The direction of securities markets in the second half of 2026 will be shaped less by corporate fundamentals and sustained earnings structures than by two forces that have until now been treated as secondary variables: debt-financed investment — credit leverage — and the direction of short selling. These two sectors, which once functioned as peripheral investment variables, are set to amplify destructive force as primary market drivers. Derivatives, which ought to function as risk management instruments, are being repositioned as aggressive platforms for profit extraction.
The earnings data that every investor watches must not lose its meaning. But concentrating all focus on earnings alone and driving investment in a single direction is an act of recklessness. Ordinary investors have limited access to specialized information, and so they cling to earnings data as their only anchor. Some segments of the mass media reinforce this by illuminating investment analysis from a single direction only. The moment investment becomes buried in one direction, difficulty takes root.
The time has come to examine closely the field indicators that have been avoided and left unexamined. Keeping this in mind and cultivating the ability to read markets simultaneously through the twin lenses of earnings and leverage is the only way to survive in equity markets. Only the investors who conduct proper analysis of this dimension in the second half will be equipped to navigate a roller coaster market.
――――――――――――――――――――
Credit Leverage at Record Levels
The U.S. securities market in the first half of 2026 recorded one of the highest leverage levels in history. The defining feature was a sharp increase in credit investment — margin — and derivatives leverage, directly coinciding with the AI and semiconductor rally.
According to FINRA’s monthly margin statistics, customer debit balances in margin accounts reached a record $1.416 trillion in May 2026 — an increase of 8.5% from April and a 53.7% rise year-over-year from May 2025’s $921 billion (FINRA Margin Statistics, May 2026; Advisor Perspectives, June 24, 2026). This 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% — also a record high, and nearly double the long-term median of approximately 2.37% (GuruFocus, FINRA Investor Margin Debt Relative to GDP, May 2026). Net credit balance — free cash accounts minus margin debt — reached a record low of negative $991.7 billion in May 2026, the thinnest cushion against forced selling on record (Advisor Perspectives, 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 — but all were late-cycle or post-surge environments in which equity risk proved more unforgiving than investors assumed (Bob Brinker, Substack, June 2026).
Market capital concentrated into AI-related names. NVIDIA, 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, illustrating the structural violence embedded in these instruments (CryptoBriefing, July 2026).
――――――――――――――――――――
Short Selling: Structure and Reality
An important structural distinction must be made here. Many commentators use the phrase “short selling volume,” but the United States has no system analogous to Korea’s — there is no official daily announcement of aggregate market-wide short selling amounts. Instead, FINRA publishes short interest data for all securities twice monthly. In the first half of 2026, short interest in leading AI names actually declined or triggered short covering as prices rose. Short positions were maintained in traditional industries. The more accurate characterization is not that short selling surged across the whole market, but that short capital was pushed out of AI names.
Short covering — in which short sellers close positions by buying, to limit losses — occurred repeatedly throughout the first half. The defining feature of the period was not an increase in short selling but an accelerating liquidation of it. The structural loop repeated: AI investment expansion → semiconductor earnings surge → results exceeding expectations → short seller losses mount → forced buying and short covering → further price appreciation.
The most illustrative case is Micron Technology. In the first half of 2026, an explosion in AI memory (HBM) demand, supply shortages, and earnings substantially above estimates drove a sharp price surge. The earnings release was recognized as a critical indicator of AI investment continuity, inducing a reduction in short positions. The result: a single-day gain of approximately 17%, a market capitalization increase of approximately $189 billion, and a new record high. As earnings came in far stronger than anticipated, short sellers liquidated positions and amplified the advance. Reuters characterized this as the trigger for a re-ignition of the AI rally. NVIDIA, Broadcom, SanDisk, Western Digital, and Qualcomm moved in parallel — not an isolated single-stock event but a simultaneous short covering and momentum chase across the entire AI semiconductor sector.
――――――――――――――――――――
Hedge Fund Leverage: The Real Picture
This is the most critical area to watch. In 2026, hedge funds used equity repo more heavily than margin loans. Primary dealers are carrying record equity repo exposure that now exceeds $220 billion (STL.News, June 29, 2026). U.S. hedge fund leverage in 2026 was evaluated as the highest since the 2008 financial crisis. The structure, however, differed from the Archegos Capital Management collapse of 2021, which involved catastrophic over-concentration in a single name. In 2026, the distinctive feature was leverage deployed across multi-strategy, relative value, Treasury basis trades, and AI and technology long-short positions (Goldman Sachs Marquee, Prime Services, January 2026). While diversified in form, the aggregate exposure is historic in scale.
――――――――――――――――――――
Why the Current Market Is Dangerous
What professionals are most concerned about is not stock prices themselves but the leverage structure underlying them. The current loop runs: AI expectations → semiconductor prices rise → leverage increases → ETF buying → index gains → additional leverage. In this structure, gains accelerate rapidly on the upside. On the downside, a chain reaction can develop: margin call → forced selling → further price decline → additional margin calls. The June 5th episode — a 3x leveraged semiconductor ETF losing 31% in a single session — is not an outlier. It is a preview.
The first half of 2026 in U.S. securities markets did not rise on AI optimism alone. Record credit investment of approximately $1.4 trillion and expanded leverage trading amplified the advance. AI lit the fire. Leverage fed the flames.
The internal substance also warrants careful review. That credit investment has reached all-time highs is confirmed by objective data. There is currently no official statistic that permits a definitive claim that overall market short selling also reached all-time highs. Short selling varied sharply by name, and in leading AI names, short covering rather than short accumulation was the dominant pattern. These are materially different dynamics, and conflating them produces a distorted picture.
A colder and more precise analysis of the current situation is the prerequisite for setting direction. At present, ordinary investors, professional daily traders, and institutional investors alike are riding a roller coaster. The April through June period saw indices and individual names — including AI and semiconductor stocks — move with volatility levels that defied prediction. This cannot be characterized as normal equity market behavior. The presence of overheating driven by credit investment and leverage strategies must be acknowledged.
In the second half, this dynamic is expected to intensify further. Corporate earnings must be held in view simultaneously with credit investment and short selling flows. Approaching the market through a biased, single-variable lens will not be sufficient. A breathless environment in which multiple variables are reflected in real time is the expected condition.
The strategic choice divides into two orientations: setting clear personal target levels and actively responding with buy and sell decisions, or trusting the market and pressing forward with a profit-maximization strategy grounded in one’s own rigorously analyzed data. Whichever direction is chosen, the most important factor is the investor’s own decision. Stock investment is by its nature a compendium of decision-making undertaken while bearing risk. In the second half, tighter and more rigorous risk management — not broader exposure — is the strategy that is absolutely required.
