#AIStockGuruReportedlyBullishOnAI

The most interesting part of the “AI Stock Guru” story is not a single stock pick. It is where the portfolio is positioned across the AI supply chain. The strategy suggests a long-term belief that AI demand will keep expanding, while also showing a very important warning: the future of AI may be bullish even when some of the most popular AI semiconductor stocks become expensive in the short term.

The portfolio of Leopold Aschenbrenner’s Situational Awareness Fund showed significant exposure to the physical infrastructure required to make AI work. SanDisk (SNDK) was increased by 85,000 shares to approximately 1.14 million shares, worth around $724 million. CoreWeave (CRWV) was increased by more than 1.07 million shares, taking the position to roughly $556 million. These are not simply bets on another AI software application—they represent exposure to storage and GPU cloud infrastructure.

That distinction matters. AI models require enormous amounts of computing power, but computing power itself depends on a much larger ecosystem: GPUs, high-speed memory, cloud capacity, data centers, electricity, networking and storage. The portfolio's exposure to Bloom Energy, CoreWeave, SanDisk, Nebius, Applied Digital and other infrastructure-related companies shows how the thesis extends beyond the headline semiconductor names.

And then comes the twist that makes this strategy much more interesting.

While the fund was building positions in AI infrastructure, it also held substantial downside protection against major semiconductor names. In its Q1 filing, put options on SMH, NVDA, ORCL, AVGO and AMD alone represented a very large portion of the portfolio's nominal exposure. Including additional semiconductor hedges, more than 60% of the reported nominal position was tied to downside protection or hedging against major AI hardware stocks.

That means “bullish on AI” should not be interpreted as “bullish on every AI stock at any price.” The portfolio was effectively separating the AI industry's long-term growth story from the short-term valuation risk of its most crowded trades. The message is subtle but important: AI demand can continue rising while semiconductor stocks still experience sharp corrections.

CoreWeave is a perfect example of the infrastructure angle. Adding more than 1.07 million shares pushed the holding to approximately $556 million, making GPU cloud infrastructure a meaningful part of the portfolio. If AI workloads continue expanding, demand for specialized cloud computing and data-center capacity could remain structurally strong. But that also creates another question for investors: how quickly can infrastructure companies turn enormous AI demand into sustainable cash flow?

SanDisk provides another piece of the puzzle. Increasing the position to about 1.14 million shares worth $724 million points toward the importance of memory and storage in the AI infrastructure cycle. AI is not only about training larger models; inference, data movement and increasingly complex workloads all require massive amounts of storage and memory capacity.

The power theme may be even more important over the next stage of the cycle. Data centers cannot scale simply because investors want more GPUs. They need electricity, cooling systems, grid connections and physical capacity. This is why exposure to companies involved in power generation and data-center infrastructure can be viewed as a second-order AI bet: if AI demand grows faster than infrastructure capacity, electricity and computing availability can become bottlenecks.

But there is a major lesson hidden inside the hedge positions. The AI boom may be structurally bullish, while the market prices of AI leaders can still move in the opposite direction. A company can benefit from rising AI demand and still decline 20% if expectations were already too high. That is why the portfolio's combination of infrastructure longs and semiconductor puts is more informative than simply saying “the AI Stock Guru is bullish.”

There is also a broader market question here. Investors have spent years focusing on who makes the chips. NVIDIA, AMD and Broadcom became the obvious beneficiaries of AI spending. The next phase could broaden the trade toward who supplies the computing capacity, storage, electricity and data-center infrastructure needed to deploy those chips at scale.

For me, the most important signal is therefore not one ticker. It is the capital flow across the entire AI infrastructure chain. If cloud demand keeps accelerating, data-center construction continues, power capacity expands and memory/storage demand remains strong, then the AI investment cycle could have much more room to run. But if infrastructure spending begins outrunning actual AI monetization, valuation pressure could return quickly.

The key distinction is simple: long-term AI demand ≠ guaranteed short-term AI stock returns. The portfolio's structure shows why professional positioning can be much more nuanced than a simple bullish or bearish label.

My takeaway is that the next major AI opportunity may increasingly move downstream from the chip itself. GPUs are the engine, but cloud capacity, electricity, storage and data centers are the roads that allow the engine to operate. If those bottlenecks become the next limiting factor, infrastructure could capture an increasingly important share of the AI capital cycle. @Gate_Square