A severe staffing and leadership crisis within the U.S. Department of Commerce’s Bureau of Industry and Security (BIS) has triggered a major logjam in export license approvals—including critical permits for AI chips and semiconductor manufacturing gear destined for key international partners. Sources cite turbulent management under BIS Undersecretary Jeffrey Kessler, and a wave of resignations and buyouts among experienced staff, as the cause of the paralysis.
This unprecedented backlog is threatening billions of dollars in export revenue, as U.S. companies are unable to fulfill hardware orders. High-profile clients like Nvidia remain unable to ship its H20 AI chips to China despite earlier assurances, disrupting deliveries and undermining confidence in U.S. competitiveness.
Read Also: https://atoztimes.com/u-s-eu-trade-pact-brings-relief-for-tech-and-auto-sectors/
Industry insiders describe the licensing backlog as the longest in more than three decades. Not only are key approvals stalled, but critical regulatory revisions—such as dismantling complex Biden-era rules—have yet to be finalized. Drafts remain unpublished, and open roles essential to export compliance remain vacant, deepening institutional dysfunction.
Under Kessler’s leadership since March 2025, the bureau reportedly adopted stringent internal controls: staff were instructed to log all meetings in shared spreadsheets, limited from interacting directly with industry representatives, and faced increased hurdles even attending inter-agency consultations.
Officials at Commerce, however, defended the approach, stating that BIS will no longer rubber‑stamp license applications that raise grave questions of national security and that Kessler enjoys Department Secretary Howard Lutnick’s full confidence.
Export-dependent companies are increasingly turning abroad. Sean Stein of the US‑China Business Council warned that segments like semiconductor manufacturing equipment are at a standstill, prompting Chinese and other international buyers to seek alternative suppliers—eroding U.S. market share with each passing week.
Trade consultants also confirm lingering denials and inconsistencies: among twenty applications submitted by one firm, four licenses were unexpectedly denied—even after months of waiting.
The downturn is especially acute for companies like Nvidia and AMD, which had planned billions in revenues from approved trade with D:5 countries, including China. Industry losses are mounting amid systemic uncertainty.
Earlier in 2025, BIS rolled out its new interim framework—the “Framework for Artificial Intelligence Diffusion”—which instituted sweeping regulations around advanced AI chips and closed-source model weights. These measures drastically expanded licensing requirements and defined new global thresholds for export eligibility.
Specifically, the rule imposes licensing obligations for advanced integrated circuits and introduces export controls on AI model weights where models exceed defined computational thresholds—even if developed abroad using U.S.-controlled chips.
While the Biden-era version was meant to take effect by mid‑May 2025, the Trump administration has paused implementation pending a streamlined replacement. Innocuous pending rule revisions—intended to simplify licensing and better align restrictions with allied vs adversarial usage—remain unsigned months later.
Despite assertions of policy reform, neither new rules nor staffing replenishments have materialized—deepening internal delays and leaving the licensing machinery stuck in limbo.
At a time of regulatory uncertainty, the AI talent market is roaring ahead. Meta recently offered breakthrough AI researcher Matt Deitke—a 24‑year‑old who dropped out of a University of Washington Ph.D. program—a record-setting $250 million compensation package over four years, roughly averaging $62 million annually, with $100 million in the first year alone. The move underscores the competition among tech giants like Meta, OpenAI, Microsoft and Google to lock in elite researchers.
To put that in perspective: Deitke’s package is more than 327 times the inflation‑adjusted income earned by physicist J. Robert Oppenheimer during the Manhattan Project, highlighting a dramatic shift in the economics of scientific talent.
This juxtaposition—exploding talent dollars on one side, chronic bureaucratic overload on the other—paints a picture of structural disconnect. U.S. firms and researchers continue to lead AI breakthroughs. Yet delays in permitting and licensing risk ceding manufactured components and systems to rival nations, eroding downstream economic influence even as upstream R&D accelerates.
The current chaos may confer short‑term protection on domestic production, but long term, it threatens to weaken America’s grip on export‑driven revenue streams and global AI infrastructure leadership.
Industry groups are calling for urgent action: prompt staffing of vacant roles, clearer guidance on licensing standards, and swift publication of revised export rules that maintain security while restoring export flow.
Without decisive resolution, business leaders warn that U.S. firms will continue losing ground to foreign suppliers—undermining both commercial returns and the broader strategic position of the country in global AI supply chains.
