Back in 1997, Mike deGyurky, a program manager at the Jet Propulsion Laboratory (JPL), had a design for a giant blimp, perhaps a mile in length. With a cargo capacity of 50,000 tons or more.
Mike and others at JPL had the skytrain concept, which were a bunch of blimp "box cars" connected together for less drag and more fuel efficiency. The largest of the box car designs was about 45,000 tons of cargo. Updated designs would be even better now because they were depending on thin film material for the skin and for thin film solar cells for power. Both of those have seen a lot of improvement and more improvement is anticipated.
JPL was and is not in the business of building airships, but was directly connected to an institution that was; The California Institute of Technology (CalTech). CalTech had been involved in Zeppelin research during the early 1930's. Theodore von Karman, a professor of aerodynamics at CalTech and a person who had participated in airship research in Germany (including the construction of the Zeppelin "Los Angeles" as part of wartime reparations to the US.) had proposed high speed dirigibles in his autobiography . This connection was first noted by deGyurky.
There were a number of design questions that arose and a number of which remained unanswered at the time. Among them, how would the system be powered? Elements of the SkyTrain could be covered with new ultra light weight solar cells to the point of being completely solar powered. A conventionally fueled backup would be necessary for staging operations. The 1994 analysis showed that this would reduce the pure solar SkyTrain cruising speed to around 43 mph. [a proposed Skycat airship design should have a speed of 97mph. The old Zeppelin's had a maximum speed of about 65 mph, There is an Aeroscraft, hybrid airship, that has a top speed of 174 miles per hour.]
1994 analysis developed the concept in a simple minded comparative analysis of three prototypical airships, each autonomously powered and controlled, able to link and unlink to sibling cars at will.
The first SkyBoxCar analyzed was a small technology demonstrator. It was sized so that 50% of its buoyant capacity was used for lifting cargo. As a demonstrator, it would fly at relatively low altitudes and would be unable to take advantage of favorable clines in wind and solar irradiance. It would cruise at 31,000 feet at a speed of 30.7 miles per hour with a cargo capacity of 4000 pounds. It would be 125 feet long and would cost $2.5 million dollars.
The second SkyBoxCar analyzed corresponded to that of a Mack truck. It was sized to carry 45,000 pounds of payload. This increase in scale produced an efficiency increase to 74% compared to the 50% of its smaller sibling. It would cruise at 31,500 feet at a speed of approximately 43.4 mph. At 244 feet long it is four fifths the length of a football field. It would cost approximately $9,485,000.
In the interest of sheer immensity, a third SkyBoxCar was modeled with a cargo capacity 45,000 metric tons. This radical increase in size produces an efficiency increase to 98% compared to the 74% of its smaller sibling. Like its sibling however it would cruise at 33,500 feet at a speed of 43.4 mph. At 2900 feet long, it would be over half a mile in length and would cost a little over 1.3 billion dollars. A SkyBoxCar of this size violates the concept of bite-sized chunks, but is of academic interest because of its lifting efficiency and flight envelope.
"With a train of 50 airships, as opposed to 50 independent airships, you could realize perhaps a 50 [-98%] percent savings in energy, and the savings go up as the speed of travel increases.
Mass produced airships could have a projected cost of 10 cents per ton-mile, compared with the 40-to-50 cents per ton-mile charged by standard air carriers.
Skycat airship. A completed ship should be flying in 2008 and production models in early 2009.
If we had cheap carbon nanotubes [prices likely falling from $200/kg to $4/kg over the next few years] that were able to provide most of the strength to the macroscale and next generation solar cells, then Skycats that were made on the scale of the giant airships would be able to travel at 300-600mph and carry over one hundred thousand tons. Something that could make sense in the 2015-2025 timeframe.
P-791, an experimental aerostatic/aerodynamic hybrid airship developed by Lockheed-Martin corporation. The first flight of the P-791 was made on 31 January 2006. The P-791 appears to be essentially identical in design to the SkyCat design
The cancelled DARPA Walrus airship (500-1000 tons of cargo) whose work continues with Skycat and P-791.
or 
The cloaking effect depends on a material's "refractive index," or its ability to influence the direction of light that passes through it. Light tends to prefer the quickest route between two points, which is normally a straight line. With metamaterials, however, the quickest path can be one that bends around an object.