The U.S. Air Force is expected to offer half of the 14 launches it had anticipated would be suitable for competition from 2015-2017, limiting the near-term opportunities for Space Exploration Technologies (SpaceX) to duel with rival United Launch Alliance.

The service inked a deal in January with ULA for 36-50 Evolved Expendable Launch Vehicle (EELV) cores over the next five years; 36 of them are guaranteed, and 14 were considered potentially open for bids.

However, the fourth launch of the Lockheed Martin Space-Based Infrared System, designed for missile warning, was pulled from those 14 and assigned to ULA to fulfill the requirement to provide the company with 36 cores worth of work, says Maj. Gen. Robert McMurry, who heads the Air Force space procurement office at the Pentagon.

Another launch is experiencing weight growth, exceeding the capabilities of SpaceX, he adds.

Also, five GPS missions slated for 2017 have been shifted beyond that date because satellites already in orbit are lasting longer than expected, relaxing a need to replenish the constellation and also slowing the tempo for launches in the coming years.

One launch for the National Reconnaissance Office (NRO) will be open for competition if SpaceX is certified in time, McMurray says. A request for proposals is expected in May, he says. The NRO is keen to compete the launch, and the Air Force is holding off as long as possible in order to keep the option open.

Last week, Lt. Gen. Ellen Pawlikowski, Air Force program executive officer for space, announced that the first of SpaceX’s three Falcon 9v1.1 missions would count toward the company’s certification requirements: SpaceX must conduct three successful missions of the Flacon 9v1.1 (two of which are consecutive) to be certified. There is still “a lot of work left” for that to be done, McMurray says. If the probability that SpaceX will be certified this year is high, then the launch is likely to be competed, McMurray says.

Meanwhile, SpaceX CEO and chief designer Elon Musk is suggesting that the Pentagon mothball its Atlas V rocket, originally designed by Lockheed Martin, and rely on the company’s Falcon 9 v1.1 and ULA’s Delta IV, designed by Boeing, for boosting national security payloads.

Testifying before the Senate Appropriations defense subcommittee March 5, Musk said the Pentagon does not need three separate rockets to maintain assured access to space. The Atlas V is the most logical vehicle to abandon because it is powered by the Russian-built RD-180 engine, making its availability suspect, he said.

Musk, whose Falcon 9 medium-lift rocket and Dragon spacecraft deliver cargo to the International Space Station under a $1.6 billion fixed-price contract with NASA, is eager to compete for a portion of military and intelligence satellite launches this year. He says competition against incumbent United Launch Alliance (ULA), which manages government missions of Delta 4 and Atlas 5 rockets, would bring down ULA’s prices.

Since 2006, when ULA was formed out of the separate launch vehicle divisions of Boeing and Lockheed, Delta 4 and Atlas 5 launch prices have skyrocketed. Though not yet certified for Pentagon and NRO launches, Air Force Undersecretary Eric Fanning says the recent deal signed with ULA for 36-50 cores over five years shows a dip in pricing, and program cost has dropped by another $1.2 billion in the fiscal 2015 budget request.

Musk repeatedly fingered Russia as a potentially unreliable partner that cannot be considered a long-term source for a U.S. company launching national security assets.

Launch program sources say that between the Atlas V and the Delta IV, the Atlas V offers significant flexibility due to the RD-180 design. But the Pentagon dashed plans more than a decade ago to develop an Atlas V Heavy to hoist the largest payloads, often crafted by the NRO.

Michael Gass, chief executive of ULA, who also testified before the subcommittee, says his company has at least two years of RD-180 engines in inventory in the U.S. ULA has already demonstrated it could reproduce the engines domestically if needed, he added. Left unsaid was the cost of setting up and maintaining a production line for the RD-180 in the U.S.

Gass said he does not mind competition and reminded the subcommittee why certain rigorous practices at ULA were developed under the Air Force’s EELV program. He argued that 15 years ago launches of military and intelligence satellites were fewer and further between, and often delivered late, factors that added to launch service provider costs.

In response, the Defense Department agreed to make annual payments to maintain Boeing and Lockheed production lines, an investment separate from the purchase price of Delta and Atlas launches.

Musk argues the Defense Department could save money by doing away with the practice and rely instead on fixed-price contracts to procure launch vehicle cores.

At the same time, he said, the government could use more expensive cost-plus contracts to pay for unique mission requirements associated with military and intelligence satellites.

To compete for such missions, Musk said SpaceX is willing to abide by whatever regulatory regime the Defense Department puts into practice for specialized missions. He reiterated that the most promising cost-saving scenario is for ULA to phase out the Atlas V and rely on Boeing’s Delta IV and SpaceX Falcon 9 v1.1, and the future Falcon Heavy, for others.

Subscribers to the Aviation Week Intelligence Network should keep visiting AWIN’s 2015 budget page for news, data and analysis of programs and priorities throughout the business day and as the proposal makes its way through Congress. The page can be found at http://www.aviationweek.com/awin/USBudget2015.aspx.