The curious case of RapidEye

In Episode 02 of #Spacedonewrong we explored how free satellite data can kill commercial initiative. In this Bonus Article we are looking on a specific example: RapidEye.

RapidEye was a German EO constellation that started as a initiative of Kayser Threde GmbH (now part of OHB group) in the mid 1990s. The company was founded in 1998 by the two founders Manfred Krischke and Stefan Scherer who convinced the owners of Kayser-Threde and the Vereinigte Hagel Insurance company to join them.

The idea was to use small satellites (150kg) to build a fully commercial observation satellite constellation. Target was to offer medium resolution with daily revisit (not daily coverage) for agriculture and insurance industries. For this a system of 5 satellites with 6.5m GSD VNIR and 77km swath was envisioned.

Launched in 2008, it became the first commercial satellite EO constellation in Germany. During its operational life it was considered a very impressive example of what small satellites can do in an operational scenario. The 5 satellites produced more than 5 million km2 of imagery per day, data that has been used by end users all over the world. Unfortunately, despite all its successes, when RapidEye was retired in 2020 it was still the only German Earth Observation constellation. Let's find out why!

Investment

The 160MEUR system was funded in 2004 with substantial private investment. Lead private investor was Deutsche Hagel (a hail storm insurance). Roughly 20MEUR were brought in as vendor financing by the prime contractor and other vendors.?80MEUR or roughly half of the investment was brought in in form of banks loans which were secured by German state guarantees. On top of that direct subsidies were provided by the German Federal State of Brandenburg (37MEUR) and the German Aerospace Centre (14MEUR).

Seeds of Destruction

The seed of its own destruction was already planted in the way RapidEye was set up and financed. Problematic were two things: first that the direct subsidies of Brandenburg required the company to set-up and maintain a large number of jobs and second that all the loans were secured by government guarantees which meant that the financing banks had very little skin in the game.

Implementation

Dr. Manfred Krischke explained to me that RapidEye approached potential suppliers with an international RFP and that the offers from established companies like DASA (today Airbus) were typically 2-3 times higher than those of SSTL (which later would win). In the final stage two companies SSTL (UK) and Qinetiq (UK) were preselected but due to intervention of the financing banks (which insisted on bank guarantees over the entire sum, which neither company was able to provide) MDA (Canada) was brought into the picture as system prime.

In 2004 when RapidEye selected the SSTL X-150 it still untested and everything that SSTL actually had flown was inferior to the available German micro satellite platforms.

What I find interesting, and I have not found a satisfactory answer yet ,why nobody in RapidEye or in the German stakeholders considered one of the two German micro satellite developments: DLR TET and the TUBSAT series of satellites.

While SSTL today is a household name, the race to market superiority was far less clear back in the early 2000s. In 2004 at the time when RapidEye was looking for a bus provider the two German developments were arguably superior to everything that SSTL had flown until this time.

You may argue that the SSTL X-150 was still superior, which is true, but the development effort to come from an X-70 to an X-150 (which was untested at that time) is much bigger than say using a DLR BIRD platform and do the necessary modifications.

If you wonder why Germany today has so little to show in terms of micro satellites and mass manufacturing one should consider that our industry likely be in a significantly better state if those 5 satellites would have been built locally using the available technology rather than to have them being manufactured in the UK.

Alternate History (what would I have done differently)

Based on a discussion in the comments I have dreamed up what a team of Prof. Udo Renner (father of the TUBSATs) and Vectronic Aerospace (spin-off of the TUBSAT group) - as Berlin Space Technologies was not yet around - for the bus and some of the less overengineering folks from DLR Optical Sensor Systems (DLR OS) could have easily built in less than 2 years for significantly less money than was spent in on the SSTL + Jena Optronik Solution. Considering the cost of the payload alone I wager you could probably build and launched twice as many satellites, easily.

Note: the DLR BIRD interestingly used the same battery system as the TUBSAT-C bus (Eagle Picher - NiH2), only difference is that Bird used 8 cells (28V) and TUBSAT C used 4 (14V). It would have been relatively straight forward to modify the power system. The modular bus with payload and bus separation would have made it simple to just add a fast payload computer (including storage) and an X-Band transmitter.

Bottom-line: an opportunity was lost when the RapidEye satellites were manufactured in the UK while similar or better technology was available in Germany, likely at significantly lower cost.

Businessmodel

The business model of RapidEye was to sell value added products - not raw data - to end users. Dr. Manfred Krischke explained to me that: "the key dimensioning factor of the system was, that for ag-applications you need to reliably have at least a full coverage of the fields every 4 weeks. Taking cloud statistics of agricultural-areas into account this could only be achieved by a daily revisit combined with a large capacity achieving full coverage in 5 cloud free days."?

In order to get guarantee a cloud free image once per 4 weeks RapidEye was sized to have a revisit time of 1 day and a capacity to image all areas of interest in 5 days

The system could generate more than 5mio km2 per day. That meant that even a middle to high single digit percentage of this capacity sold at the target price of 1EUR/km2 would already be able to refinance the whole venture.

Highly Processed Image data (>Level 3) was priced at 1EUR/km2. Selling a fraction of the 5mio km2/day would have been sufficient.

Unfortunately, as RapidEye found out the market was not yet mature enough for this business model. So instead of value added products what they were mostly selling was raw data, probably at a much lower price, mainly to institutions which were used to do their own post processing.

Unfortunately the market was not yet ready and two years after launch RapidEye had only achieved 14MEUR annual turnover.

So while they were making 14MEUR turnover pear year in 2010, they also were facing 10+MEUR running cost for their 120+ people team and likely 4+MEUR in interest for the credits (assuming 3% at reduced government backed rate).

This was obviously not sufficient to make sufficient money to justify the running expenditure.

Dr. Manfred Krischke explains: "There has been a considerable cash reserve in the RapidEye funding plan for unexpected delays and slower market entry. All delays due to technical issues and resulting operational cost were contractually covered by the prime contract and the prime contractor [MDA] was supposed to cover those cost. Unfortunately, a delay in the camera development [by Jena Optronic] led to a consequential problem with the launch, which was put into questions because of tensions between Russia and Kazakhstan at that time. This problem led to the option for the prime contractor, to get out of his obligations and to call for a force major case. This ended up in the fact that the cash reserves of RapidEye were consumed by issues purely caused by manufacturing delays."

"RapidEye was protected by contract against cost overruns. However, due to a loophole main contractor MDA was able to escape its obligations citing force majeure. Therefore this cash reserve was no longer available when it was needed.

It is thus easily visible that with these expenses, slower growth and no cash reserves to cover those,?they had no real chance to pay back the loans let alone acquire money for the next generation of satellites. It is apparent that two things would have been required to make it a more profitable business under the existing market conditions, sell more data and reduce cost of labour.

Why did they not pivot

As indicated earlier the subsidies came at a price of hiring and maintaining a workforce >100 people. The problem was the business that RapidEye was able to realize was hardly worth the upkeep of such a large team. Truth to be told the technical management of the entire constellation probably needed significantly less than 10 people.

For the government investments RapidEye promised to set up location in Brandenburg and to create 120 jobs. This job requirement was a constant source of their burn rate which contributed to their untimely demise.

As long as you are only selling raw data you also do not need more. Unfortunately if RapidEye did not want to risk paying back the 37MEUR that they had received from Brandenburg they needed to maintain it. While the company managed to increase their sales year after year it was still not enough to stop the bloodletting by high operation cost and interest, therefore RapidEye was looking for new investors since 2010.

Meet the black swan

After launch RapidEye was facing a troubling set of conditions. In addition to the slower than anticipated sales (likely <1% of their capacity in 2010) an untimely event happened. The black swan came in the year that RapidEye was finally launched. In 2008 a competing satellite system was announced: Sentinel 2 of the GMES (later EU Copernicus) was slated to deliver wide swath medium resolution images. Normally government funded missions like LandSat focus on lower resolutions but at 10m VNIR and 5days global repeat cycle the Sentinel 2 hit to close at home. To add insult to injury it was further rumoured that the EU was considering to make the data access free. This was a change from earlier government sponsored missions which usually offered data under the COFUR (cost of fulfilling user request). It is not hard to imagine the horror of the RapidEye management when facing the looming Sentinel 2. As I have written in 2nd Episode of SpaceDoneWrong where governments give away services for free, business will die.

Bancrupcy

Therefore, in addition to trying increase business before meeting the impeding doom, RapidEye ran internal studies how to escape this government sponsored competitor. Two avenues were analysed: same resolution but drastically increased number of spectral channels or much higher resolution. Ultimately neither of these two roads were travelled. RapidEye started the process to look for new investors in 2010 but to no avail. Therefore, in May 2011 the financing banks pulled the plug.

Why did they do that? Well, the guarantees that the Federal state of Brandenburg had provided meant that they could step out virtually without losses.

The bank guarantees that the government had provided did two things: the drag RapidEye down via the burn rate of the job requirements and allow an easy way out for the banks when things went south.

Originally, the repayment of the loans was foreseen over a 5 year period starting from 2009 but the dim business outlook (aggravated by the upcoming Copernicus Sentinel 2) meant that they could either wait and see whether RapidEye would make the turn around or take the easy way out.

Firesale - a bargain is to be made

After filing for bankruptcy the courts appointed an insolvency administrators whose task it was to sell the company and its assets and try to satisfy (at least in part) the liabilities of the company. Dr. Manfred Krischke, who at this point was the largest share holder in the supervisory board points out that: "initially there was a considerable interest" and so in order to prevent the total loss of all investments he together with his business partner "Peter Müller-Brühl approached many players to see if they would support the reacquisition of the assets to continue the operations."?

150MEUR investment, sold for 14MEUR in a fire sale. I believe a better deal could have been made, especially since RapidEye was sold later for 100MEUR to planet.

The constellation was finally sold to Blackbridge (Canada) for 14MEUR or roughly 10% of the initial investments. Dr. Manfred Krischke points out that: "In the midst of the summer timing was to the favour for bidders who could quickly get a full picture of the situation and would have quick decision time." For me this all sounds like a firesale: Blackbridge bought the whole constellation, free of debt, for roughly 14MEUR , less than 10% of the initial investment and roughly the same amount that the company had as turnover the year prior.

After restructuring and without the debt RapidEye in the hands of Blackbridge was immediately profitable.

Since there were only much smaller loans to be serviced (14MEUR vs 80MEUR) the new RapidEye was immediately profitable with the satellites in space as assets. Considering the value that RapidEye was sold for a few years later I think its safe to assume that Blackbridge as the new owners struct quite a bargain.

A new life

Since the existence of RapidEye was still in danger due to the Sentinel 2 satellites the new owner followed the strategy of the old management. They announced a new constellation in 2014. The RapidEye+ which was to bring together 14 multispectral channels (1 more than Sentinel 2) and in addition a panchromatic channel of 1m GSD. The new constellation was announced for 2019 but never materialized. Instead RapidEye was sold for more than $100M to Planet in 2015.

I can only guess why Planet bought RapidEye. Probably in part for their archive and likely to acquire both algorithms and the trained team of 100+ earth imaging experts. The actual RapidEye image data (5 mio km2 per day) even though of higher quality than those of the Dove (100+ mio km2 per day) satellites would not significantly contributed to the Planet constellation.

Considering the massive profits that were made by selling the RapidEye constellation from Blackbridge to Planet one has to ask, why was there no sell-on clause?

What can be said is that the German government most notably the state of Brandenburg lost a lot of money. Whereas Blackbridge gained significantly a 5x increase over its purchase of the constellation. In its defence, Dr. Manfred Krischke points out "that the government invested the money to sustainably create qualified and tax paying jobs. This goal has actually been achieved (at least in Berlin) since still today after almost 20 years the Berlin Planet team with more than 100 highly qualified employees is a very essential part of the overall Planet company."

However, I personally think that while this is all well and good, it still does not answer the question why the government officials did not put a sell-on-clause in the contract. This would have them benefit at least in part from future sales of the constellation. I mean this is a tool used in almost any transfer of soccer players. If sport clubs can come up with contract clauses to protect their interest in future re-sales why is that not possible here.

Rest in Peace RapidEye

Finally after 12 years of operation, Planet decommissioned the satellites in April 2020. To an end came the first fully commercial EO constellation. Maybe they were too early, maybe they got the business wrong. They should however serve as an inspiration to us all.