원문 보기
But the reality is that everyone in this ecosystem, including me, is trying to figure out how to price it all so that the end buyer makes more money and grows faster, and that's still up for debate. Well, I think you're getting to the point that Alex Karp, who was here a week and a half ago, was talking about, about a kind of token maximization , a concept that is that some of these companies —and this is Anthropic and OpenAI—where there are CEOs that I've also talked to now, are saying that the math doesn't add up.
In my opinion, CEOs and CFOs probably have no idea how much token maximization is happening inside their organizations. I suspect that one day you'll miss your target, your EPS will be a few cents lower, and the CEO will ask the CFO, "What happened? Where did all these extra operating expenses come from?" And they'll trace it back to a $50 barrel of intelligence versus a $1.50 barrel of intelligence.
This has not happened yet. This is interesting. These are software stocks . What else is interesting? So, this is a component of Dow, IBM. So, the Dow fell 500 points. Just do the math. Multiply 50 points by whatever divisor it is. The NASDAQ rose by more than 100. And it is still rising. So if there's a doubt about technology in general, you're not seeing it in the NASDAQ yet, and the big 500-point drop in the Dow is because, well, I think it's fair to say there's a lot more to the NASDAQ that's not software.
All the hardware, all the memory, all the chips. This whole complex will still make a ton of money, right? The question is again in the future, right? If you buy oil , you need to pour it into a new engine. This engine should make you go faster, right? It seems like you're assuming that the high-end LLM complex, which is essentially represented by Anthropic and OpenAI in terms of pricing all of this, and which is still in the private markets, is going to collapse.
Is that what you're trying to imply? I think that's what you're trying to push me towards. Let me be very clear. The hardware and memory complex will work for several more years, as there are huge limitations in this ecosystem. So, there is a huge deficit. These things will continue to be successful.
The NASDAQ will likely outperform the S&P. Good. If you look at the only other niche where there is real growth, it is the LLM sector. As you said, I don't know what the long-term valuation is going to be , but if you see a lot of public companies underperforming in the next few quarters because of operating expense mistakes , right?
And they will trace it to uncontrolled costs that were not even suspected within the organization, you will be a little mistaken. Number spreads for growth. So, for Exxon Mobil, I'm looking at the monthly expiration date of September 18th. I plan to buy an in-the-money call option at the 155th call, and then sell the 170th call, just below that historical high.
Earlier today, you could pay about six dollars for this debit spread, and that would give you a risk-to-reward ratio of one and a half to one. And I specifically use such an "in the money" debit spread because the breakeven price is only a couple of dollars higher than the current one. So I'm not overpaying for the time decay because right now, you know, oil volatility remains relatively high and I want to be able to access that.
Okay, I know you like Oracle too. So tell us about your thesis on Oracle, which is still in the red since the beginning of the year , although it is trading higher today. Yes, Oracle has lost over 50% of its value in the last few months. But as CoreWeave's financial results earlier this week proved, the development of AI infrastructure is not actually slowing down.
On the contrary, it is gaining momentum, and Oracle is gaining support along with the results of CoreWeave. It managed to break through the 150 level. This was a key level that we were watching as a potential reason or level to increase the position. So, now that we've broken the 150 mark, especially on high volume, ahead of the S&P 500, that suggests a target of $185 for Oracle, especially when looking at the fundamentals after the company lost 50% over the past few months.
Oracle looks much more attractive from a valuation perspective, especially when you consider the growth profile and how profitable Oracle remains. This is where I think Oracle among the so-called " hyperscalers" looks most attractive in terms of risk-reward ratio towards growth. OK. So, you like Oracle, you like Exxon.
What else are you looking at? I know you mentioned the commodities market . Are there any specific strategies there: is it an ETF or do you work with physical assets? Yes. So, we definitely use ETFs like GLD and SLV to gain access and exposure to the potential growth of the commodities themselves. When it comes to mining companies, as options traders we usually see that most of them are not very liquid.
So, other than Barrick Gold, we don't take many individual positions in the miners themselves, except potentially through GDX, the gold mining index. But in the uranium sector, we are looking at CCJs with an eye on growth. It's obviously the largest uranium miner in Canada, and again , from an options liquidity perspective , unless you're using the uranium ETF URA, that's how we play the commodities sector.
You remain heavily invested in technology, even when looking at the market more broadly. Where do you see the greatest strength right now, Tony, where are you looking for the best opportunities? Yeah, I mean, the big surprise for everyone was how strong the software turned out to be , right? Because software has been the sector that has been hit hard over the past few years by the AI boom.
But we see that many of them demonstrate considerable growth potential. These are Adobe, Zoom, Salesforce. These are companies that could have potentially been wiped out of the market because AI completely destroyed their business, and now they are coming back to life. You know, we're actually really interested in Nvidia just changing the rules of the game in AI infrastructure.
Nvidia's latest moves raise a much broader question than whether the company will be able to sell more GPUs . What happens when the company that sells the technology also helps fund the infrastructure needed to deploy it. This is exactly what we are beginning to observe. Nvidia is reportedly backing a large-scale AI data center project in Ohio announced by OpenAI, with the potential capacity of the project reaching several gigawatts.
And Nvidia is no longer just a chip supplier; it becomes part of the financial structure, as the main constraint in AI may no longer be demand, but capital. Building data centers for AI requires incredible amounts of money. You need energy, land, power lines, cooling, networks, the physical data center itself, and then billions of dollars in accelerators.
So Nvidia has an interesting initiative. If customers have difficulty financing infrastructure, they won't be able to buy Nvidia chips. So, helping to resolve financial problems could potentially create more demand for Nvidia's core business. This is the flywheel of financing that creates infrastructure.
Infrastructure creates computing power, and computing power creates AI services. AI services generate revenue, revenue justifies . But this is where the story gets dangerous, because financing can accelerate both investment and overinvestment. If demand for AI continues to grow rapidly, this could be one of Nvidia's smartest strategic moves.
But if companies create too much capacity , data centers will eventually start competing for customers, GPU rental prices will fall, and infrastructure profitability will decrease. This is why the idea of "dark" GPUs becomes important. Imagine spending billions on computing power that is not actually being used enough to generate attractive profits.
The equipment doesn't disappear, the data center doesn't disappear, but the economy gets much worse, which can create a problem not only for the operating company, but also for lenders and investors. So, there are two possible futures: an optimistic scenario where AI adoption accelerates, token prices fall, making AI cheaper, and more businesses start using it.
The demand for computing is growing, and GPU usage remains high. Rental prices remain stable, and the huge capital invested continues to generate attractive returns. In such a scenario, Nvidia's help in funding infrastructure could actually prolong the " golden era" of AI. Pessimistic scenario: capital becomes too accessible.
Everyone starts building. Data center capacity is growing faster than the demand for AI. GPU rental prices are falling, profits are falling, and eventually the market realizes that some of the infrastructure was built years ahead of real demand. That's when the financial structure becomes a potential amplifier of recession.
And there is another aspect that investors should not ignore—it is electricity. The OIO project is not just about GPUs and large-scale AI.
Comments 0