Paul Spudis visits the launch cost issue and finds a new lower limit on the price of getting to LEO: The Moon: Creating Capability in Space and Getting Value for our Money.

Dr. Spudis shows the cover of a 1994 report on cheap access to LEO and claims that "nothing has changed" and goes on to argue that it never will. However, I could just as easily find a report from the 1970s hailing, as Dr. Spudis does so often today, lunar resources as the magic ingredients that will allow for vast human enterprises in space. In 40 years, however, that hasn't happened. Does that prove it never will? Of course not. There are good reasons for the failure of the Moon to be exploited and there are good reasons for the failure to achieve low cost access to orbit.

Dr Spudis believes that the concern over the high expense of is merely an assumed "verity" with groups like the Augustine panel. He says also that the

“New Space” community is aggressively campaigning for commercial launch services, pronouncing any NASA-designed/NASA-built space system politically “unsustainable” due to the agency’s inability to receive adequate funding from Congress over the lifetime of a flight program.

No, this not what anyone in the NewSpace community is pronouncing. The problem with NASA in-house systems is that they involve both enormous development costs (e.g. GAO est. for Ares I/Orion is $35B-$49B) and enormous operational costs (e.g. the Augustine panel predicted recurring costs of $1B per flight for Ares I/Orion). New Space proponents do not say that a too small NASA budget is the problem. The problem is the staggering waste and the huge opportunity cost arising from losing all the productive work (such as proving in situ resource utilization on the Moon) that could have been accomplished with that money.

He points out that NASA's budget is large in absolute value but small relative to other programs. I'll add, however, that NASA is fighting for funding with hundreds of other programs in a non-security discretionary budget of about $550B. The possibility of a substantial increase in NASA's budget is nil. It's in fact a historical anomaly that NASA's budget of $19B is as big as it is. The budget for human spaceflight alone is probably larger than the civil space spending in the rest of the world combined. Either NASA learns to spend its money more cost-effectively or it will accomplish less and less.

Dr. Spudis goes on to correctly point out that fuel contributes only a tiny fraction to the cost of a launch. However, he then claims that high costs are all due to the expense of operating complex systems with an army of expensive technicians and uses the Shuttle as the standard for this. However, the Shuttle is exactly the sort of NASA in-house system whose design led directly to extremely high operational costs. (By chance, a post today by Alan K. Henderson includes a quote from this article by Rand Simberg that nicely summarizes how the high costs of the Shuttle came about.) A partially refurbishable, grossly over-sized vehicle with a major component thrown away on every trip was guaranteed to be expensive to operate.

He notes correctly that high launch rates would bring down the costs of expendables but fails to see that high launch rates for fully reusable launch vehicles would lower costs even more.

He cites the currently quoted price for the SpaceX Falcon 9 of around $5000/kg to LEO and then notes that this price is similar to those for other vehicles. Thus there must be be a plateau in launch prices around this value. (This is at least an improvement over John Pike who has always claimed that $10k/kg is the plateau.) However, that Falcon 9 price assumes no reuse of any components. If SpaceX achieves even partial reuse of the boosters or their components, this will lead already to a price break in the Spudis Limit.

SpaceX is attempting to add reusability to an expendable. The suborbital RLV efforts are going at low costs from the other direction. First prove that suborbital spaceflight can be done routinely, even several times a day, with highly robust fly-return-refuel-refly rocket vehicles. Then go to orbital systems. All of the suborbital RLV projects have orbital designs in mind, and some have them at quite an advance stage. None of their plans require any new physics or changes to the rocket equation.

Airliners once all had four engines. This was not because they needed all that power but because it was quite possible for one or two engines to die out during a long flight. It was not new physics that led to two engine airliners. It was the relentless pursuit of high reliability and robustness in jet engines, which are extremely complex machines. There are no fundamental reasons blocking rocket vehicles from eventually attaining similar high levels of robustness and reliability.

Yes, during the past decades there has been lots of talk about cheap access to space but the actions necessary to bring it about did not take place. Incremental, systematic development was required to overcome the many technical challenges involved in spaceflight. The DC-X, for example, planned for a series of vehicles leading to orbit (TSTO if SSTO was not possible) but was sidetracked by the X-33 fiasco. Many private projects got derailed by financial issues totally unrelated to the technical viability of their vehicles. The current NewSpace projects, ranging from the suborbital guys to SpaceX, are the first to have the financial and technical prowess to accomplish their goals.

Those 1970s plans for using lunar resources were based on an expectation of low cost space access with the Shuttle. If the cost to LEO remains at $5K/kg as Dr Spudis predicts, then his plans for a "space-based economy" generated by utilizing the Moon will go nowhere as well and be remembered as nothing but "hype". I'm quite confident, though, that such a lunar industry will in fact happen someday because I know that rocket pioneers will persevere and eventually achieve "LEO on the Cheap".

Phoenix Integration will offer a "Webinar" on Tuesday, August 24th on Analysis of Alternatives and Conceptual Design of a Two-Stage to Orbit (TSTO) Vehicle.

The work presented in this webinar will demonstrate conceptual design analysis of a Two Stage to Orbit (TSTO) Reusable Launch Vehicle (RLV) model to determine feasibility for future studies. An analysis of alternatives will be conducted within PHX ModelCenter for sizing and cost analysis, specifically featuring data visualization, trade studies, and design of experiments. PHX ModelCenter® offers unique trade study, visualization, and optimization tools specifically designed for AoA and providing a decision support environment for performance, cost, and operational assessments.

I found the X-37B article in Air & Space Magazine that Jim Oberg referenced in which some details are given about the TPS and other aspects of the vehicle: Space Shuttle Jr. : After 2010, the only spaceplane in the U.S. inventory will be the Air Force's mysterious X-37 - by Michael Klesius - Air & Space Magazine - Jan.1.10.

The TPS uses

silica tiles impregnated with the latest version of Toughened Uni-Piece Fibrous Insulation (TUFI), some of which have flown on the shuttle since the 1994 mission STS-59. The tiles will provide most of the thermal protection for the X-37's underside, and are more durable than earlier shuttle tiles, which have been pocked by debris as light as paint chips. In a TUFI tile, the surface material permeates the underlying insulation, which supports and reinforces the outer surface and renders it more resistant to impacts. In contrast to the shuttle's older, more rigid glass-fiber composite tiles, TUFI tiles have a porous nature that stops cracks from spreading.

The X-37 will also demonstrate a new-generation Conformal Reusable Insulation blanket technology, which provides better protection for top surfaces, along with a hard, smooth finish that produces less drag than the shuttle's 1970s-era thermal blankets.

The X-37's most notable thermal advance is on the wing leading edge. On the shuttle, that vulnerable area was covered with reinforced carbon-carbon; the X-37 uses a different material, called TUFROC, for Toughened Uni-piece Fibrous Reinforced Oxidation-Resistant Composite. TUFROC (pronounced "tough rock") was developed at NASA's Ames Research Center in California by a group led by David Stewart, who has worked on thermal protection systems since the shuttle program.

Stewart explains that during reentry, heat is generated not just by friction of the vehicle against the atmosphere, but also by atoms on the surface recombining. In the shuttle's case, the carbon-carbon oxidizes. As the name implies, the new material resists oxidative damage. The surface of the shuttle's tiles heats up very fast because the insulator's high-density coating is very thin. TUFROC's surface material is thicker, and therefore takes longer to heat up. And the new material will reduce weight, which will enable the spaceplane to carry more payload.

Valtteri Maja analyzes the Air Force rocket-back booster project mentioned here earlier: The US Air Force Tries To Do Reusables - Gravity Loss.

This sounds like a fascinating project, especially considering that it involves real hardware demos. Hope we hear something about the project at this meeting.
/-- Doing a 180 - AFRL's Rocket-back Pathfinder - Ares/Aviation Week - Apr.7.10
/-- Air Force wants reusable rocket ships: Air Force also works with NASA to develop reusable rockets - NetworkWorld.com - Apr.7.10

Here's the solicitation info:
/-- Reusable Booster System (RBS) Pathfinder - Federal Business Opportunities: Opportunities
/-- Reusable Booster System (doc)

The latter document includes the following:
"Brief Program Summary: The Air Force has identified the Reusable Booster System (RBS) concept as a promising approach to meet its future spacelift needs. The RBS consists of an autonomous, reusable, rocket-powered first stage with an expendable upper stage stack. The reusable first stage launches vertically and carries the expendable stack to the staging point. From the staging point, the reusable first stage returns directly to the launch base, landing aircraft-style on a runway.

To return the booster to the launch base, the Government is considering a rocket-back maneuver. The rocket-back maneuver consists of reorienting the vehicle so that it can use its main rocket engines to accelerate the vehicle back towards the launch site. The return to launch site maneuver is completed with an unpowered reentry and gliding flight and landing.

More about the secretive X-37 project and what its mission(s) might be: U.S. Air Force Launches Secret Flying Twinkie - IEEE Spectrum - Mar.31.10.
===
Rand Simberg has doubts about the goals for the Indian RLV-TD vehicle, which resembles the X-37 and which I mentioned the other day: A Reusable Indian Rocket? - Transterrestrial Musings - Mar.31.10.

As a commenter there notes, the RLV-TD is a demonstrator for the Avatar, which is a vehicle project that has been around since the late 1990s.

Kirk Sorensen provides another way to show mathematically that single-stage-to-orbit (SSTO) is hard: A Simple Modification of the Rocket Equation - Selenian Boondocks.

Kirk relates this to the failure of the X-33 but that was a suborbital project that was canceled after it overran its budget following various delays and problems, especially with construction of the peculiar shaped composite fuel tanks (see X-33/VentureStar - What really happened - NasaSpaceFlight.com - Jan.4.06). On the other hand, the SSTO VentureStar, which was supposedly going to follow the X-33, was certainly a step too far.

The X-33 was a hijacking of a plan initiated with the DC-X project at SDIO to develop fully reusable launch vehicles incrementally in a low-cost, small team, X-project "fly a little, break a little" manner. A series of VTVL vehicles would go to higher and higher altitudes and demonstrate capabilities such as full reusability and fast turnaround between flights. The emphasis would be on overall system capabilities and low cost operations, and not on bleeding-edge technologies. If eventually SSTO with a useful size payload proved too hard, they would still be in excellent shape for a robust two-stage-to-orbit (TSTO) system.

The DC-X fulfilled all of its goals (and at a cost that was about ten times below the standard procurement cost model) and should have been followed by a larger DC-Y. However, NASA took over the program and the DC-X became the DC-XA, which later was destroyed on a test flight, and the DC-Y became the X-33. Instead of a system that could "fly a little" and take the next step in RLV development, the X-33 turned into a program to develop various component technologies like composite multi-lobed LH2 fuel tanks and aerospike engines, none of which were essential to a practical low cost RLV.

Thankfully, companies like Armadillo, Masten Space, Blue Origin, [TGV Rockets] and XCOR are now following the incremental, low-cost, DC-X style approach to development of RLVs. They all have orbit as their long range goal and TSTO designs will almost certainly be the way they get there.

Stephen Metschan of the DIRECT project responds to Rand Simberg's recent article in The New Atlantis: A Defense Of DIRECT - Transterrestrial Musings.

I'm surprised that Metschan claims, or at least implies, that the higher mass fraction of RLVs versus ELVs necessarily means that it will cost more to put mass into orbit with RLVs. Physics determines the mass fraction but it is engineering that determines whether a RLV can fly once every three months or once a week or once a day or multiple times per day. It is engineering that determines how many people it will take to do a turnaround and that in turn determines the all important operational costs. Physics certainly does not in principle prevent a RLV over multiple flights from putting as much or more mass into orbit and at lower cost than a big expendable.

So far, one partially refurbishable vehicle has been developed and it was far too big, fragile, and complex to reduce costs, especially operational costs. I think the private suborbital projects will move RLV technology forward towards highly robust vehicles that can carry out multiple space flights per day. What they learn from this will eventually lead to similar capability for orbital space vehicles. The money they make from suborbital tourism and other markets will help to fund them. However, there are various ways that NASA could contribute to the development of this technology, as in the way NACA helped with development of aviation technology. Throwing money away on yet another big expensive throwaway vehicle is not one of the ways.

With respect to Metschan's political point, I've certainly noted here a few times that the choice may be between funding a dumb NASA launch system or no funding at all. It would be a shame, though, if it is impossible for NASA centers to focus on projects like fuel depots that actually contribute to the development of real spacefaring and also keep lots of people employed and their Congressional delegations happy.

Regarding launch vs spacecraft costs, as long as space access opportunities are infrequent and expensive and there is no in-space infrastructure to fix problems and/or to return components (if not whole spacecraft) to the ground, there will be a strong inclination to build spacecraft that are costly and uni-modular. They are this way because they must have multiple redundancies, contain expensive components, and endure endless testing all to insure they survive launch and last long enough in space to complete their function, untouched ever again. Low cost space access will change this paradigm. Design and operations of spacecraft will change substantially as they become modular, serviceable, and repairable.

Jonathan Amos lauds the potential of the Reaction Engines Skylon vehicle. The UK spaceplane aiming to go to a new level - Spaceman/BBC .
===
XCOR has been expanding some of the resource material on its web site. This includes more information and some excellent photos about the XCOR Prototype X-Racer.

Scramjets remain the space access technology of the future: Scramjets promise space travel for all - New Scientist - July.22.09.

I'm all for scramjet research and it appears that military applications will lead to continued funding for it. However, these sorts of articles always exaggerate the shortcomings of rockets and overstate the advantages of air-breathers for space access. In a New Scientist blog post a few months ago, Henry Spencer gave the counter argument: Rockets, not air-breathing planes, will be tomorrow's spaceships - Short Sharp Science/New Scientist - Mar.4.09.

I'll also note that the above article begins oddly with a focus on the SS1 and SS2. The SS1 flew only twice three times to space because owner Paul Allen wanted it to go to the Smithsonian. Burt Rutan also wanted to get going on the WK2/SS2 project, which has taken longer than hoped but this has been mostly driven by the much larger passenger capacity than the SS1. The accident had some affect but I doubt they would have had the vehicles flying sooner even if it had not happened. The WK2 alone is a major vehicle project and would take most companies several years to develop.

With respect to the relative difficulty of suborbital vs orbital, as I've noted before, orbital is hard but so is developing a space vehicle that is truly reusable. That is, one that can fly, return, refuel, and re-fly multiple times a day. Creating such a challenging system first for suborbital is a perfectly reasonable approach to developing the robust hardware, operational techniques and financial wherewithal to build a fully reusable vehicle that is capable of reaching orbit. If getting to orbit on a shorter development time scale is the priority, then a Falcon 9/Dragon is a perfectly reasonable approach.
===
The magazine includes this image gallery: Spaceplanes and scramjets: A 50-year history . The subtitle, "Engineers have struggled for decades to make reusable spaceplanes " is misleading if it implies that there have been continuous efforts over decades to build spaceplanes. Following the Shuttle development in the late 70s, for example, NASA did no new vehicle development until the mid-90s when it half-heartedly carried out a handful of X-projects, none of which it stuck with and finished. The US DoD made a significant push with the National Aerospace Plane (NASP) in the late 80s early 90s, but it failed and was followed by the low level scramjet R&D as discussed in the above article. Until this decade, private efforts to develop rocketplanes never had the funding to accomplish much.
===
The magazine also has this article about the VASIMR project: Ion engine could one day power 39-day trips to Mars - New Scientist - July.22.09. Seems like a fine research project but as the article points out at the end, the engine needs a nuclear reactor to power it over those 39 day trips to Mars. Currently, there is no serious space nuclear reactor development program going on in the US.

Jon Goff continues a series of discussions about reusable launch vehicles with an examination of where RLVs have big advantages and where they do not: RLV Markets Part II: The Black Aluminum Analogy - Selenian Boondocks.

While the Rocket Racing League has slowed down during the recession (see Alan Boyle's update on the RRL mentioned here awhile back), the Red Bull Air Race circuit seems to be doing well as it heads into the seventh season: Red Bull Air Racing Season Kicks Off In Abu Dhabi: Defending Champion Hannes Arch Wins First Race - Aero-News Network. Here are some videos of the events. Sure seems like rocket racers could be at least as exciting if not a lot more so.
===
A posting on the aRocket forum pointed to an interesting paper by David Ashford from a couple of years ago: New Commercial Opportunities in Space - Space Future - Feb.07. (I see I linked to it back in 2007.) He reviews various commercial space applications and then focuses on how development of a series of increasingly capable spaceplanes could make them viable.

At the end he speculates on what might have happened if development of the British Saunders-Roe SR.53 rocket/jet combo powered vehicle had not been canceled.

If it had entered service, the RAF would soon have had a mature rocketplane with long life and rapid turnaround. With straight-forward development, the SR.53 could have had sub-orbital performance. Indeed, when it was cancelled as a fighter in 1958, Saunders Roe did propose a space research variant(17).

A commercial development could have been built to carry passengers on space experience flights, much as now planned by Virgin Galactic using the US SpaceShipTwo. Thus, routine suborbital flights could have been achieved by the late 1960s. With economies of scale and maturing technology, the cost per seat would eventually have approached that of a long-range business jet at just a few thousand pounds. Orbital spaceplanes could have followed a few years later, probably based on one of the 1960s European Aerospace Transporter projects.

If a development along these lines had happened, spaceflight would have evolved naturally into an everyday and widely affordable business with large new commercial markets, especially space tourism. Present spaceplane initiatives from the private sector offer the prospect of catching up rapidly with what might have been.

If these conclusions seem unduly optimistic, this is perhaps due to the divergence over the years between aviation and spaceflight. Operating aeroplanes is so different from operating ballistic missiles and expendable launchers that aviation and spaceflight have developed into businesses with little overlap. An opportunity to merge aviation and spaceflight (as far as low Earth orbit) arose in the 1960s, when spaceplanes first became feasible, but was not taken.

A couple of interesting articles about observations of high temperature air flow along the bottom of Discovery during its re-entry:
/-- Tripping the boundary layer: shuttle experiment for entry - Spaceflight Now - Mar.28.09
/-- Tripping the boundary layer: shuttle's re-entry experiment - Spaceflight Now - Mar.29.09

The February issue of Scientific American has an interesting overview of electric propulsion and they have made it available on line: The Efficient Future of Deep-Space Travel--Electric Rockets: Efficient electric plasma engines are propelling the next generation of space probes to the outer solar system - Scientific American - Feb.09 (via spacetoday.net). See also the Video.
===
Rocketeers.co.uk notes that Reaction Engines Ltd has posted their News Update for January 2009. Items include a statement that an internal study shows that space based solar power would become economically feasible when the Skylon reduces launch costs

by a factor of 50-100 compared to today's expendable launchers and about 5 compared to reusable TSTO rockets.

===
They also point to a new website devoted to Project STERN (Static Test Expansion deflection Rocket Nozzle).

The NASA - Armadillo award ceremony is now was on NASA TV. Here are my notes:

Doug Comstock of NASA's Centennial Challenges program gives an introduction.

Mike Griffin comes up to give some comments.
- Supporter of prize programs in areas where the technology is within reach.
- Not a supporter of multi-billion dollar prizes for big efforts like a human landing on the Moon as some members of the "chattering classes" have suggested.
- The LLC produced a 6 to 1 payoff in terms of hours worked and money invested by the various teams.
- Hopes with the Challenges to attract people and organizations that wouldn't normally participate in NASA contracting.
- Peter Homer who won the space glove competition is a great example of this.
- Scaled Composites is building on its X PRIZE win to build a commercial suborbital space vehicle.
- NASA not in the business of space tourism but should encourage those who are.
- Should encourage commercial partnerships.
- NASA plans to announce the selection of the commercial ISS cargo contract winners later this month.
- Mentions success of SpaceX Falcon 1 launch and the 9 engine firing and notes Orbital's progress in meeting their Taurus II milestones.
- NASA does have a way to get crew to ISS during the gap but doesn't for cargo; so that is the critical issue. Don't want to raid other accounts to pay for crew capability.
- Ares I/Orion is not intended for LEO access. Only to be used if the commercial guys can't fulfill their promise.
- Will purchase seats for experiments, tests and perhaps astronaut training on suborbital vehicles.
- Lots of discussion within NASA on how to take advantage of the new suborbital RLVs.
- Will leverage these capabilities to improve current science efforts, not simply to subsidize such projects.
- Describes ZERO-G contract for micro-gravity projects. Going well and plan eventually to shift all parabolic work to them. [Need to double check this. I think I understood him correctly.]
- Starts describing the LLC program.
- Edison said invention comes from 1% inspiration and 99% perspiration. Here to recognize the inspiration and perspiration of Armadillo Aerospace.
- Griffin leaves the stage. Said he must leave the news conference due to other obligations.

They show a X PRIZE video about the Northop Grumman Lunar Lander Challenge starts.

Doug comes back to the podium. Notes that Ken Davidian, who led the CC program previously, was in the audience.

George Nield of the FAA Commercial Space Transportation Office, comes to the podium.
- AST has a dual role of insuring public safety and promotion of the industry.
- The LLC effort displayed both roles.
- Setting up the event involved permits for the competitors, setting up the airport on short notice, resolving insurance issues, etc.
- Of the 5 experiment permits issued so far, Armadillo has gotten 3 of them.
- Look forward to working with Armadillo as they move towards crew vehicles and sending payloads to orbit.

Peter Diamandis comes up the podium.
- Thanks NASA for supporting prizes.
- NASA will be a major beneficiary of the technologies developed via prize programs.
- Helps make the most of NASA dollars.
- Need to get space technology development on a Moore's Law growth curve.
- Thanks Northrop Grumman for its support.
- Describes the mad scramble to move the event from Holliman AFB, which had withdrawn its support at the last minute, to the Las Cruces Airport in just 30 days.
- Thanks the NM partners and the other LLC teams.
- Notes that Google has a rep in the audience and mentions the Google Lunar X PRIZE

Carl Meade of Northrop Grumman comes to the podium
- Congratulates John Carmack directly. Says he can come down and join NG anytime.
- Hails entrepreneurial spirit of the teams and the X PRIZE program.
- Looks forward to awarding Level 2 prize next year.

Doug returns to the podium:
- John Carmack invited to the stage.
- The BIG check is given to John.

John Carmack makes some comments:
- When first began looking at rocketry, he decided that rockets are not all that complicated.
- Today he still believes they are not complicated, as compared to big software projects, but they are not easy.
- Making rockets work reliably every time is really hard.
- Effort in 2006 was a good practice and felt very confident going into 2007 LLC
- E.g. they did the complete LLC flight profile in 2007 in Oklahoma
- However, had several unexpected problems during the 2007 LLC that prevented winning it.
- In 2008, they had been focused on rocket racing and then last minute permit problems reduced testing time. (He thanks George Nield for his help in helping them overcome these problems.)
- Level II effort ran into a valve problem that has now been fixed.
- Could do Level II today.
- Comments on general LLC program.
- Believes it has been a huge success.
- Half a dozen serious teams with lots of money and work time put into their vehicle development.
- Real vehicles have been built and test flown by these teams.
- Can now find people to hire who have skill sets not available a few years ago.
- The competition has been everything that was initially hoped for it.
- Not winning in 2006 and 2007 probably benefited the program since it gave time for others to participate.
- Expects to win the $1M Level II first prize but LLC activities could go on for another 2 years before the second prizes are won.

The news conference closes.

Rob Coppinger reports on preparations of the WK2 for its first flight: EXCLUSIVE PICTURE: Virgin Galactic's WhiteKnightTwo is prepared for flight - Flight Global. (I've also been hearing other reports like this one in the past month or so about WK2 being seen outside of its hangar.)

Here he reports on evidence that the WK2 was missing interior systems when it rolled out last July: Did Virgin Galactic do a Boeing 787 with the WhiteKnightTwo? - Hyperbola
===
Rob also posts a video of a talk at the recent IAC 2008 about Expansion Deflection rocket nozzles: VIDEO: UK nozzle could enhance SSTO propulsion - Hyperbola.

More about the ED nozzles can be found here:
/-- Project STERN - www.airborneengineering.co.uk
/-- First firing of STERN rocket - Bristol University - Mar.26.08

John Hare considers an unusual design for propellants tanks and for the rocket vehicle to hold them : Toroid Tanks - Selenian Boondocks - Oct.14.08.

As noted in a comment there, see also these Lenticular Vehicle concepts studied by NASA and the Air Force during 1959-1964.

Jon Goff addresses the issues of flight rates and markets for RLVs : RLV Markets Part I: The Importance of High Flight Rates - Selenian Boondocks.
===
John Hare discusses attitude compensating nozzles and proposes a particular design: Saddlespike Nozzle - Selenian Boondocks

John Hare joins the crew at Selenian Boondocks. Here are his first couple of postings:
/-- Admission of Failure
/-- SpaceX and RLV

In this article - Elon Musk [in a list of the 75 most important people of the 21st Century ] - Esquire - Oct.08 (link via spacetoday.net) - Elon Musk points out the huge impact on launch costs if they can succeed in reusing the Falcon 9 stages:

We're making progress. If we succeed in recovery and reflight of our Falcon 9 rocket, which carries 11 tons of payload into orbit, it will be the first fully reusable orbital rocket and one of the most significant developments since the dawn of rocketry. At $35 million to manufacture, it's already four times cheaper than comparable single-use vehicles from Boeing or Lockheed. However, since Falcon 9 costs only $200,000 to refuel (and reoxidize), an efficient refurbishment and launch operation would allow the production costs to be amortized over many flights. This has the potential to bring the per-launch price down to about $1 million, a hundredfold improvement over current costs. And if that happens, life will become sustainably multiplanetary in less than a century.

According to the Falcon 9 specs, it can place 12,500 kg (27,500 lb) into LEO (200 mile circular). So at $1M for a flight, that means $80/kg ($36/lb).

Of course, they must start by proving the Falcon 1 first stage can be reused after it parachutes into the ocean and floats there until recovered. Then this must be demonstrated for each of the two Falcon 9 stages. Reusing the second stage seems especially challenging. For flight four of the F1, they did not add the necessary thermal protection to allow for recovery but they will do that for flight 5.

At worst, they should be able to reuse major parts and materials from the stages. That wouldn't get them to $1M per flight but should cut costs to some degree.