Why Is a $13.7B AI Fund Loading Up on Crypto Miners While Shorting $8.46B of Chipmakers?
Leopold Aschenbrenner is a 24-year-old former OpenAI researcher who left in 2024 to found Situational Awareness LP — a hedge fund focused on AI infrastructure that has raised $13.7 billion in under two years. Last week the fund filed its first quarterly 13F disclosure with the SEC.
What the 13F reveals is a contrarian position at scale: $8.46 billion in put options against Nvidia and the broader semiconductor sector, alongside increased long positions in four crypto mining companies — CleanSpark, Riot Platforms, Applied Digital, and IREN.
The thesis is infrastructure arbitrage: short the companies selling AI hardware, long the companies that already own the physical assets required to run it. We have looked at the portfolio construction and the key risk levels to watch. Here is what traders need to know.
Who Is Leopold Aschenbrenner?
Leopold Aschenbrenner was born in Germany in 2001 or 2002. He enrolled at Columbia University to study economics and graduated at 19, completing the degree in roughly two and a half years through an accelerated course load.
He joined OpenAI in 2023 as part of the Superalignment team, a group focused on long-term AI safety research. In April 2024, OpenAI fired him over an alleged information security leak related to internal concerns he raised about the company's safety practices. Aschenbrenner disputes the characterisation and says he was dismissed for raising legitimate concerns. The firing drew significant attention across the AI and tech industries.
Weeks after leaving OpenAI he published "Situational Awareness: The Decade Ahead" — a 165-page essay arguing that AGI would arrive by 2027 to 2030 and that the physical infrastructure buildout was the defining investment opportunity of the era. He founded Situational Awareness LP in late 2024.
His fiancée is Avital Balwit, Chief of Staff to the CEO at Anthropic — OpenAI's primary competitor. He lives in San Francisco.
His personal net worth is not publicly disclosed. Reports indicate he has invested nearly all of his personal capital into the fund. As general partner of a $13.7 billion portfolio, his stake in the economics of the partnership is substantial — but no specific figure has been confirmed by Aschenbrenner or verified by public filings.
What Does the 13F Filing Actually Reveal?
Situational Awareness LP held the following positions at the end of Q1 2026:
Chipmaker Puts (Short Bets):
VanEck Semiconductor ETF (SMH): $2.0 billion notional put exposure
Nvidia (NVDA): $1.6 billion notional put exposure
Oracle (ORCL): $1.1 billion notional put exposure
Broadcom (AVGO): $1.0 billion notional put exposure
Advanced Micro Devices (AMD): $969 million notional put exposure
Micron (MU): $584 million notional put exposure
Taiwan Semiconductor (TSM): $535 million notional put exposure
ASML (ASML): $494 million notional put exposure
Intel (INTC): $159 million notional put exposure
Corning (GLW): $21 million notional put exposure
Total put exposure: $8.47 billion
Bullish Bets:
Bloom Energy (BE): 6.5 million shares ($879 million) plus call options on 409,000 shares ($55 million)
Sandisk (SNDK): 1,140,119 shares ($724 million) plus call options on 611,900 shares ($389 million) — $1.1 billion combined
CoreWeave (CRWV): 7,177,919 shares ($556 million) plus call options on 1,814,500 shares ($141 million) — $697 million combined
Micron (MU) and Taiwan Semiconductor (TSM): Selective call positions ($422 million and $355 million respectively)
Crypto Miners (Increased Positions):
IREN Limited (IREN): $401 million
Core Scientific (CORZ): $389 million
Applied Digital (APLD): $320 million
Riot Platforms (RIOT): $142 million
CleanSpark (CLSK): $104 million
Bitfarms (BITF): $39 million
BitDeer Technologies (BTDR): $30 million
Why Is He Shorting Chipmakers but Long Crypto Miners?
The apparent contradiction resolves when you understand the infrastructure thesis.
Aschenbrenner is not bearish on AI compute. He is bearish on chipmaker valuations at current prices. The $8.46 billion in puts is a bet that semiconductor stocks are overvalued relative to the actual build-out of AI infrastructure.
But he is bullish on the physical infrastructure itself. Data centres. Power. Land. And critically, the companies that already own these assets.
Crypto miners are not bitcoin plays in this framework. They are real estate and power plays with existing data centre footprints. CleanSpark, Riot, Applied Digital, and IREN own facilities, power contracts, and land that can be repurposed for AI compute.
The trade is: short the chipmakers who sell shovels, long the miners who already own the goldfields.
How Does This Affect Crypto Markets?
The direct link to crypto perps is through miner equity sentiment. When institutional capital flows into miner stocks, it creates a feedback loop:
1. Miner equity rises → Miner balance sheets strengthen → Less forced BTC selling
When miners are well-capitalised, they do not need to sell bitcoin to fund operations. This reduces spot selling pressure. Spot tightens. Perp funding turns positive.
2. Miner equity rises → Hash rate expansion → More network security → BTC narrative strengthens
Rising hash rate is a bullish narrative for BTC. Narrative drives retail positioning. Retail goes long perps. Funding rises.
3. The contrarian risk: If Aschenbrenner is wrong about chipmakers, he is probably wrong about miners
The $8.46 billion in chipmaker puts is a leveraged bet. If NVDA reports strong earnings and the puts expire worthless, the fund faces massive losses. Forced redemption could trigger selling across the portfolio, including the miner positions. This would reverse the feedback loop.
Why Are Crypto Miners the Hidden AI Trade?
The market still treats crypto miners as bitcoin proxies. Aschenbrenner's filing reveals they are something else: distressed infrastructure assets with power contracts in a world where AI data centre demand is exploding.
The numbers:
A new AI data centre costs $500 million to $2 billion to build
A crypto miner already owns the land, power substations, and cooling infrastructure
Repurposing a mining facility for AI compute costs 60-70% less than greenfield construction
The transition takes 12-18 months, not 3-5 years
This is why Aschenbrenner increased positions in CLSK, RIOT, APLD, and IREN. He is betting that the market will revalue these companies from "miners" to "data centre operators".
What Are the Risks to This Thesis?
Three conditions break the miner-as-infrastructure narrative:
1. Bitcoin price collapse: If BTC continues to fall below the cost of mining, miner margins compress regardless of AI pivot potential. The equity prices would fall, forcing Aschenbrenner to mark down positions or sell.
2. Power contract renegotiation: Many miner power contracts are at industrial rates that assume constant baseload consumption. AI data centres have variable load profiles. Utilities may refuse to renew contracts at favourable terms.
3. Regulatory blocking: Local governments that welcomed bitcoin mining may resist AI data centres due to noise, water usage, or land-use concerns. Zoning changes could strand assets.
How Should Crypto Traders Position Around This?
For directional traders:
Monitor miner equity flows as a leading indicator for BTC sentiment. When CLSK and RIOT outperform the S&P 500, BTC funding typically turns positive. When they underperform, funding turns flat or negative.
For pairs traders:
The chipmaker-miner divergence is tradable. Long miner stocks or crypto versus short NVDA Equity Perps on BitMEX.
On BitMEX, you can trade Nvidia Equity Perps ( NVDAUSDT) with up to 20x leverage. With Equity Perps, traders can use crypto as collateral and trade even when traditional markets are closed. Nvidia reports earnings after the market closes, locking out traditional traders.
Learn more about Equity Perps here.
For risk management:
Aschenbrenner's $8.46 billion in chipmaker puts is a crowded short. If chipmakers rally on earnings, the forced covering could cascade into selling across his portfolio. This would hit miner stocks and, by extension, BTC sentiment.
How to Track the Situational Awareness LP Portfolio
As an institutional manager with more than $100 million in US-listed assets, Situational Awareness LP must file a Form 13F with the SEC within 45 days of each calendar quarter end. The filing lists all long equity positions and options exposure as of the final day of the quarter.
Direct SEC filing: Search CIK 0002045724 on SEC.gov EDGAR, or access the Q1 2026 13F directly via the index URL in the Sources section below.
Third-party trackers:
hedgefollow.com — tracks Situational Awareness LP holdings with quarter-over-quarter changes and new position flags
whalewisdom.com — historical 13F data with performance attribution
Yahoo Finance — search any stock the fund holds (IREN, Applied Digital, CoreWeave) and Situational Awareness LP appears under Institutional Holders
Filing calendar for 2026:
Q1 (March 31) → filed May 18, 2026 ✓
Q2 (June 30) → due by August 14, 2026
Q3 (September 30) → due by November 14, 2026
Q4 (December 31) → due by February 14, 2027
Important limitation: 13Fs only show long equity positions and listed options. They do not show short positions, bonds, cash, or private holdings. The $8.47 billion in put options appears because puts are listed equity derivatives — but the fund's full short book is not visible from a 13F alone.
What Is the Situational Awareness Essay?
"Situational Awareness: The Decade Ahead" is a 165-page essay series Aschenbrenner published in June 2024, weeks after being fired from OpenAI. It is the intellectual foundation for everything in the 13F — understanding it makes the portfolio positions considerably less surprising.
The argument runs in three parts. First, AI capability is scaling predictably: each order-of-magnitude increase in compute has produced reliably smarter models, the scaling laws have held, and there is no obvious wall in sight. Second, the physical infrastructure required to run these systems — power, data centres, semiconductors — is the genuine bottleneck, not the algorithms themselves. This is not a software race; it is an energy and real estate race. Third, the geopolitical stakes are existential: the US and China are competing to reach AGI first, and whoever gets there first shapes the 21st century.
The investment implications are direct. If the bottleneck is infrastructure, then the companies that own or build that infrastructure are the trade, not necessarily the chip designers or the chatbot companies.
The full essay is freely available as a PDF at situational-awareness.ai. It became one of the most widely read documents in the AI investment community in 2024 and 2025. The fund Aschenbrenner subsequently launched takes its name directly from it.
Frequently Asked Questions
Who is Leopold Aschenbrenner?
Leopold Aschenbrenner is a former OpenAI researcher who left in 2024 to found Situational Awareness LP, a hedge fund focused on AI infrastructure. He gained prominence through his "Situational Awareness" essay series, which argued that artificial general intelligence would arrive within a decade and that the AI infrastructure buildout would be the defining investment opportunity of the era. Born in Germany in 2001 or 2002, he graduated from Columbia University at 19 and is one of the youngest managers to oversee a fund at this scale — $13.7 billion in reported portfolio exposure. His background is academic and technical rather than financial, which may explain the unconventional portfolio in his first 13F: $8.47 billion in semiconductor puts alongside long positions in crypto miners treated as data centre infrastructure proxies, and a $697 million direct bet on CoreWeave.
What is a 13F filing?
A 13F is a quarterly disclosure that institutional investment managers with more than $100 million in assets must file with the SEC. It lists all equity holdings, including long positions and options exposure, as of the final day of each quarter. Managers have 45 days from quarter end to submit. 13Fs are closely tracked by investors and journalists because they reveal how major funds are positioned — which sectors they favour, what options leverage they are applying, and where they are concentrating risk. Situational Awareness LP's first 13F was accepted by EDGAR on May 15, 2026 — the 45-day deadline — but carries an official filing date of May 18, three days later. Late filings can trigger civil penalties up to $750,000. The filing disclosed $8.47 billion in chipmaker put options alongside long crypto miner and AI infrastructure positions.
Why are crypto miners considered AI infrastructure plays?
Crypto miners own three assets that AI data centres urgently need: land in power-rich locations, high-voltage electrical substations already connected to the grid, and cooling infrastructure designed for sustained high-density compute loads. AI GPU clusters require an almost identical physical envelope. A greenfield AI data centre costs $500 million to $2 billion and takes 3–5 years to permit, construct, and connect. Repurposing a miner facility costs 60–70% less and takes 12–18 months. Core Scientific's 200MW deal with CoreWeave and Applied Digital's pivot from mining hosting to AI hosting are the clearest proof of concept already in the market. Aschenbrenner's 13F suggests he views IREN, Core Scientific, Applied Digital, RIOT, and CLSK as distressed infrastructure assets mispriced as bitcoin speculation.
What is Situational Awareness LP?
Situational Awareness LP is a hedge fund founded by Leopold Aschenbrenner in 2024 after he departed OpenAI. Named after his influential essay series on AI safety and development timelines, the fund focuses on AI infrastructure investments — semiconductors, power, data centres, and the companies building the physical layer of the AI economy. It raised $13.7 billion, an extraordinary sum for a debut fund. Its first SEC 13F filing, submitted for Q1 2026, revealed a portfolio that defies the obvious AI trade: rather than buying semiconductor companies supplying AI hardware, the fund holds $8.47 billion in put exposure against them, while going long crypto miners as infrastructure proxies, Bloom Energy as a power play, and CoreWeave as the direct AI compute beneficiary. The approach is infrastructure arbitrage.
How do miner equity prices affect BTC perp funding?
When miner equity prices rise — as happens when institutional capital like Aschenbrenner's flows into IREN, Core Scientific, RIOT, APLD, or CLSK — miner balance sheets strengthen. Well-capitalised miners sell less bitcoin to fund operations, reducing spot selling pressure. Spot price firms, the perp market premiums above spot, and funding turns positive as longs pay shorts to hold positions. The reverse applies when miner equity falls: compressed balance sheets force BTC sales, spot supply rises, and the perp flips to discount, driving funding negative. The lag between miner equity moves and funding rate shifts is typically 24–48 hours — giving traders who monitor CLSK and RIOT as lead indicators a structural edge before the signal fully prices into crypto perp funding.
What happens to BTC if Aschenbrenner's chipmaker shorts go wrong?
The $8.47 billion in chipmaker puts represents 62% of the fund's assets — far beyond a typical hedge. If semiconductor stocks rally on strong earnings, the puts move deeply out of the money and losses accumulate. Forced redemptions could require selling across the entire portfolio, including the long positions in IREN, Core Scientific, RIOT, APLD, and CLSK. Miner stock declines would compress miner balance sheets, increasing BTC selling pressure and pushing perp funding negative. The timing risk is acute: if NVDA's near-term earnings beat consensus significantly, the cascading effect through the fund's portfolio could hit crypto markets within days. For perp traders, NVDA earnings are the primary risk trigger for this thesis — monitor the print before sizing up.
What is the historical precedent for miner-to-data-centre conversions?
The pivot from bitcoin mining to AI compute is already proven at scale. Core Scientific signed a 200MW deal with CoreWeave in 2024 to repurpose its mining facilities; the stock rose 300% on the announcement. Hut 8 rebranded as a digital infrastructure company in 2023 and now runs both mining and cloud compute from the same physical sites — the narrative shift added approximately $500 million to its market cap. Applied Digital never mined bitcoin directly; it hosted miners, then pivoted entirely to AI hosting and is up 400% since the announcement. These precedents validate the thesis that the market undervalues miner infrastructure. Aschenbrenner is betting the revaluation repeats across IREN, Core Scientific, RIOT, and CLSK, not just the early movers.
How do I trade the miner-AI thesis on BitMEX?
The most direct expression of this thesis in crypto markets is through the BTC perpetual swap on BitMEX (XBTUSD). When CLSK and RIOT outperform the S&P 500 — signalling institutional inflows into miner equity — BTC perp funding typically turns positive within 24–48 hours. Going long XBTUSD in that window lets you collect positive funding while holding directional exposure. Step 1: deposit crypto to your BitMEX wallet. Step 2: open the XBTUSD perpetual from the contract selector under Crypto. Step 3: go long when miner equities trend up; reduce exposure when chipmaker puts face forced covering. Step 4: monitor funding rates in the position overview — positive funding above 0.02% per 8-hour period signals strong institutional demand. Size for 0.5–1% moves, not 5%, given the 24–48 hour lag between equity signals and perp funding shifts.
What is Leopold Aschenbrenner's net worth?
Leopold Aschenbrenner's exact personal net worth is not publicly disclosed. He keeps his private finances confidential, and as a private fund manager he has no obligation to report personal wealth. Reports from Fortune and Phemex indicate he has invested nearly all of his personal capital into Situational Awareness LP, tying his wealth directly to the fund's performance. As general partner, he earns standard management fees — typically around 2% of assets under management — and a performance allocation of approximately 20% of profits. With the fund reporting $13.7 billion in portfolio exposure as of Q1 2026, the economics of the general partnership represent substantial paper wealth, though the actual AUM and fund performance figures are not publicly available. He is widely regarded as one of the wealthiest people under 25 in the AI investment space.
How do I track Leopold Aschenbrenner's portfolio?
The most direct method is via SEC EDGAR. Situational Awareness LP (CIK: 0002045724) files a quarterly 13F disclosure within 45 days of each quarter end. For a more readable format, hedgefollow.com and whalewisdom.com both track the fund and display quarter-over-quarter position changes with new entries flagged. Yahoo Finance surfaces Situational Awareness LP under the Institutional Holders tab for any stock it holds — search IREN, CoreWeave, or Applied Digital and the fund's position size will appear. Note that 13Fs only disclose long equity and listed options; short positions and private holdings are not included. The next 13F covering Q2 2026 positions (as of June 30, 2026) is due by August 14, 2026.
How did Leopold Aschenbrenner graduate from Columbia at 19?
Aschenbrenner enrolled at Columbia University to study economics and completed the degree in approximately two and a half years — graduating at 19, roughly two years ahead of the standard schedule. Wikipedia notes this but does not specify the exact mechanism. Early graduation at US universities typically involves a combination of AP credit, dual enrollment, and compressed course loads — all of which allow a student to arrive with significant credits and finish in fewer semesters. He was at or joining OpenAI around 2022 or 2023, consistent with a graduation in that window. The Columbia story is frequently cited alongside his OpenAI firing and the Situational Awareness essay as evidence of an unusually accelerated trajectory — from academic prodigy to managing one of the largest debut hedge funds in recent memory.
Disclaimer: This article is for informational purposes only and does not constitute financial advice. Trading forex with leverage involves substantial risk of loss. Past performance is not indicative of future results. Always conduct your own research and consider your risk tolerance before trading.
