The U.S. Commodity Futures Trading Commission has lost 24% of its workforce since early 2025, dropping to 535 employees — a 15-year low. Chairman Michael Selig, the agency's sole sitting commissioner, told CoinDesk on April 27, 2026 that the CFTC is deploying Microsoft 365 Copilot across its rema...
"If I was a different person I would launch a crypto scam right now, because there's no cops on the beat." — Former CFTC Chief Trial Attorney (26-year veteran, laid off in 2025), quoted in Barron's
The U.S. Commodity Futures Trading Commission has lost 24% of its workforce since early 2025, dropping to 535 employees — a 15-year low. Chairman Michael Selig, the agency's sole sitting commissioner, told CoinDesk on April 27, 2026 that the CFTC is deploying Microsoft 365 Copilot across its remaining staff and building proprietary AI tools to review crypto registration applications, flag incomplete filings, and surveil trading activity. The AI systems will auto-reject materially incomplete submissions and prioritize well-formed applications, effectively delegating initial triage to machine review.
The bet carries high stakes. The CFTC's enforcement division secured $17.1 billion in monetary relief in fiscal year 2024. In FY 2025, that figure collapsed to less than $10 million — a 99.9% decline — as staffing cuts, leadership turnover, and a philosophical shift away from "regulation by enforcement" gutted the agency's capacity. The Chicago enforcement office, birthplace of modern futures regulation, went from 20 trial attorneys to zero. Meanwhile, the CFTC's regulatory mandate is expanding: the March 2026 SEC-CFTC joint interpretation designated approximately 70% of traded digital assets as commodities under CFTC jurisdiction, and prediction market volumes now run in the tens of billions of dollars monthly.
Whether AI-augmented supervision can compensate for a workforce that is smaller than at any point during Trump's first term remains an open question. The CFTC is simultaneously building new rulemaking infrastructure (Innovation Task Force, prediction market ANPRM), litigating preemption lawsuits against five states, and attempting to rebuild its enforcement ranks — all with a single commissioner and a $410 million budget request that Congress has not yet approved.
The CFTC entered 2025 with approximately 700 employees. By February 2026, that number stood at 535, according to agency records cited by CNN. The reductions occurred under Acting Chair Caroline Pham through layoffs, buyout offers, and early-retirement incentives pushed by the Department of Government Efficiency (DOGE).
The enforcement division absorbed disproportionate losses. Chairman Selig's FY 2027 budget request seeks funding for 108 enforcement staff — down from 140 filled positions in FY 2025, a 23% reduction in the unit responsible for prosecuting market abuse. The Chicago Division of Enforcement, which participated in most major CFTC cases since the agency's founding in 1975, lost all 20 of its enforcement attorneys. Four of its five chief trial attorneys were dismissed on a single day in July 2025, along with the division's deputy director, per Barron's reporting from February 2026.
The SEC, by comparison, employs roughly six times as many staff. Selig is the only sitting commissioner on the CFTC's five-member panel; the other four seats are vacant following departures throughout 2025, with no nominations pending.
Even if Congress approves Selig's request for $410 million and 650 full-time employees, the CFTC would remain smaller than during most of Trump's first term.
Selig described three categories of AI deployment during his April 27 interview with CoinDesk and his April 16 testimony before the House Agriculture Committee:
1. Registration Application Review AI tools will screen crypto registration applications for completeness. "AI tools can be used to review the applications, flag certain things for the staff, make their jobs easier, make it much faster for them to provide feedback and also reject certain things that aren't materially complete," Selig stated. Incomplete submissions — those with blank fields, inadequate descriptions, or obvious errors — will be deprioritized or pushed to the back of the processing queue without human review.
2. Market Surveillance The agency is developing in-house systems to analyze swap data and trading activity. Selig said the CFTC has created tools that "can help staff draw conclusions about some trades," though he provided no detail on the underlying models, data sources, or validation processes.
3. General Productivity All remaining CFTC staff are being trained on Microsoft 365 Copilot. Staff use the tool to draft memos, reports, and presentation materials. The agency is also using AI to review the nearly 1,200 public comments submitted on its prediction market rulemaking, with the comment period closing April 30.
Selig has not disclosed the vendor or architecture of the proprietary surveillance tools, whether the AI systems will be subject to independent audit, or what error tolerances the agency considers acceptable for automated application rejection.
The collapse in CFTC enforcement output is stark:
| Metric | FY 2024 | FY 2025 | |--------|---------|---------| | New enforcement actions | 58 | 13 | | Total monetary relief | $17.1 billion | < $10 million | | Whistleblower tips received | 1,700+ | Not disclosed | | Whistleblower awards | $42 million (15 awards) | Not disclosed |
The FY 2024 total was inflated by one-time resolutions: $12.7 billion from the FTX/Alameda case and $2.7 billion from Binance. Stripping those out, the agency still secured approximately $1.7 billion from 56 other cases — a figure that dwarfs the FY 2025 total.
Under Acting Chair Pham's "back-to-basics" approach in 2025, the Division of Enforcement consolidated its nine specialized task forces into two, closed approximately half of its open enforcement matters, and reduced case filings dramatically. Between October and December 2025, the CFTC filed a single case against three defendants — compared to 29 SEC actions during the same period.
New Enforcement Director David Miller, a former CIA officer and SDNY prosecutor appointed under Selig, announced five priority areas on March 31, 2026: insider trading (including prediction markets), energy market manipulation, disruptive trading, retail fraud, and AML/KYC violations. Miller said the division is "actively hiring" to rebuild capacity, but the budget request tells a different story — 108 funded positions versus 140 in FY 2025.
The CFTC has not published its FY 2025 annual enforcement report. At the April 16 hearing, Selig did not provide a clear answer when asked whether that was his decision.
The CFTC's oversight burden has grown substantially in 2026:
Crypto Asset Jurisdiction. On March 17, 2026, the SEC and CFTC issued a joint interpretation establishing a five-part crypto asset taxonomy: digital commodities, digital collectibles, digital tools, stablecoins, and digital securities. Of these, only digital securities fall under SEC jurisdiction. The interpretation names 16 specific tokens — including BTC, ETH, SOL, and XRP — as digital commodities and states, per Senator John Boozman, that approximately 70% of traded digital assets qualify as commodities under CFTC oversight.
The pending CLARITY Act, if enacted, would codify this framework and formally elevate the CFTC as primary regulator over non-securities crypto trading. The agency is simultaneously processing a backlog of registration applications that accumulated during a 43-day federal government shutdown.
Prediction Markets. The CFTC issued an Advanced Notice of Proposed Rulemaking on prediction markets on March 12, 2026. Monthly volumes on platforms like Kalshi and Polymarket now run in the tens of billions of dollars. Selig has filed preemption lawsuits against Arizona, Connecticut, Illinois, New York, and Wisconsin, seeking declaratory judgments that the CFTC has exclusive federal jurisdiction over event contracts.
The most notable prediction market enforcement action to date involved a U.S. Army special forces soldier who allegedly profited more than $400,000 by trading on Kalshi using advance knowledge of a military operation to capture Venezuelan leader Nicolás Maduro. The CFTC and SDNY announced parallel civil and criminal actions on April 23, 2026.
On March 24, 2026, Selig formed the Innovation Task Force (ITF), led by senior adviser Michael J. Passalacqua. On April 10, five staff members were named:
The ITF covers three domains: crypto/blockchain, AI/autonomous systems, and prediction markets. It operates in an advisory capacity and does not issue binding rules, though it is charged with producing proposed rulemaking in each area. It coordinates with the SEC's Crypto Task Force on overlapping jurisdictional matters.
Criticism has been bipartisan.
Representative Angie Craig (D-MN), ranking member of the House Agriculture Committee, stated at the April 16 hearing that "the agency's workforce is stretched too thin," noting its role as "primary regulator of two of the fastest growing and most volatile markets." She questioned whether the CFTC can fulfill its mission with a single commissioner advancing major rulemaking without bipartisan input.
Congresswoman Alma Adams (D-NC) raised concerns about Selig's "willingness to advance major new regulations without a full complement of commissioners" and noted that the FY 2025 enforcement report remained unpublished six months after fiscal year end.
Representative Eugene Vindman (D-VA) told CNN: "I think he is somebody we can work with. But cutting staff at a time when the mission is growing is not a good idea. This is creating real vulnerabilities."
Committee Chairman Glenn Thompson (R-PA) acknowledged the growing burden — "We're putting a lot on your plate with digital assets" — and said he and Craig would send a letter to the White House urging prompt commissioner nominations.
David Slovick, a former Chicago enforcement attorney, told Barron's: "Chicago is the spiritual home of the futures markets; it's where it all began. To wipe out the enforcement staff in a place like Chicago sends a very bad signal to market participants."
Selig's position: "There are no gaps in our ability to fulfill our mission." He told lawmakers the agency is "running more efficiently and effectively than ever before."
The CFTC's AI strategy introduces several categories of risk:
Operational. Automated rejection of registration applications without human review creates potential for false negatives — legitimate applicants deprioritized by pattern-matching errors. The CFTC has not disclosed error rates, appeal mechanisms, or whether applicants will be notified that AI, not a human, flagged their submission.
Legal. Administrative law generally requires reasoned decision-making by agency officials. Delegating initial application triage to AI tools may face legal challenges if rejected applicants argue their filings received inadequate human review.
Surveillance Accuracy. AI-driven trade surveillance depends on training data, model architecture, and threshold calibration. The CFTC has not published its methodology, validation approach, or false-positive rates. Automated errors that generate enforcement referrals could impose significant costs on market participants.
Capacity Gap. Even with AI augmentation, 535 staff overseeing commodity derivatives, 70% of crypto assets, prediction markets, and agricultural commodities represents a mandate-to-resource mismatch that software alone may not resolve. The agency's request for 650 employees in FY 2027 — if approved — would still leave it below its pre-2025 headcount.
Single-Commissioner Governance. Major rulemaking and enforcement policy is being set by a single commissioner, raising questions about institutional legitimacy and durability. A future administration could reverse any interpretation or rulemaking that was not codified by Congress.
The CFTC is attempting to resolve a structural contradiction: its regulatory mandate is expanding while its workforce contracts. The agency's answer — AI-augmented supervision, Copilot-assisted productivity, and a six-person Innovation Task Force — represents an experiment with no regulatory precedent. No major U.S. financial regulator has previously delegated initial application screening to machine review or relied on AI surveillance to compensate for a workforce reduced by nearly a quarter.
The experiment's outcome will shape not just crypto market oversight but the broader question of whether AI can substitute for regulatory headcount across federal agencies. The data needed to evaluate that proposition — error rates, false-positive frequencies, processing time improvements, enforcement referral quality — has not been disclosed. Until it is, the CFTC's AI strategy remains a bet on technology solving a problem created by policy.