SouthernValue

Bullish cloud market datapoints from a call with the CEO of cloud cost management firm Archera, hosted by $JEF. Terms becoming more favorable for CSP: - Customers trying to renew 3yr deals are being told they can only renew 1yr, meaning smaller discounts (higher price) - To reserve capacity for 3yrs are being told they need to make bigger commitments (min cluster size going up) with an annual spend cadence (helps clouds optimize capacity utilization) - Prices going up for GPUs and CPUs - Supply > demand, so if customer push on these terms, clouds are “pushing clients to $ORCL or the Neoclouds” Neocloud momentum isn’t at the clouds expense… it’s because they can’t satisfy all of the current market demand. Positive for basically everything in AI/DC trade, $AMZN, $GOOG… but sentiment on $MSFT particularly is still too negative relative to the AI trade. And people don’t appreciate how much ROIC on the existing DC fleets for CSPs will go up.

In this particular case (SBUX headline on potentially replacing two inventory systems from IBM/MSFT w/ home built app, beginning in 2027), I'm not sure there is much signal. Its not clear what $MSFT product $SBUX is replacing - but SBUX was a known AX 2012 customer at some point, a perpetual license product that went ex-support in 2023. Linkedin job postings since mention $ORCL / $WDAY / $WK as core accounting systems, but not Dynamics 365. SBUX may still be on an old MSFT perpetual license, perhaps AX or perhaps Great Plains which is being sunset in 2029, or perhaps they built a custom app on Azure. If they are replacing 1 custom app with another one, its a net-neutral to MSFT. If they replace an old perpetual license with a new modern app on Azure, its a net positive to MSFT. Migrating off old systems is a tale as old as time in IT. And in horizontal categories like inventory mgmt, warehouse mgmt, for Fortune 500 companies, internally-built systems will always be on the table to consider bc few SaaS apps can scale to that degree and serve all the particular needs of a customer that size (why $TOST is an SMB product). Personally, I dont see this as evidence of disruption.

$IBM - surprised this is taking all of SaaS down with it. My ~subjective~ attribution of the IBM miss: • 50–60% IBM-specific execution and mainframe product-cycle issues • 25–35% emerging structural pressure on legacy/mainframe software • 10–20% broader enterprise deal timing and budget reprioritization Honestly, probably net neutral-to-bullish $MSFT considering the commentary around cybersecurity (people always forget $MSFT has the largest cybersecurity business in the world). And it’s the hyperscalers who should benefit from mainframe workload migration.

$APO feels well positioned to help finance this DC buildout. Hyperscaler bond issuance: ~$28B/yr from 2020-24. $121B in 2025. $159B in the first five months of 2026. Street expects 250-300B+ this year, and ~$1.5T of AI-related paper over five years (which would make five tickers 15-20% of the IG index). Public DCM can’t smoothly absorb that kind of concentrated net supply. Index limits, duration appetite, and single-name exposure, etc. The answer is probably manufacturing the paper into shapes every pocket of capital can hold… IG-rated SPV bonds for insurers, ABS, mezz for credit funds. Originate, structure, tranche, distribute, etc. Apollo seems pretty well positioned as a solutions provider here. Will take their cut along the way.

$IBM issue clearly mainframe related… mainframe hardware and mainframe software down, RedHat accelerating. “Given this was the strongest start to a mainframe program in our history, we expected infrastructure revenue to decline low-single digits for the year, beginning this quarter," Krishna wrote on Tuesday. "What played out was worse than our expectations, driven by a shortfall in our Z performance and the associated software stack, primarily in transaction processing. In the last few weeks of June, we saw clients shift their quarterly capex spend toward servers, storage, and memory purchases to secure supply-constrained infrastructure ahead of expected price increases. This dynamic impacted client buying patterns." He added that the company did not anticipate the "magnitude of the capex reprioritization." Meanwhile, IBM's adjusted earnings climbed 5% to $2.93 per share, compared to expectations for $3.01 per share. "While our second-quarter results are disappointing, our performance in many areas showed strength, reinforcing the conviction we have in our portfolio and strategy," Krishna wrote. He pointed to Red Hat software growth accelerating to 11%.

"Be greedy when IG Corporates are fearful" $400B of cash with no ability to deploy it... meet the AI datacenter buildout. There might be a chance for some classic $BRK.B preferred equity deals if IG markets tap out while hyperscalers are still <1x levered. $GOOG offering involvement makes me think more likely.