Margin debt hit a record $1.416 trillion in May 2026, amplifying the AI rally. Short covering, not short selling, defined the first half for AI names. The second half will test whether investors can manage the leverage they have built.
The U.S. securities market in the first half of 2026 did not rise on AI optimism alone. Record credit investment and expanded leverage trading amplified the advance. AI lit the fire. Leverage fed the flames.
According to FINRA's monthly margin statistics, customer debit balances in margin accounts reached a record $1.416 trillion in May 2026. That is an increase of 8.5% from April and a 53.7% rise year-over-year from May 2025's $921 billion. The figure is the highest nominal level in the FINRA data series, which begins in January 1997.
Expressed relative to GDP, margin debt now stands at approximately 4.45%. That is also a record high, and nearly double the long-term median of approximately 2.37%, according to GuruFocus data. Net credit balance – free cash accounts minus margin debt – reached a record low of negative $991.7 billion in May 2026. That is the thinnest cushion against forced selling on record, Advisor Perspectives reported on June 24, 2026.
Since 1997, only three prior periods saw margin debt growing at a comparable or faster year-over-year rate: late 1999 to early 2000, mid-2007, and spring 2021. None of those periods aligned precisely with market tops. All were late-cycle or post-surge environments in which equity risk proved more unforgiving than investors assumed, wrote Bob Brinker on Substack in June 2026.
Market capital concentrated into AI-related names. NVDA, Broadcom, AMD, and Micron Technology led the advance. Investors aggressively expanded leverage through margin loans, leveraged ETFs, options, and equity repo structures. Leveraged ETF assets expanded to approximately $220 billion in the first half of 2026. The concentrated nature of these flows amplified market gains further – and amplified the downside risk simultaneously.
On June 5, 2026, a 3x leveraged semiconductor ETF dropped 31% in a single trading session. The episode illustrates the structural violence embedded in these instruments, CryptoBriefing wrote in July 2026.
Short covering, not short selling, defined the first half of 2026 for AI names. Short positions in leading AI stocks declined or triggered covering as prices rose. The structural loop repeated: AI investment expansion, semiconductor earnings surge, results exceeding estimates, short seller losses mounting, forced buying, further price appreciation.
Micron Technology illustrated the pattern. Explosive demand for AI memory chips, supply shortages, and earnings substantially above estimates drove a sharp price surge. The earnings release triggered a reduction in short positions. The stock gained roughly 17% in a single day, adding about $189 billion in market capitalization, Reuters reported. The move was characterized as re-igniting the AI rally. NVIDIA, Broadcom, SanDisk, Western Digital, and Qualcomm moved in parallel.
Hedge fund leverage in 2026 is the highest since the 2008 financial crisis, according to Goldman Sachs Marquee data. The structure differs from the Archegos Capital Management collapse of 2021, which involved catastrophic over-concentration in a single name. In 2026, the distinctive feature is leverage deployed across multi-strategy, relative value, Treasury basis trades, and AI and technology long-short positions, Goldman Sachs said in a January 2026 report. Primary dealers are carrying equity repo exposure that now exceeds $220 billion, STL.News reported on June 29, 2026.
Professional investors are more concerned about the leverage structure than stock prices themselves. The current loop runs: AI expectations, semiconductor prices rise, leverage increases, ETF buying, index gains, additional leverage. On the upside, gains accelerate rapidly. On the downside, a chain reaction can develop: margin calls, forced selling, further price declines, additional margin calls.
The second half of 2026 will test whether investors can manage the leverage they have built. Corporate earnings, credit investment, and short selling flows must be held in view simultaneously. A single-variable lens will not be sufficient.
Prepared with AlphaScala editorial tooling from the source reporting linked above. Indexable analysis may include a cited Alpha Score value. Publishing checks screen each story before release. Educational coverage, not personalized advice.