like I said, this is not one of those ideas. Rather, plain and simple, I think AVGO is a high-quality business with a strong mgmt team at a good price. A “long-term compounder” style idea, whose success is too often solely attributed to being a roll-up, which I think unfairly characterizes the true quality of the underlying business and mgmt team (i.e., their earnings growth/success isn’t only due to unsustainable financial engineering).
AVGO can be purchased today at an attractive valuation both on an absolute basis and relative to peers. While AVGO’s valuation has never been all that expensive, I find the combination of 1) undemanding absolute valuation; 2) excellent relative valuation; and 3) the prospect of rising rates taking a bite out of market valuations (thus favoring ideas that have both 1 and 2 above) providing a nice opportunity to own a quality name at a good price.
Company Description and History
(In their 2016 merger, Avago and Broadcom decided to compromise on company name / ticker by calling the combined co. Broadcom but using the ticker AVGO - just an FYI to avoid any potential reader confusion.)
AVGO is a mostly fabless semiconductor company, focused on all of digital, analog, and mixed signal devices, and diversified across multiple end-markets and product lines.
Unlike most “diversified” semis players, AVGO is not a broadline provider. In other words, as opposed to TXN, ADI, MCHP, MXIM and others, AVGO doesn’t make a gazillion different products, many at low-ASPs, that go into everything from toasters to supercomputers. While AVGO does have many products, the portfolio can be more appropriately characterized as several specific product “franchises” that are complex, have clear secular growth drivers, and face minimal competition (either due to high barriers to entry or a consolidated, duopolistic market niche, sometimes both). This differentiates AVGO in multiple ways: 1) AVGO is less cyclically-driven than other, diversified peers (the broadline guys in particular); 2) AVGO faces far less price / commoditization risk than other semis players; and 3) AVGO enjoys high(er) returns on capital.
AVGO has a long, interesting history that bears relevance to understanding the company today. AVGO was originally the semis division of HP and was originally focused on LEDs, fiber-optic transmitters, and optical mouse sensors. In 1999, HP spun off Agilent Technologies, which was mostly focused on test and measurement equipment (and also included what is today KEYS), but which also included HP’s semis division. In 2005, Agilent sold the semis division to private equity - led by Silver Lake and KKR - for $2.7B, making AVGO a pure-play semis company, albeit private, until IPO-ing the unit in 2009.
After acquiring AVGO from Agilent, Silver Lake / KKR installed Hock Tan as CEO. Tan came to AVGO from IDTI/ICSI, where he was chairman of IDTI after running ICSI (Pres and CEO) and selling the company to IDTI. Interestingly, Tan was never an R&D runt who rose the ranks, but rather a Harvard MBA who worked in financial-focused mgmt roles at Commodore, Pepsi, GM, and in venture capital before moving into the semis world.
Back to AVGO. Even before Hock was appointed CEO (in the few months between the Silver Lake acq closing in Dec 2005 and his appointment in April 2006), but certainly continuing under his leadership, AVGO has been very active in asset mgmt, not just on buying assets, but discarding those that didn’t meet specific parameters, regardless of their general quality (those parameters generally being high barriers to entry, minimal competition, strong market position, and complex/high-value-add). At the time of the Silver Lake deal, AVGO’s primary franchises were LEDs, optical components, and wireless filters (known as FBAR - chips that filter RF frequencies; the better the filter, the higher quality of wireless connectivity. These chips are a basic building block of wireless connectivity - particularly in smartphones - and increasingly important with every cellular technology advancement).
Following Tan’s arrival, AVGO made a few smaller moves from 2006-2012, before starting a string of meaningful acquisitions. In 2013, AVGO acquired CyOptics ($400M) and LSI ($6.6B), boosting AVGO’s networking franchise and re-starting the storage franchise (AVGO had previously exited storage in 2006, selling their prior storage business to PMC Sierra - now part of MSCC). In 2014, AVGO bought PLX for $300M (networking), and in 2015, AVGO took down ELX for $600M (storage). Finally, the Big Kahuna, AVGO merged w/ BRCM in a $37B deal in 2016, which was a leader in networking and connectivity. Since the BRCM deal, AVGO recently acquired the storage networking business of BRCD for $5.5B (though they jettisoned the ethernet and wireless networking pieces to EXTR and ARRS), in a deal that recently closed at the end of 2017.
What we are left with today is 4 primary franchises: wired (45% of rev; connectivity, networking), wireless (31% of rev; connectivity, mobile), storage (20% of rev), and industrials/other (5% of rev), with leading positions in high-end switching/routing ASICs/ASSPs (e.g., what ANET standardizes on), SoCs for cable/telco access equipment, wireless connectivity (bluetooth, WiFi, GPS, and wireless front-end/filters), server-storage connectivity, and HDD and SSD controllers, among others. More-or-less, high-end datacenter equipment is rife with AVGO products.
Financially, AVGO is a ~$19B company w/ industry-leading margins (GM% ~64%, EBITDA% in the high-40s, and net income margin in the low-40s) and a strong dividend (2.8%) with a >GDP LT rev growth profile.
Recent History / QCOM
Late in 2017, AVGO went hostile on QCOM, first offering $70/share ($60 cash / $10 in stock) and its now up to $82 ($60 cash / $22 in stock). I don’t want to re-hash the entire saga and certainly there is enough literature on it, given its become a public soap opera. A shareholder vote on AVGO’s nominated slate (for which Glass Lewis is in full favor of and ISS is in partial favor of) was supposed to be March 6, but has since been delayed 30 days or so by CFIUS looking into everything.
Bottom line is the following:
Let me start by saying that I don’t think NXPI - even if QCOM can close that deal - ever becomes a meaningful part of AVGO. Hock has implied as much and, though a good company, doesn’t fit AVGO’s acq criteria or mold (they are broad-line guys). Thus, I’m going to focus on only QCOM, not NXPI. If AVGO does take down the whole 9 yards, my guess is they divest a significant portion of the NXPI assets.
QCOM fits the AVGO mold quite well - strong secular drivers, duopolistic market niche, and clear market leader. Further, for the most part, I agree w/ AVGO’s assessment that QCOM has been managed far too liberally for far too long (likely due to the rich licensing cash flow stream) and could really use a new look. Other than the very public issues on the licensing front, QCOM’s chip biz does $17B in rev, good for 3rd-largest after INTC and AVGO (among non-memory guys), and yet carries segment EBT margin <20% and implied GM% of <47% (assuming licensing rev is 100% GM% and costs are all allocated to opex line).
Both metrics are troubling - application processors are pretty value-added devices, especially given that all smartphones run on them and smartphones have overtaken PCs as the connected device of choice. Secondly, QCOM has far and away the market-leading product - multiple media reports of AAPL throttling performance on INTC-based phones punctuated this point.
How are the GM% so low? I think its both due to mismanagement AND pricing games played in conjunction w/ the licensing biz. I think the argument of AVGO being able to negotiate lower licensing rates is a good one, not because QCOM can’t do it, but because AVGO would have a better negotiating position to jack prices on the processors. As a thought exercise - if QCOM slashed its licensing biz by 30%, to make up the difference, GM% on chips would have to be ~59%. That doesn’t seem all that ambitious. At a 50% cut, chip GM% would have to be ~69%, which even though pretty high, doesn’t seem all that crazy to me either.
More importantly, the opex intensity of QCOM’s chip business is ~29%:
By contrast, AVGO is at ~17% opex intensity. Obviously, the opportunity would be for AVGO to bring the chip biz opex intensity to AVGO levels (from a combo of synergies and scale), which should be entirely possible unless there is some structural factor of QCOM’s chip biz opex that makes it more intensive. While there could be a reason why QCOM’s chip biz is more opex intensive (certainly if there is funny business in the R&D accounting between the chip and licensing segments), anecdotally QCOM has long been known for being liberal on spending. Ultimately, that potential opportunity is quite large - closing that 12% gap on $17B of rev = $2B.
Of course the other aspect would be getting back all the lost licensing revs, or at least a portion of it, that is currently going unpaid by AAPL. Using the low-end of QCOM’s estimates ($1.50-2.25/share, not including catch-up payments) on what a future relationship would look like, would yield another $2.2B. Together, that’s $4.2B (it could potentially be even higher, plus catch-up payments), which would even make a $90/share (with $30 of that in stock) deal accretive, with accretion growing as debt is paid down.
This all says nothing of the value of QCOM long-term based on its growth prospects and the quality of its offerings. Bottom line, I think a deal for QCOM would add an entire other leg to the story from which to extract value over many years, but which the thesis is not dependent on.
Thesis / Investment Positives
In order to properly set the stage for the thesis, I wanted to put the comp table upfront, given my view is more-or-less predicated on the idea that AVGO is one of the highest quality companies in semis (and in general), making the valuation compelling.
Variant Perception / Bear Case / Things I Don’t Like
Estimates / Valuation
This is a tricky one too, given the style of the idea. I tend to look at these types of idea through the lense of multiple more than anything else. Meaning, instead of trying to call a given earnings number (and thus significant near/medium-term price appreciation as that number becomes reality), I will own as long as the value prop is good at current valuations. Put another way, as long as I feel good about AVGO’s ability to grow rev ~5% and earnings/FCF ~10% annually, an ~8% earnings yield and ~3% div yield look good to me, particularly in light of the peers and the market.
Thus, valuation is a moving target - at a discount or in-line with both peers and the market, I’m staying put, possibly buying more; at a premium to both I may trim; at a large premium to both I may sell. It all depends on where things are in both the semis and broader markets.
Ultimately, I can make the case that Street numbers are too low and that EPS will be >$20/share this year (w/o QCOM), with 5-year earnings power >$25/share, but I don’t own it because I think everyone is so way off, so its not my primary focus. Put another way, the crux of this idea isn’t that the Street is way off.