Pentagon's B-2 Cost History Echoes in $58.5 Billion AI Budget Request
The B-2's final price tag of $2 billion per plane offers a measuring stick for Washington's latest defense technology spending surge.
The Pentagon's fiscal year 2027 budget requests $58.5 billion for artificial intelligence and combined joint all-domain command-and-control, a figure that lands with more context when set against the history of what Washington once paid to fly a stealth bomber.3
Gold closed at $4,346.47 per troy ounce on Monday (2026-08-10). At that price, a 70-ton B-2 Spirit would be worth roughly $10 billion in gold. As Time reported in 1989, estimates from then-Congressman Les Aspin already put the cost per plane at over a billion dollars — meaning each aircraft was worth more than its weight in the metal. That comparison aged poorly for the program's defenders.3
The B-2 program did not hold its initial ambition. Total acquisitions were cut to just 20 planes, with a then-total program cost of $44.7 billion, according to War on the Rocks reporting published Monday (2026-08-10). The resulting unit cost — approximately $2.235 billion per aircraft — gave rise to the aircraft's enduring label: the $2 billion bomber.3
What made the B-2 case instructive was not just the cost overrun. It was the sequence: classified development, limited congressional visibility, then a slow public reckoning as oversight mechanisms caught up. Congress began expanding scrutiny of stealth programs through riders to the annual defense authorization acts for 1988 and 1989, directing that certain features of the B-2 and other weapons systems remain classified while other budget information became subject to review.3
That pattern is now visible in how Washington is approaching its next generation of high-cost defense technology. The FY2027 AI and command-and-control request of $58.5 billion is not a single program — it spans a portfolio that includes Project Maven and its descendants. War on the Rocks noted Monday (2026-08-10) that "there are many more Mavens," suggesting the AI spending surge is distributed across programs with varying degrees of public accountability.3
The Golden Dome missile defense system is the clearest parallel. In early May 2026, the Congressional Budget Office put the cost of the program at $1.2 trillion — a figure that stood in sharp contrast to the $185 billion estimate Pentagon senior officials provided to the House Subcommittee on Strategic Forces in April 2026. The CBO's own assessment found that even at $1.2 trillion, the resulting system "would not be able to fully counter a large attack of the sort that Russia or China might be able to launch."1
Defense industry attention has nonetheless mobilized quickly. The April 2026 award of $3.2 billion to twelve companies for concept development and demonstrations of space-based interceptors has spawned significant contractor activity, according to the Atlantic Council.1
The mineral supply chain behind these systems adds a further cost variable that was absent from the B-2 era. Washington has spent approximately $46 billion since the Trump administration took office trying to build a critical mineral supply chain less dependent on China, according to OilPrice.com reporting from early August 2026 (2026-08-04). Tungsten — essential to weapons systems and hardened components — is a specific pressure point. China controls 80% of the global tungsten market, and prices for the metal have risen 600% over roughly the past year, driven by export curbs and wartime demand. A deposit discovered in the United States holds an estimated 1.78 million tons, five times larger than the next domestic tungsten deposit, with a value estimated at $152 billion by its developer, 3PL, because of surging prices.2
Whether that deposit moves into production fast enough to matter for programs now in concept development is unclear. Mining timelines in the United States rarely compress on demand.2
The FY2027 AI budget request will draw the same congressional scrutiny that eventually caught up with the B-2. The question for defense appropriators is whether the oversight architecture has improved enough since the 1980s to identify cost divergence before acquisition commitments harden. Pentagon scrapped a $300 million lithium tender for defense stockpiles this year, a sign that procurement on critical inputs can stall even when the strategic case is clear.2
The CBO's $1.2 trillion Golden Dome estimate — versus the Pentagon's $185 billion — is the nearest live data point on that gap. It is a spread of roughly $1 trillion, and concept development contracts are already signed.1