Intro. Last week at the GEOINT Symposium in Washington D.C. National Geospatial-Intelligence Agency (NGA) Director Robert Cardillo gave the keynote speech billed as the introduction of a new NGA strategy. A key highlight of that speech is as follows (emphasis ours):
One of the most challenging things affecting our people in the near future is the SmallSat revolution. Some are uncomfortable with this seemingly uncontrolled movement of more and more sensors into space. And while I recognize that there are two sides to the world’s growing transparency, I am energized and enthused about this development. Frankly, it has pushed GEOINT to an inflection point. In the next five years, more than a dozen constellations, with hundreds of SmallSats, will launch, and continuously scan the earth. The SmallSat revolution will do just that: revolutionize the way we sense the planet. And I stress “sense,” as it’s much more than just images. It means our analysis of world events is going to be holistic and persistent. In fact, the democratization of GEOINT and the “darkening of the skies” is the opportunity of our time for our people. So we in government have to pivot, and change our mindset, to investigate multiple possibilities, and to better understand this complex situation. We won’t need to balance a finite collection capability against a seemingly infinite set of GEOINT requirements.
This is a big deal, particularly for DigitalGlobe (DGI) as the company currently generates approximately 50% of its revenue from the EnhancedView SLA via the NGA. The consensus view has been that this revenue stream is bullet-proof; we think this speech should break that view. But wait, there’s more…
Background. With the benefit of nearly two additional years of following DGI since our first post (VIC, Sept 13), we have gathered a few insights into the company and the market in which it operates. The first insight is that there are only two “killer apps” for satellite imagery that represent DGI’s core markets: (i) intelligence and (ii) mapping (or “LBS” as DGI euphemistically refers to it). While there are many other applications for satellite imagery, these are generally niche and either do not represent significant incremental revenue opportunities or they are highly fragmented and the cost of customer acquisition is high enough to be unprofitable for DGI to pursue. Furthermore, the two key applications are mature with respect to their demand for DGI’s imagery, which is quite clear based on DGI’s reported financials. If a third, sizable application emerges, we may be forced to change our view on the company, but until that point we remain highly skeptical that there is a secular growth tailwind of any kind supporting DGI’s markets (more on this below). The second insight relates to the value proposition of orthogonal imagery in the intelligence market and increasingly applies to LBS: resolution does matter, but it matters to different degrees for different customers for different use cases and, as such, higher resolution is not in-and-of-itself an effective recipe for higher demand and/or willingness-to-pay. In fact, we would argue that the vast minority of DGI’s customers’ use cases value the difference between 50cm and 30cm resolution; if they did value higher resolution, they would have likely already utilized aerial imagery for these applications (more on this below). A critical corollary to our second insight is that once a “good enough” resolution limit is reached for each customer and each use case (that is, the point where “I can see what I need to see”), then, as a general rule, all the incremental value of this imagery is generated by revisit, the change between images (or, “persistence”; that is, how many times you get to see the same image over a given time period), not the image itself. This is critical to keep in mind as we discuss the impact that small satellites are likely to have on DGI over time.
Thesis Revisit Point 1: DGI is not a growth company, silly. Part 1a: R.I.P, commercial growth story. Despite the sell-side’s unwavering belief in the “commercial growth story” at every point since the early 2013 merger between DGI and GeoEye, the company’s reported results provide compelling evidence it doesn’t in fact exist. Estimated pro forma commercial revenue in 2012 was approximately $260m (GeoEye + DGI, pre-merger excluding US Government revenue) and, two years later, in 2014 commercial revenue was also $260m. 1Q15 commercial revenue was $55m, down 7% year-over-year, and this is despite the launch of the highly-touted (by management) WorldView-3 (WV3) satellite in the summer of 2014 that sports 30cm resolution, the highest available anywhere. Commercial growth has failed to materialize to date for four primary reasons: First, new capacity at 70cm resolution entered the market in 2013 in the form of Airbus/Astrium’s two Pleiades satellites, which lacked the NGA’s preemptive access rights that burdens DGI and allows Astrium to differentiate itself with better service (primarily through faster tasking and image delivery times). Luckily for DGI (so far), Astrium used DGI’s pricing as an umbrella and thus has not disrupted industry pricing to date. Second, new capacity was launched by foreign governments that drove in-sourcing and cost DGI revenue, the most meaningful in its impact being the June 2013 launch of the Russian satellite Resurs-P, capable of 70cm resolution. (Two notes here: first, DGI classifies revenue from friendly foreign governments as “commercial” revenue; second, DGI guided well below the street’s 2014 revenue expectations in early 2014 and blamed – and continues to blame – US-Russian tensions as the primary driver of this lost demand, when, in fact, it has been largely driven by this incremental capacity.) Third, LBS is no longer a growth engine as Google has invested in its own air force to capture aerial oblique imagery (taken at an angle, giving image greater richness and more depth) at resolutions significantly higher than DGI’s satellites are capable of delivering. As Google sets the standard for consumer-facing digital maps, the rest of the industry is forced to keep up, and, as a result, the inherent value of DGI’s imagery in the eyes of the LBS industry has been degraded; satellite imagery is simply no longer a key point of differentiation in the eyes of the user. Which brings us to the final reason the commercial revenue growth story here is dead: DGI’s management. We acknowledge that there is the theoretical potential for new killer apps, but we expect these to come as a result of the SmallSat revolution combined with software innovation out of Silicon Valley, not from a group of government contractors in Denver utilizing decades old hardware. DGI’s management has thrown the proverbial spaghetti against the wall here (and sold some unicorn and rainbow shaped pasta to The Street to boot) and nothing has stuck despite ample servings of trendy buzzwords (“cloud”, “crowd-sourced” and – our favorite – “geospatial big data” just to name a few). Boiling it all down, DGI is not an innovative company; it is a company with a legacy built around two core tenets: serving one big customer (the US government) and fooling the public markets to help the government fund additional spy satellites. And so far, these are the only two things they have excelled at. Which brings us to Part 1b: NGA (US Government) revenue is now at risk. With the emergence of small satellites at 1m resolution, the industry is now at the precipice of having “good enough” imagery from a resolution standpoint that is available at significantly lower capital cost and offer a superior value proposition to most of DGI’s customers, including the NGA, with its dramatically higher revisit rate. Cardillo said it himself just last week in the quote above: a SmallSat revolution is coming very, very soon. Against this backdrop DGI will find it nearly impossible to retain its $300m per year subsidy from the NGA. This viewpoint is not ours alone, it is supported by many industry and government experts we have spoken to and also has the benefit of common sense behind it. The US government budget pressures are relentless and the NGA’s budget isn’t growing. In addition, if the incremental value is in revisits – particularly for intelligence use cases – then of course the NGA would shift budget. The capacity is coming from private market funding (see: Google’s purchase of Skybox, and the venture funded Planet Labs, Deimos and BlackSky constellations to name a few), so the NGA stands to benefit from private capital at an even lower cost to them than with DGI. Finally, while the EnhancedView contract runs through 2020, it is cancelable annually and we have a clear precedent in the NGA’s cancellation of GeoEye’s comparable contract. While we do not pretend to know when exactly the SmallSat revolution will hit DGI’s top line, with Cardillos speech last week coming directly on the heels of the NGA’s May 4th formal request for information (RFI) from SmallSat providers (available at FedBizOpps.gov by searching solicitation number HM0476-15-NexGen-Imagery) – we now feel extremely confident that the risk to DGI’s US Government revenue is clearly to the downside over time and, by extension, the discount rate one should put on that revenue stream has gone up significantly.
Thesis Revisit Point 2: Sensor supply about to explode, pricing to implode. We see three sources of significant increased capacity: SmallSats, foreign governments launching traditional (DGI-class) ultra-high resolution satellites and commercial drones. For SmallSats, the good NGA Director has already made a much more credible case than we ever could for the coming flood of supply. We won't belabor the point other than to highlight again that this capacity is truly disruptive to DGI. If the SmallSat's capital cost being approximately 1/20th of the cost of DGI's wasn't problem enough, they also redefine the vector of competition in the industry away from DGI's historical competitive advantage (resolution) and towards revisit where DGI cannot compete effectively. The second source of supply growth occurring is foreign government in-sourcing DGI-level earth imagery satellites (maybe not leading edge 30cm resolution, but good enough at 70cm +/- 20cm resolution). The motivations for these investments can range from (i) being overly dependent on DGI's capacity (through DGI’s Direct Access Program (DAP), ~15% of total revenue) yet being stuck behind the US governments preemption rights on said capacity, (ii) the desire for greater secrecy and not trusting any third-party capacity, particularly DGI’s in a post-Snowden world, to (iii) long-term aspirations to increase competitiveness in space technology. Examples of this from our research include China, Russia, UAE and South Korea - and those are the ones that are unclassified! What is particularly pernicious for DGI is that these are actors launching capacity for non-commercial motives yet who, in most cases, only have demand for imagery over a small subset of their owned supply; because, by their very nature, these satellites will only add to the structural overcapacity of imaging sensors over 95% of the earth’s surface that already exists from DGI's fleet alone (even DGI’s management will admit that demand > supply only in a subset of regions and even then often in only a subset of seasons). UAE, for example, has been public about justifying the cost for two (two!) of these satellites by having Astrium sell their excess capacity into the commercial market. Even more dangerous for DGI, China’s long-term space aspirations are quite aggressive, as is their near-term launch schedule. China has been open about their plans to directly enter the commercial satellite imagery market and do what China does best, compete on price. Finally, the short-term (in emerging markets) and medium-term (in developed markets) capacity growth from drones should also not be underestimated. First, they will compete with DGI on resolution as the only limit to their resolution is the cost of the on-board imaging sensor and, on average they will image at or under 12cm resolution. Second, drones have the benefit of being able to fly under cloud cover which is quite often an important limitation to tasking. Finally, the fact that the drone industry is well funded and developing hardware for myriad applications outside of imagery, means that the development cost advantage for imagery service providers leveraging drones should dwarf that of SmallSats; also, time to market will be much more rapid once regulations allow flight in most developed markets. Our research suggests drones are rapidly growing market share in developing markets where regulations are much less stringent or non-existent and that developed market regulation changes are around the corner, in the US at least. Taken together, this capacity growth should be terrifying to DGI (and their investors) given DGI operates the highest cost, longest lead-time assets in a very high fixed cost, near zero marginal cost industry. Add to this recipe the extremely short economic value half-life that DGI's imagery experiences in the existing supply environment and pricing here could get very, very ugly.
Thesis Revisit Point 3: Catalysts abound (again). The sell side analysts have had zero qualms taking DGI’s management team’s rosy prognostication for 15%-20% growth in commercial revenue and continuously projecting an inflection point in that area, despite the disappointments to date. As such, DGI IR has repeatedly needed to walk a fine line of managing down near-term expectations without diminishing long-term bullishness. Going into the initial 2015 guide early this year, IR managed down consensus’ 2015 revenue as much as they could (~10%) as the beleaguered CEO, Jeffrey Tarr, likely wouldn’t survive a second consecutive year of disappointing revenue guidance, having already thrown his previous CFO under the bus (ed note: as much as this new CFO looks and smells like fresh, dear-in-headlights sacrificial meat, he was smart enough to defer his joining of DGI until immediately after the full-year 2015 guidance was released, so as not to dirty his hands with it). Tarr therefore chose an alternative route with 2015 guidance: straddling consensus’ 2015 estimates yet including cadence in the guidance that cleared the decks for 1H15 but offset this with hockey stick revenue projection for 2H15. The hockey stick was justified by the expected ramp of WV-3 (despite its launch in 3Q14) and came with a host of explanations as to why it takes so long to ramp despite all the hype around massive demand resulting from the 30cm resolution. Never mind that the NGA has already purchased over 50% of the capacity of WV-3 and that that 50% is probably more like 80%-90% of the desirable capacity (e.g., time over the Middle East) from our checks. DGI then printed a disappointing 1Q and has been talking down both 2Q and 3Q estimates, further bending the hockey stick towards 4Q. And while you might think that low expectations for 2Q and 3Q results does not provide a great set-up – we’ll offer up a different view. First, given the underlying fundamental challenges facing this business and its ludicrous valuation (more on this below), we think management has set such an obscenely high bar for 4Q results (to be printed in Feb 2016) that there is very little incentive for rational longs to add after 2Q results. Second, the more likely outcome is that management will have to lower full year guidance in conjunction with either 2Q or 3Q results. Third, even if management does happen to get to the mid-point of the range for 2015,what is clear to us is that consensus estimates for 2016 are completely unachievable as the sell-side has generally plugged in either high-teens or 20% year-over-year growth for commercial revenues next year on top of what is already aggressive full year 2015 revenue estimates based on our research. There is no evidence we can find as to why, after nearly three years of zero growth in DGI’s commercial revenue, we should suddenly see it grow even 10%, say nothing of 20%. Also remember that 2015 had the benefit of the step-up in NGA recognized revenue from $255m to $337m which won’t repeat in 2016. As a final note, DGI is scrambling to launch WV-4 (formerly GeoEye-2, the “ground spare”) in an attempt to improve their refresh rate and to have a satellite in orbit that does not suffer from the NGA’s preemption rights. However, management has admitted that the earliest this will generate revenue is 1H17.
Thesis Revisit Point 4: DGI’s valuation and its multiple layers of lunacy would make Escher proud. The bull case on valuation is built on the following argument: on the US Government portion of the business DGI should get at least a defense industry 8x-9x EBITDA multiple and maybe even above that multiple range because they are a sole provider post the GeoEye acquisition. On the commercial portion of the business DGI should get an information services multiple of 12x-14x EBITDA because it’s going to grow that revenue 20%. So for a second we’ll put aside how foolish it is to value this business on an EBITDA multiple at all (but we’ll come back to that) and point out that (a) DGI’s US government business at best will grow at 0%-1% CAGR for the foreseeable future, well below the defense comps; (b) though we don’t have EBITDA by segment, we would guess that the vast majority of DGI’s EBITDA is generated by the US government as the commercial market is more competitive and has a higher cost to serve and, of course, customer acquisition costs; (c) the US government business faces significant risks from Cardillo’s SmallSat revolution that has no analogy for the highly diversified defense businesses in the comp set; (d) the list of reasons why DGI’s low-quality, no-growth commercial business is nothing like the information services companies in the analysts’ comp set is not short and this write-up is long enough as it is, but suffice it to say our objections have something to do with growth prospects, barriers-to-entry, availability of substitutes, capital intensity, and management track record, among others. One point we will emphasize, however, is capital intensity. Because DGI’s assets become obsolete and fall out of the sky every 12 years (though we think obsolescence timing for DGI’s satellites could compresses meaningfully in a post-SmallSat world) it makes much more sense to value DGI on a EBITDA less capex basis – or even better on an EBIT basis as D&A is a pretty good proxy for normalized capex as near-term capex is not given the “ground spare” acquired in the GeoEye merger. We recently noticed that DGI trades at 27x consensus’ (unattainable) 2016E EBIT which we think is flat at best but consensus believes can grow 9%. Note, for example, that Facebook (an arguably unfair analogy) trades at 20x 2016 EBIT and is expect to grow EBIT at >25%. DGI trades at a massive TEV/EBIT premium to even the sell-side’s defense and information services comps…and that’s on consensus’ estimates. Oh yeah, and we don’t see any viable path to sustainably positive EPS here, given the leverage. To hit this point home, another entertaining valuation exercise is DGI’s unit economics. Take the $250m of pre-WV-3 NGA SLA revenue and divide it by three and then take the rest of the non-SLA 2014 revenue and also divide it by three. Then, for WV-3, give it only $50m of NGA (cash) revenue and perhaps even give them a nice premium for 30cm pricing vs. the first three satellite’s non-NGA revenues and have higher incremental GM% for WV-3. Finally do the same for WV-4 but, obviously, don’t give it any NGA revenue. Even when we put what we feel is an aggressive EBIT multiple on each bird’s unit economics, we can barely cover the debt implying DGI’s satellites do not cover their cost of capital. Our view is that businesses that don’t cover their cost of capital should trade at less than their tangible book value, which for DGI is $11/share.
I do not hold a position with the issuer such as employment, directorship, or consultancy.
I and/or others I advise hold a material investment in the issuer's securities.
Earnings missing and negative guidance revisions.