To The Moon

$IBM(IBM)$

IBM delivered one of the clearest signals yet that the AI boom is reshaping corporate technology budgets.

The company’s shares plunged about 25% after it released preliminary second-quarter results below Wall Street expectations. IBM expects quarterly revenue of roughly $17.2 billion, up only 1% year over year and below the $17.86 billion analysts expected. Adjusted earnings are projected at $2.93 per share, versus the $3.02 consensus estimate. (Reuters)

The headline numbers were disappointing, but the explanation was even more important.

IBM CEO Arvind Krishna said that during the final weeks of June, corporate clients redirected part of their quarterly capital spending toward servers, storage and memory. Customers wanted to secure supply-constrained infrastructure before anticipated price increases, while several planned IBM software and mainframe-related deals were delayed. (Reuters)

That raises a much bigger question for the technology sector:

Are companies shifting limited IT budgets away from traditional software projects and toward AI infrastructure?

What actually went wrong at IBM?

IBM’s warning reflects several issues happening at the same time.

First, corporate spending priorities changed faster than IBM expected. Clients moved money toward physical infrastructure that could become more expensive or harder to secure.

Second, IBM admitted it did not adjust quickly enough. Numerous large deals failed to close on schedule, contributing to the quarterly shortfall.

Third, the market has become far less tolerant of weak execution among traditional software and IT-services companies.

IBM’s software revenue is still expected to grow around 5%, while infrastructure revenue is projected to decline 7%. That suggests demand has not collapsed across the entire company. The market reaction reflects concern that IBM may be losing budget priority during the most aggressive phase of the AI infrastructure buildout. (Business Insider)

The selloff also spread beyond IBM. Microsoft, ServiceNow, Salesforce and Intuit fell between roughly 2% and 5%, showing that investors viewed the warning as a broader software-sector signal. (Reuters)

Why are companies buying hardware first?

Enterprise AI projects require more than a software subscription.

Companies may need to secure:

• GPUs and AI accelerators

• Servers and storage systems

• DRAM, HBM and flash memory

• Networking switches and optical connections

• Data-center capacity, electricity and cooling

When supply is tight and prices are expected to rise, procurement teams naturally prioritize equipment with long delivery times.

A software upgrade can often be delayed for one quarter. A scarce server cluster or memory allocation may be much harder to replace later.

IBM’s warning suggests that the infrastructure shortage is now affecting real corporate purchasing decisions, rather than only the capital-expenditure plans of Microsoft, Amazon, Meta and Google.

The AI boom may be creating a “budget migration”

Until now, investors mainly focused on how much hyperscalers were spending on AI.

IBM provides another perspective: some of that spending may be coming at the expense of other technology projects.

Corporate IT budgets are not unlimited. When more money goes toward servers, chips, memory and networking, less may be available in the same quarter for:

• Software subscriptions

• Consulting projects

• Mainframe upgrades

• Digital-transformation programs

• Non-essential IT spending

That could deepen the gap between AI infrastructure winners and software companies still trying to prove that their AI products generate incremental revenue.

Potential beneficiaries

AI chips and custom computing

$NVIDIA(NVDA)$ remains the core supplier of AI accelerators.

$Advanced Micro Devices(AMD)$ offers AI accelerators and data-center processors.

$Broadcom(AVGO)$ benefits from custom AI chips and data-center networking.

$Taiwan Semiconductor(TSM)$ remains central to advanced chip manufacturing and packaging.

If companies continue prioritizing AI capacity, compute remains near the top of the purchasing list.

Servers and networking

$Dell Technologies(DELL)$ supplies enterprise and AI servers.

$Hewlett Packard Enterprise(HPE)$ provides servers, networking, storage and enterprise infrastructure.

$Super Micro Computer(SMCI)$ offers high AI-server exposure, although its margins and order execution can be volatile.

$Arista Networks(ANET)$ supplies high-speed data-center networking.

$Marvell Technology(MRVL)$ and $Credo Technology(CRDO)$ provide exposure to custom silicon and high-speed connectivity.

AI systems need complete clusters. GPUs alone cannot operate without servers, switches and fast interconnects.

Memory and storage

$Micron Technology(MU)$ is one of the clearest U.S.-listed plays on HBM, DRAM and data-center memory.

$SanDisk(SNDK)$ provides NAND flash exposure.

$Western Digital(WDC)$ is tied to enterprise storage and data-center demand.

$Seagate Technology(STX)$ benefits from demand for high-capacity data storage.

IBM specifically cited clients prioritizing servers, storage and memory ahead of potential price increases, making this group particularly relevant to the current budget shift. (Reuters)

Which software companies now face a higher bar?

$IBM(IBM)$ must show that delayed contracts eventually close and that Red Hat and software growth remain intact.

$ServiceNow(NOW)$ needs to prove that AI agents can increase contract values rather than simply add features to existing subscriptions.

$Salesforce(CRM)$ must demonstrate meaningful revenue from Agentforce and other AI products.

$Adobe(ADBE)$ needs clearer evidence that generative AI supports pricing power and user growth.

$Intuit(INTU)$ must show that AI strengthens customer retention and monetization.

$Atlassian(TEAM)$ faces pressure from both corporate budget changes and new AI coding and collaboration tools.

For software companies, launching an AI feature is no longer enough.

Investors increasingly want answers to four questions:

Will customers pay extra?
Will AI increase contract values?
Can AI revenue cover additional development costs?
Will infrastructure spending delay software renewals?

Will hardware keep winning over software?

The current budget migration may be temporary.

Companies first need the infrastructure required to run AI. Once chips, servers and storage are installed, they will still need databases, cybersecurity, workflow software, observability tools and business applications.

The spending cycle may develop in three stages:

1. Secure chips, servers, memory and networking

2. Build data platforms, security and model infrastructure

3. Expand enterprise AI applications and agent-based software

Hardware is currently receiving the highest purchasing priority. Software could regain momentum once the infrastructure layer is in place.

The greatest risk sits with software companies whose AI products lack differentiation, fail to generate additional spending and cannot protect their existing businesses.

TigerComments Take

IBM’s 25% plunge looks bigger than a single-company earnings miss.

It suggests the AI boom is causing a real reallocation of corporate IT budgets.

In the near term, companies are prioritizing scarce infrastructure with long delivery times: chips, servers, storage, memory and networking.

Software companies now have to prove that their AI products can generate new revenue rather than simply consume more customer budget.

The next signals to watch are:

• Whether IBM’s delayed deals return when it reports full results

• Whether server and memory demand remains elevated

• Whether ServiceNow, Salesforce and Adobe report measurable AI revenue

• Whether software spending recovers after the current hardware procurement cycle

Where would you position for this IT-budget shift?

A. AI chips: NVDA / AMD / AVGO
B. Servers and networking: DELL / HPE / ANET
C. Memory and storage: MU / SNDK / WDC
D. Software rebound: IBM / NOW / CRM

Do you think companies are only temporarily prioritizing hardware, or is AI creating a longer-term squeeze on traditional software budgets?

Disclaimer: This post is for market discussion only and does not constitute investment advice. Investing carries risk.