For example, in July two enterprising Middle Eastern firms started offering cellphone service in Baghdad, setting up jury-rigged systems compatible with those of neighboring countries. Since the collapse of Baghdad's phone system has been a major source of postwar problems, coalition authorities should have been pleased.

But no: the authorities promptly shut down the services. Cell service, they said, could be offered only by the winners in a bidding process — one whose rules, revealed on July 31, seemed carefully designed to shut out any non-American companies. (In the face of strenuous protests the rules were revised, but still seem to favor the usual suspects.) Oddly, the announcement of the winners, originally scheduled for Sept. 5, keeps being delayed. Meanwhile, only Paul Bremer and his people have cellphones — and, thanks to the baffling decision to give that contract to MCI, even those phones don't work very well. (Aside from the fact that its management perpetrated history's biggest accounting fraud, MCI has no experience in building cell networks.)

For a few days at the end of July, a couple of companies from Bahrain and Kuwait actually did set up a working cell-phone network for the public. But Bremer shut it down right away. According to the briefer sent out to meet the press, the “illegal” Bahrain and Kuwait phone service was interfering with U.S. military communications and the MCI network. The public’s phones were causing problems, in other words, for all those folks in the Bubble. Now we’re told “legal” phones won’t be in operation until mid-November at the earliest.

When a colleague of mine pressed one of the Bubble’s senior officials about this problem the other night, this up-and-coming Washingtonian said he didn’t see what all the fuss was about. “Quite frankly, we are in no hurry to establish a cell net here,” the official confided. “We don’t want the terrorists talking to each other.”

No bidder could be more than 5% government-owned. That meant that nearly every mobile operator in the Middle East and Europe was out of contention. Plus, the three winners would be required to post bonds covering the full cost of construction -- possibly as high as $150 million each. Such demanding terms, says telecom-equipment analyst Jason Chapman for researcher Gartner Inc., "would have made the short list very short."

But a surprising thing happened over the next few weeks. Responding to the outcry from potential bidders, the U.S.-led Coalition Provisional Authority (CPA) twice relaxed its terms. By the time bids were due on Aug. 21, the CPA had lowered the bond requirement to $30 million and agreed to permit up to 10% government ownership in any bidding consortium. That let state-owned carriers enter the running in conjunction with private investors. More than a half-dozen Middle Eastern telcos have thrown their hats in the ring.

Seeking a lightning-fast rollout, the CPA devised a scheme to carve Iraq into three wireless regions. Each will be served by different carriers. After 12 months, the operators are encouraged to invade other territories in the hopes of fostering competition.

The decision to limit the contracts to two years was perhaps the boldest stroke. After that, a new Iraqi government is expected to organize its own tender -- and there's no guarantee September's winners will be picked again. It's an assurance to Iraqis "that we're not giving it away forever," says a U.S. Senate Foreign Relations Committee staffer. Thus, most bidders have lined up Iraqi partners. "Having local investors in a winning consortium is seen as an insurance policy when a subsequent Iraqi government takes over," says Norman Sandler, director of global strategic issues for Motorola Inc. (MOT ).

On the other hand, the two-year term could make it nearly impossible for companies to recoup their investments. Iraq's infrastructure is so shattered that operators may need to build pricey fiber or microwave backbones to connect cellular towers. The threat of sabotage by rebels necessitates expensive security. And market potential is limited by widespread poverty. Motorola figures on only about 500,000 mobile users initially -- roughly a $60 million annual business if customers spend an average of $10 per month. Such small returns suggest that only the very brave -- or foolhardy -- are plunging into the fray.