Description

ViaSat sells satellite-based communication services – mostly plain-vanilla broadband internet access – and related equipment.  It is a bad business with poor returns on capital, a promotional management team, operational results that have disappointed investors for years, and an overvalued stock.  We bought ViaSat’s stock 2010 but sold in 2011 when we became disillusioned with management’s over-optimistic predictions of future benefits that they claimed would flow from their new satellite launch.  We are now short ViaSat because of . . . management’s over-optimistic predictions of future benefits that they claim will flow from their new satellite launch.  They will likely disappoint expectations badly.  Although that disappointment will probably unfold slowly, we believe now is still a good entry point for a short, largely because of this one-year stock chart of ViaSat, Inmarsat, Eutelsat, and SES, which I will explain more below:

This is my second consecutive write-up of a telecom short based on a generic thesis that, in telecom, future competition usually exceeds what management expects (or at least claims to expect).  The last write-up was short-WIFI in August 2015.  What I said then also applies now:

This position is likely a longer-term short.  I am posting a recommendation now because it looks like a great entry point. . . .  My thesis depends less on the specific details of what is happening now – no bearish channel checks, no looming specific problem, no “they are going to miss the next quarter” – and more on the structural analysis.  I can’t tell you what the catalyst will be, only that odds are excellent that there will be one.  I have 17 years of experience as a strategy and operations consultant for telecom and technology companies and then as a tech/telecom-focused hedge fund manager.  I have seen this movie before.

My exaggerated and knowingly imprecise description of the average telecom business, from the owner’s perspective, is this:  Dig a hole in the ground, throw tens or hundreds or thousands of millions of dollars into the hole (capex), and pray.  Pray that you can earn back your capital cost before either the technology you just bought becomes obsolete or new entrants pile in and destroy the margins.  In the old days, people literally dug holes in the ground to hold the wires.  Now you might be launching the money into the sky (satellites), hanging it on towers (cellular), hiding it in corners or behind stadium seats (wifi), etc.  To put it in academic terms, telecom businesses tend to have high up-front capital requirements, high technology risk, and low barriers to entry. The big boys like VZ and T have enough structural advantages, technology diversification, government help, and cash flows such that their risks end up being fairly low.  For almost everyone else, owning the equity of a telecom business, on average, sucks. 

Instead of being a longer-term position, WIFI returned 25% in three weeks, which was 21% alpha.  That particular outcome was lucky.  But when the structural situation, valuation, chart, and news flow all line up in your favor, your odds of getting some good luck go way up.  I think that is true with ViaSat now.

THE BUSINESS, FINANCIALLY

Quality businesses have high returns on invested capital.  ViaSat’s ROIC for the last five years, even excluding intangible assets from invested capital, has been 0%, -1%, 0%, 4%, and 3%.  This is a crappy business.  You would never know it from management’s press releases and calls, because they only talk about revenues and “adjusted EBITDA.”  ViaSat has been growing those numbers each year, but it is nearly profitless, nearly worthless revenue and EBITDA growth.

ViaSat has a $4.5 billion enterprise value, which is supported by:

• Two old and nearly obsolete satellites that it acquired in 2009, which now handle only 10% of its traffic.

• One newer operational satellite, called ViaSat-1.  It launched in October 2011 for a total cost of $500 million.

• A coming satellite, ViaSat-2, which will be launched in early 2017.  ViaSat-2 will have 2.5x the capacity of ViaSat-1, 350 gigabits/sec versus 140 for ViaSat-1.  Total costs are estimated at $650m, much of that already spent but with almost a year’s worth of spending to go before ViaSat-2 goes live in mid-2017.

• The promise of three more satellites, all called ViaSat-3, the first of which will (maybe) be launched in late 2019 or 2020.  (Satellite launch dates are often pushed back; ViaSat-2’s launch has been pushed back by 6-9 months so far.)

• An equipment-selling business.

In terms of profits, the $4.5 billion EV is supported, sort-of, by $325m of forward-year EBITDA (FY17 with a March year-end).  I say “sort-of” because the real EBITDA is not even that strong; they capitalize and then depreciate a lot of their equipment costs in the systems they sell.  Their FY17 EBIT will be about $55m, only one-sixth as much as EBITDA.  EBIT margin is set to grow from roughly 0% in FY12-FY14 to a still-pathetic 3.6% in FY17.  After interest and taxes, they will report around $25m of net income, 0.49 per share.

Sometimes for a capex-intensive business, looking at EBIT or EPS is unfair for valuation purposes because the capex bulge is in the past and D&A runs well above capex.  Not here.  ViaSat’s capex has been well above D&A every year for the last eight years (FY09-FY16) and will be again in FY17 and probably in FY18 and FY19.  The last five years have seen capex – including the purchase of “patents, licenses, and other assets,” which they make in size every year – of $229m, $202m, $352m, $419m, and $451m.  The company’s free cash flow has been strongly negative in each of those years.  Over five years, the combined free cash flow (even ignoring stock comp) has been -$567m, and net debt has ballooned from $295m to $882m.  On top of that, they have issued $169m worth of stock-based compensation, at a rapidly increasing pace.  In FY15 and FY16, revenues grew 2.3% and 2.5%, but stock comp grew 17% and 21%, to $48m in FY16.

To recap, this relatively mature, middling-growth business has the following FY17 valuation multiples:

• 13.9x EV/EBITDA

• 83x EV/EBIT

• 152x EPS

• Negative FCF multiple because FCF is strongly negative

THE BUSINESS, OPERATIONALLY

ViaSat reports three business segments, all of which provide satellite communications services and equipment to various customers:

• Satellite Services – 39% of FY16 revenues, 12% growth in FY16, 24% CAGR last three years

• Government Systems – 43% of FY16 revenues, 13% growth in FY16, 0% CAGR last three years (it shrank in FY14 and FY15)

• Commercial Services – 18% of FY16 revenues, -28% growth in FY16, -19% CAGR last three years

The best segment is Satellite Services, which itself has four sub-parts.  The largest part is selling broadband under the Exede brand name to consumers, largely those who can’t get their broadband from a landline (fiber, cable, or DSL).  That business has been an enormous disappointment (more on that later) and is stalled out due to ViaSat-1’s capacity constraints; it grew subscribers by only 1.6% over the past year.   Based on average subscribers and average ARPU, this business generated roughly $470m of the total $559m Satellite Services revenues in FY16.

The newer, smaller (around $70m revenues in FY16), but more growthy and sexy business is providing broadband to consumers on airplanes, some private jets but mostly commercial jets.  The incumbent in this niche is Gogo, whose original service uses a low-capacity terrestrial network.  I assume you have all used Gogo’s legacy service on an airplane at least once, although if you are like me, you have given up using it even though the alternative on any given airplane is to go without any internet access; it doesn’t take very many users per airplane before download speeds become so slow as to make the service literally unusable.  ViaSat’s satellite-based access has far more capacity and therefore a far better service.  Gogo scrambled to catch up and is now offering a competing higher-capacity satellite-based service.  It seems likely that, over the next few years, every existing domestic commercial plane with Gogo’s legacy system will either be retired or refitted with someone’s satellite service.

ViaSat’s foundation customer has been JetBlue, which has already outfitted most if not all of its planes.  ViaSat now has 476 commercial aircraft in service and is adding about 30 per quarter.  (That is 44% growth for 4QFY16, but the % number is shrinking rapidly as the base gets larger.)  They also have another ~600 government and private jets, which generate much lower revenue per aircraft.  Last week American Airlines, which until now has exclusively used Gogo’s ground-based network, announced that it would use ViaSat on its future already-ordered 737s (~100 planes) and use Gogo’s new satellite-based service on its future Airbus planes (134 planes).  American will also switch out 400 older planes from legacy Gogo to satellite and hasn’t decided whose satellite service it will use for them.  ViaSat has also signed Virgin America and Qantas.  A reasonable assumption here is that ViaSat will continue to add planes at a good clip as its market share on U.S. flights grows from 0% to 50% and as broadband’s total penetration of commercial flights continues to increase, particularly internationally.

The other two Satellite Services pieces are a mobile broadband service (maritime and ground-mobile) and various enterprise broadband services.  I am guessing these smaller bits total $20m of revenue.  If that number is too low, then the high-value aircraft business’s revenues are smaller and ViaSat’s total value is smaller.

Satellite Services has a sexy-looking 44% “adjusted EBITDA” margin, but only because it is by far the most capital-intensive segment.  (D&A for Satellite Services is 29% of its revenues, compared to 9% for Government Systems.)  Management of course trumpets “adjusted EBITDA” (which adjusts normal EBITDA by excluding the hefty stock comp).  It buries in tables each segment’s operating earnings excluding amortization of intangibles (EBITA).  The segment EBITA margin is only 15%.  Margins have been grinding higher as the business adds more revenue at high incremental margins, but they may stall out or even reverse as the company starts expensing more ViaSat-2 and ViaSat-3 costs.

While Satellite Services gets 90% of the attention, Government Systems remains ViaSat’s largest segment.  ViaSat sells satellite communications equipment and services to governments (revenue is 2/3 product sales, 1/3 services).  The 10K has a detailed breakdown of the various offerings in this segment, but frankly those details make my eyes glaze over and are not material to the investment case.  This business chugs along at lumpy but single-digit average annual growth and a consistent 13-14% EBITA margin.

The final segment, Commercial Networks, is a black hole that sucks in any cash that gets too close to its event horizon.  ViaSat sells various types of communications equipment to non-government customers.  (90% of segment revenues are product sales.)  This segment has shrunk each of the last three years and has lost money each of the last six years.  It has shrunk enough to become almost immaterial to the investment case and will become material only if management can turn the segment around; there is no sign of that occurring any time soon.

VIASAT HAS BEEN A CONSISTENT DISAPPOINTMENT

ViaSat essentially entered the satellite-internet consumer market by paying $568m to acquire Wildblue in December 2009.  Wildblue already had two operational satellites and 424,000 subscribers.  On this base they added ViaSat-1, which, as management repeatedly trumpeted for years leading up to its launch, had “more capacity than all the other communications satellites covering North America combined.”  The dream in 2010, heavily flogged by management and underlying everyone’s decision to own the stock at that time (including ours), was that ViaSat-1 would allow ViaSat to triple its subscriber count at high ARPUs and incremental margins (because the customers, almost by definition, couldn’t get broadband from a competing terrestrial source).

We lost faith in that dream by mid-2011, even before ViaSat-1’s launch, because it seemed inevitable that communications technology would advance as usual and competing sources of internet access would eat away ViaSat’s purported advantages long before it could reap the expected profits.  That fear turned out to be more justified than even we thought.  The combination of Wildblue and ViaSat-1, relative to expectations, has been a disaster.  I still have some sell-side predictions from 2011.  Here are the forecasts for FY16 compared to the actual results:

I may be understating the subscriber disappointment because it is not clear from my old notes whether the 1,200,000 estimate was for ViaSat-1 subs alone, on top of the existing 424,000 Wildblue subs, for a total of over 1,600,000.  I know that at least one analyst forecast 1,117,000 subs in FY16 for ViaSat-1 alone.

The total-company financial disappointment has been “helped” by Commercial Networks but is mostly due to Satellite Services.  At a $55 ARPU, the 514,000 subscriber shortfall represents $339m of annual revenue not generated, with something like an 85% incremental profit margin (both EBIT and EBITDA).  ViaSat paid $568m for Wildblue’s satellites and 424,000 subscribers and $500m for ViaSat-1, and it has ended up with only 697,000 subscribers.  ViaSat-1 has been saved from more-total financial disaster by the aircraft business, which did not even exist in the forecasts back in 2011.  Even with the aircraft business, the financial numbers are horrendous.  ViaSat’s cumulative Satellite Services EBITA since acquiring Wildblue (FY11-FY16) is only $41m.  If you assume that 16% of the Government Systems EBITA comes from satellite connectivity – services are 1/3 of revenues; assume connectivity is half of that and systems development and installations are the other half – that is another $64m of EBITA, and the total EBITA from the satellites is $105m over six years, an average of $18m per year.  On a $1068m investment (which excludes any later capex that has supported this business), that is a 1.6% ROI without any time discount.  Admittedly these cumulative and annualized returns numbers should rise modestly over the coming few years, but they will still be bad.

The over-optimism did not stop in 2011.  On their 4QFY15 earnings call, management gave guidance for FY16 of $360m adjusted EBITDA.  By the 3Q16 call in February, they had lowered the guidance to $345-350m.  Three months later, they reported actual adjusted EBITDA of $331m.  Revenue came in at $1417m, versus consensus estimates of $1467m six months earlier.

And the hits keep coming.  On the February earnings call, management gave FY17 adjusted EBITDA guidance of $380-390m.  On the next call three months later, their “outlook” gave no numbers for adjusted EBITDA but instead gave several oblique nudges and winks that, once pieced together, implied adjusted EBITDA of around $363m.  Even when pressed by the sell-side analysts during the call’s Q&A portion, management would not admit that it had lowered its guidance.  Good grief.

HISTORY WILL REPEAT

Although the most recent EBITDA downgrade was due to higher spending, ViaSat’s six consecutive years of disappointments have largely been caused by the iron law of telecom: the industry’s bandwidth capacity constantly increases and its unit costs decrease, often at “faster than expected” rates.  The consumers that ViaSat thought would use its satellite service instead were either able to get a landline or, more likely, felt well-enough served by terrestrial wireless speeds that went from a relative snail’s pace in 2011 to LTE supporting high-quality streaming video today.  The capacity/throughput arms race will continue.  Yet ViaSat’s management continues to display no understanding of this dynamic.  The 4Q earnings call was jarring for management’s implicit assumption that no one but ViaSat will materially increase their bandwidth offering in the coming years:

“[Our] segments results [remember, they missed the implied 4Q guidance, lowered FY17 guidance, and grew subscribers by 1.6%] very effectively show the competitive advantage of our highly productive satellites and the end-to-end vertical integration, as well as our overall business strategies. . . .

In general, we like to find business strategies that are disruptive in the sense that they change the dimensions of competition and in the underlying business models in our target markets. That means bringing value propositions to market that are fundamentally different than those that currently exist. It's not just same things overpriced or even same price for incremental performance improvements.  When we can bring true abundance where incumbents have depended on scarcity in markets that crave bandwidth, it's likely going to be a good bet. Incumbents with substantial assets built on the scarcity model generally find our approach difficult to defend and we can leverage powerful macro market forces. That theme is most obvious in our high capacity satellite services but there are similar effects in tactical data links and information security. Finally, each of these markets already involve significant entry barriers, and we believe that the technologies we inject into those markets can create durable competitive advantages and yield attractive returns. So we've got good success with ViaSat-1 in using disruptive bandwidth economics to move up market in satellite services. . . .  

Our satellite assets and partnerships along with scale from our defense business makes us uniquely suited to serve the bandwidth abundance model, and we anticipate continued strong growth ahead. And we're having striking success in the government market by delivering unprecedented amounts of bandwidth at very competitive cost to hundreds of previously un- or under-served airborne platforms and by investing in making the powerful situational awareness data embedded in Link 16, available to thousands and ultimately tens of thousands more participants, we are accomplishing similar effects in digital terrestrial networks. . . .

[T]o recap the investment thesis behind the ViaSat-3 constellation. Fundamentally, we think that maximizing the productivity of expensive satellite assets, in terms of delivering the most gigabits per second focused on the high demand geographic regions per unit capital cost invested, is a really powerful source of competitive advantage. . . .  We believe it's extremely difficult to attack the technology needed to reach this kind of productivity without the level of vertical integration we've achieved. It requires close coupling of multiple technology and business domains in ways that are not practiced by incumbent satellite operators, satellite manufacturers, or other members of the existing business ecosystems. . . .  

All of these statements are largely false.  ViaSat offers a quasi-commodity product of IP-based bandwidth.  Providing that bandwidth is complex for any provider, satellite or terrestrial, but the consumer doesn’t care about the underlying technology and buys based on throughput, data caps, and price.  There is no evidence that ViaSat has much of a competitive advantage over other satellite companies in providing bandwidth, either from its past results or from comparing what ViaSat tangibly does with what competitors do.  ViaSat is launching a new higher-capacity satellite because everyone in telecom must constantly invest more capex to keep up with the competition in offering more bandwidth per second, more bandwidth per month, and more bandwidth per dollar of revenue.  

Far from ushering in an era of “true abundance,” ViaSat is not even keeping up with its current capacity needs.  History is repeating here, too.  For a full year before ViaSat launched ViaSat-1, it had to stop signing new subscribers because it ran out of capacity on its older satellites.  At that time, management originally estimated that ViaSat-1 could handle 2.0 million subscribers.  Before ViaSat-1 even launched, management reduced the estimate to 1.5 million.  In the years after launch management found – surprise!! – that competitive advances forced it to increase the bandwidth offered to each subscriber.  In a replay from 2010, for the last four quarters – starting two full years before ViaSat-2 will go live – management has been blaming its inability to add any new subscribers beyond 697,000 on capacity constraints with ViaSat-1.

Today ViaSat’s Exede subscriber offering is merely on par with cellular wireless offerings; it’s sometimes faster but always a lot more expensive.  The most direct cellular competitor is Verizon, because Exede’s target customers are in more remote areas and Verizon has the best network coverage.  Exede offers 12 Mbps speeds with 10 Gb/month for $60 or 18 Gb/month for $100.  Subscribers must also pay $5/month to rent the modem.  Verizon’s LTE service offers speeds of “5-12 Mbps with peak speeds approaching 50 Mbps.”  Verizon’s most popular wireless plan has 3 Gb/month of data (plus unlimited voice and texts) for $45; the 12 Gb/month plan costs $80.  Put differently, upgrading plans to add an incremental 9Gb of data on Verizon costs $35, versus $55 (including modem) for 10 Gb from Exede.  Verizon also offers 18Gb/month for $100; i.e. buying an extra 15GB beyond the popular plan costs an extra $55, versus $105 for 18 Gb on Exede.  Exede will probably offer better capacity in a year or two once ViaSat-2 is operational – but so will Verizon as LTE keeps improving.

Even in aircraft, ViaSat has lost its competitive edge quickly.  A year ago ViaSat’s service was much better than Gogo’s terrestrial-based service.  Gogo responded by putting together a competing satellite-based service that uses Intelsat’s and SES’s Ku-band satellites.  ViaSat management spent part of its May earnings call arguing that Ku-band offerings (such as Gogo’s) are far inferior to ViaSat’s Ka-band service.  Yet a week later, American Airlines announced it was splitting its new-plane build-out between Gogo and ViaSat and was giving more planes to Gogo than to ViaSat.

So what is ViaSat-2’s $650m cost likely to get the company?

• It will expand the land-based service footprint to the Caribbean, Central America, and the top of South America (Venezuela and Columbia).  That will allow modest additions to ViaSat’s various non-consumer business lines such as maritime and oil and gas.  It could expand the consumer business as well, but the revenue would likely all be wholesale to a third-party sales channel rather than retail, at much lower ARPUs.

• It will provide coverage over the Atlantic, so that ViaSat can launch aircraft service for cross-Atlantic flights via a joint venture with Eutelsat, who will cover the over-Europe portions of flights.  (This JV deal has already been signed.)

• It will triple ViaSat’s total capacity.  But in the existing markets, what will the company use this capacity for?  For the land-based consumer business, there is unlikely to be much unmet need for satellite-provided broadband.  ViaSat-2 will mostly allow the company to increase its throughput (megabits per second) and monthly data caps (gigabits per month) to keep up with competition yet again.  The government business will probably keep growing at its usual pace, which it could not do without the new capacity.  Both of these “benefits” are essentially maintenance capex.

• That essentially leaves aircraft, the only true growth business.  Let’s say ViaSat does fabulously with aircraft and quadruples its aircraft in service over the next several years, to a number higher than Gogo has ever had even though ViaSat and Gogo will probably split the business.  If my estimate of $70m of FY16 aircraft revenues is correct, that would be $210m of new revenue.  At a 45% EBITDA margin, that would be $95m of new EBITDA, which will take years to ramp up.  That is good money, but not enough by itself, not for an asset with maybe a ten-year useful life.

VALUATION

Given the broad range of possible outcomes, I don’t know what VSAT equity will end up being worth, but I am confident that it is worth much less than its current price.  I like technology shorts for which a very bullish bull case produces an equity value that is at or below the current stock price.  ViaSat meets that requirement.  Here are the assumptions needed to get to a DCF value near the current stock price:

• Consensus revenue growth of 6% in FY17 and 9% in FY18, 12% each for FY19-20-21 as ViaSat-2 and ViaSat-3 ramp up, then trailing down to 10/9/8/7/6% in the out years

• EBIT margin triples from 3% in FY16 to 9% by FY21, higher than it has ever been, and stays there

• 20% tax rate in FY17, 30% afterwards.  (As a reference point, management says FY17 tax rate will be the ~39% statutory rate minus $8m for the R&D tax credit.)

• Capex, which has far exceeded D&A for each of the last eight years, somehow drops below D&A starting in FY18 and runs at least $100m below D&A each later year all the way until the terminal year

• The ratio of net working capital to revenues stays at its current 17%

• 3.5% terminal growth rate, 10% WACC

If I keep all those assumptions but make capex merely equal to D&A – still much better than any of the last seven years – the DCF valuation already yields 20% downside.

A more hit-you-in-the-face valuation approach is comparing VSAT’s EV/EBITDA multiple to that of its peers.  As detailed above, VSAT is trading at almost 14x EV/EBIT.  Inmarsat, Eutelsat, and SES are all at about 5x-7x.

WHY SHORT IT NOW?

Look again at that stock chart above.  VSAT’s stock went from $56 to $70 in February on the news that it might win a 200-plane order from American Airlines (plus a stock market rebound).  It then went from $70 to $74 in early June on the news that ViaSat did win the business for only half of those 200 planes, with the other half going to Gogo.  If you try to strip out the market dip and recovery, it looks like the American Airlines news pushed the stock from $60 to $74.  That is a $700m increase in market cap.  The American Airlines win is a good sign that future wins will come, but future wins like this are already mostly included in analysts’ estimates.  For comparison, Gogo, with 1,500 aircraft in service today, has a market cap of $825m and an EV of $1.1 billion.  The good news is far more than in baked into the stock price.

Meanwhile, the stocks of Inmarsat, Eutelsat, and SES have collapsed over this same time period, primarily because the inevitable force of telecom price deflation and increased capacity needs – the same force that regularly hits ViaSat – has hit them even harder than usual lately.  ViaSat’s CEO was asked about this disconnect during the last earnings call, and he went straight back to the fantasy of unconstrained bandwidth that no one else will ever provide:

Summing up:  The disconnect between ViaSat’s reality and its stock price is currently as high as I have ever seen it.  The short may take some time to “work,” but I see unusually little upside risk, relative to the market indexes, unless a truly unforeseen blue-sky bull scenario unfolds.

Catalyst

• Years of steady investor disappointments in both revenues and profits.
• Baupost owns 24% of the stock and has owned it for many years.  Like many value investors, we first looked at VSAT in 2010 because of that ownership.  We have done very well over time shorting technology stocks that are owned by successful big-name investors, so when we run across a highly-respected owner like this, we worry less than we would have in the past that we are missing something important. That is especially true here where we are former longs and have known the business for years.  However, there is no question that Baupost's continued ownership helps other investors "keep the faith" and helps keep the stock price elevated.  That fact is a double-edged sword.  If Baupost ever sells, this stock will probably be down 30% in a few weeks, on top of whatever decline Baupost's selling pressure causes before it discloses the sales.