A tale of two tech sectors

Good morning,
Some wonder whether they have space for a sectional couch, others wonder where to put their 38-foot-long dinosaur skeleton. A T. rex fossil sold for a record $50M at auction to a mystery phone bidder.
In today’s edition:
- IBM’s bad day raises reallocation worries
- Corporate America flies on Starbucks
- A trending category in top SaaS vendors
IBM’s unsavory earnings appetizer raises the capex reallocation question

“We did not anticipate the magnitude of the capex reprioritization…” IBM is expected to report earnings tomorrow, and the Q2 appetizer didn’t sit right with investors: IBM stock had its worst day in nearly 60 years last week (down 25%) after the enterprise software provider reported disappointing preliminary quarterly sales.
In his letter to investors, CEO Arvind Krishna attributed the miss partly to customers shifting their capex spending to servers, chips, and storage ahead of expected price increases. While hardware supply has been tight for a while amid the AI boom, Krishna said IBM didn’t anticipate a capex reprioritization of this magnitude.
The memory hardware shortage and subsequent price hikes raise the question of whether AI hardware spending could start to cannibalize software budgets, as we saw with IBM. TrendForce expects that chip costs could rise by another 63% this quarter.
A tale of two tech sectors… Despite its name (International Business Machines), most of IBM’s revenue today comes from software and consulting. Companies have been boosting their overall capex spend to pay for AI infrastructure, but as costs rise, we could see more firms shifting capex allocation away from software and consulting services. Add in the long-running SaaSpocalypse discourse, and investors are jittery:
- The S&P Software & Services Select Industry Index is down 13% year-to-date, while the S&P Technology Hardware Select Industry Index is up 30%.
The companies selling shovels to the AI goldrush are on a tear: ASML, which makes machines used to produce cutting-edge AI chips, hiked its sales forecast last week for the second time this year on “extremely strong” orders. And TSMC recently smashed second-quarter earnings estimates, with profit up 23% from Q1 and 77% from a year earlier.
It’s not a zero-sum game… It’s too early to say that IBM is the canary in the coal mine of a broad capex reallocation. Ramp Rate data shows that for traditional SaaS vendors, the share of seat-based revenue has increased, rising from 65% in April 2025 to 76% in February of this year. Consumption-based revenue was roughly flat. And Gartner forecasts that software spend growth will accelerate to 15% this year, the second-fastest growing category after data centers. We’ll be watching upcoming software earnings for more signals.
Watercooler Data

Traveling employees need their caffeine fix. Whether they’re at the airport or in between meetings, the chain they’re charging to the company card the most often is Starbucks.
- Starbucks is the most-expensed restaurant on business trips booked on Ramp, and it leads by a latte (sorry, had to). Over the past 18 months, the java giant had more transactions than the next three chains combined.
Starbs sees about 6x more travel transactions than coffee rival Dunkin, though that’s likely in part due to its larger international footprint (it has around 41K global locations, vs Dunkin’s ~14K).
Starbucks has returned to growth under CEO Brian Niccol’s turnaround plan, which includes bringing back the original coffeehouse experience (ceramic mugs, free condiment bar, Sharpie-signed cups).
In April, Starbucks reported its second-straight quarter of traffic growth following a two-year slump in transactions. We'll see if that growth streak stuck when the company reports next week.
How may AI help you? AI support bots emerge as one of the clearest enterprise AI use cases

AI support bots PolyAI and Sierra are trending on Ramp’s latest list of Top Software Vendors.
We often talk about the difficulty in measuring AI’s ROI… In most cases, there isn’t yet a clear way for companies to measure the relationship between token spend and productivity.
But one area where AI’s effect can be directly measured is with AI support bots. You can track the share of customer service issues fixed by AI, the average time to close a ticket, and customer satisfaction. Then you compare that to pre-AI metrics.
“Outside of coding agents, AI support bots are likely the most commercially-advanced use-case for AI,” says Ramp’s lead economist Ara Kharazian.
Read more from Ramp Economics Lab.
Signals Shortlist
- BlackRock leads $12B financing for new Meta data centers in Texas (The Wall Street Journal)
- Big banks' record-smashing Wall Street profits are increasingly tied to AI (Yahoo Finance)
- Gas prices hit $4 again after oil soars more than 15% in a week (NBC News)
- US producer prices post largest drop in 14 months; inflation risk still tilted to upside (Reuters)
- OpenAI’s first device will be a movable, screenless AI speaker (Bloomberg)
🗓️ Leading Events:
Tuesday, July 21: Earnings expected from GM
Wednesday, July 22: Earnings expected from Alphabet, Tesla,IBM, Texas Instruments, Alaska Airlines, and ServiceNow
Thursday, July 23: Earnings expected from Intel, American Express, Lockheed Martin, Southwest Airlines, American Airlines, T-Mobile, Verizon, Honeywell, and Blackstone
Friday, July 24: Happy Friday¯\_(ツ)_/¯
🤝 Like what you read? Share Leading Indicators with a friend.


