Oracle's Data Center Build-Out Faces Investor Anxiety Over Financing

Key Quotes

"The big context on oracle is that they are embarking on a historic build out of data centers where it's really a logistical feat it's quite a um tight schedule they need to be on and if all goes right it's transformational for oracle and the entire industry but all that needs to happen for that to get off is a couple of things falling out right and that's the issue with what we're seeing today is that the ft is reporting that there's been some issues in getting the financing for one of their big centers and that's giving investors a lot of anxiety that hey are are these lenders seeing something that we are not that's what's driving the fear this morning."

Brody Ford explains that Oracle's ambitious data center expansion faces significant logistical challenges and tight timelines. Ford highlights that issues in securing financing for a major center are causing investor anxiety, suggesting that lenders might perceive risks not immediately apparent to the market.

"What's interesting is you've been writing about the sheer scale of leases and i think leases is something different because it's off balance sheet and always think we're hearing it in a 10q that they could have up to 300 billion worth thereabouts in terms of their overall leases that eventually they're on the hook for paying but how is this different from capital expenditure and the like right so we're used to hearing about capital expenditures which is i'm going out and buying computers i'm buying wires i'm buying the data centers themselves but what's becoming more popular is renting the data centers and that allows you to not pay everything upfront and not have that be on your books directly and so what we've seen is a big spike in these kind of future dated commitments that oracle is the example that they have about 250 billion that they're on the hook for over the next 20 years or so and most likely this michigan data center would be one of them it's just kind of you know one more pool of spending that we realize quite how large this build out is going to be."

Brody Ford clarifies the distinction between capital expenditures and leases in the context of data center financing. Ford notes that leases, being off-balance sheet, allow companies like Oracle to commit to significant future spending without immediate on-book liabilities, as demonstrated by Oracle's substantial future lease commitments.

"Yeah it's right amazon also a close partner of anthropic they've invested 8 billion in that company as part of a deal that got anthropic to use amazon's own homegrown ai chip for their models right so this would definitely be a coup for amazon's in house chip making effort um which analysts are still trying to uh to struggle a little bit to um understand the effectiveness of to date in part because there's just few enormous customers for this stuff there's not a whole lot of big model trainers out there so it goes without saying that if openai can make you know the gpt suite work on top of amazon's silicon that's that's a big deal for for aws and for amazon because it was such a big deal for alphabet and we think about the endorsement that its vertical model seems to be getting from the market."

Matt Day explains that Amazon's potential investment in OpenAI, similar to its investment in Anthropic, could serve as a significant endorsement for Amazon's in-house AI chips. Day points out that the effectiveness of these chips is still being assessed by analysts due to the limited number of large-scale customers, making OpenAI's adoption a crucial validation for Amazon's silicon efforts.

"Yeah well i think the first is the market is growing and so um you know this is an exponential growth curve so i don't think that it's a winner take all and so you know multiple companies can can do well and in some ways i think when you step back and think about what's going on um you know there are there is kind of a space race for ai but they're going to multiple areas multiple moons um and so they're all doing their own kind of tech domain expertise i think like with the tpu it's very optimized um you know for a specific stack and for the google ecosystem and then i think nvidia is like kind of more the merchant like broader platform uh can run everything and i think both can have a place um and you can see that you know google's been doing really well over the last five years uh so has nvidia uh so has amd so has broadcom and so to me i think that um you know there there's not a winner take all here."

Tony Wang argues that the rapidly growing AI market is not a zero-sum game, allowing multiple companies to succeed. Wang likens the AI landscape to a "space race" with "multiple moons," where different companies, like Google with its optimized TPUs and Nvidia with its broader platform, can carve out distinct niches. Wang supports this by noting the success of various chipmakers over the past five years.

"Yeah we're definitely on an upward trajectory and memory prices so i think what's going on there is that hbm the stuff that goes into these gpus and ai systems you know they soak up a lot of wafer capacity and so as a result creating shortages in broader dram and then also that's creating shortages in broader nand um and so that you're seeing this ripple effect into other areas like sandisk for example has seen you know tremendous appreciation in their stock price so you know i think that there will be these areas where um you know the economic profit can be distributed to even more that didn't participate as much over the last few years."

Kim Forrest explains the current upward trend in memory prices, attributing it to high demand for High Bandwidth Memory (HBM) used in AI systems. Forrest details how this demand strains wafer capacity, leading to shortages in DRAM and NAND, which in turn positively impacts related companies like SanDisk. Forrest suggests that this ripple effect will distribute economic profits to a broader range of companies in the sector.

"Yeah we think about it as kind of two sides of the coin right one is the customer experience that you've mentioned and one is how we build the company on customer experience automation and ai allows us you know to personalize how much rewards you get how we talk to the customer how the experience even shape shifts for the customer and more and more investing in that so it feels radically different to anything that a bank has ever provided in the past for these brands or for these customers the really interesting thing and maybe it's not um as apparent on the surface is we get to rebuild what looks like a financial company or even a fintech um in the time of automation and ai so over the last year we've grown the business by almost 300 and we've grown headcount by 20 and every turn we get to choose are we going to build a new team here are we going to invest in technology and automation so we don't have to hire a team of people we can actually scale the company using technology which is an amazing opportunity that i don't think anybody has ever had."

Dara Murphy discusses how Imprint