Beat, Raise, and Sell: A Narrow Software Miss Buries a z17 Blowout and an FCF Guide-Up

Results vs. Consensus

Quality of Beat/Miss

• Revenue: Of the ~$400M total-revenue beat, management attributed roughly $375M to genuine constant-currency business performance and the balance to FX, so this was an operational beat, not a currency mirage. But the composition is the problem: the beat sat in Infrastructure (z17), which the Street treats as cyclical, while Software, the secular-growth engine, missed by ~$40M.
• Margins: High quality. The 230bps of operating gross-margin expansion came from portfolio mix (software is ~45% of revenue at 84% gross margin) and structural productivity savings the company is now annualizing at ~$4.5B, not from a one-time item. Adjusted EBITDA grew 16% to $4.7B.
• EPS: The $2.80 operating result is clean. The effective operating tax rate of 17% is in line with the mid-teens full-year framework, and the 15% operating-EPS growth tracked profit growth rather than leaning on buybacks (diluted share count actually rose ~1.5% YoY to 948M as IBM prioritizes dividends and M&A over repurchases).

Year-over-Year and Sequential Detail

Sequentially (against a derived Q1 2025), the headline jump overstates momentum because of seasonality and the z17 launch: Total revenue rose from ~$14.54B to $16.98B (+16.7%), Infrastructure from ~$2.89B to $4.14B (+43.6%) as the mainframe shipped, and operating EPS from $1.60 to $2.80. The cleaner read is the year-over-year line above, where the dispersion across segments, software decelerating organically while infrastructure inflects, is the whole story of the quarter.

Segment Performance

Software — the engine missed a beat

Software revenue of $7.387B grew 10% (8% at constant currency) with segment profit of $2,296M at a 31.1% margin. Underneath, the dispersion matters. Red Hat (reported within Hybrid Cloud) accelerated a point sequentially to ~14.5% constant-currency growth, with OpenShift revenue up more than 20% and its ARR now at $1.7B; Automation grew 14% at constant currency with HashiCorp, closed in Q1, contributing its first full quarter and management citing a second-half pipeline 3x the prior year; Data grew 7%. The drag was Transaction Processing, down 2% at constant currency, as mainframe clients reprioritize budget toward hardware at the start of a cycle.

"Software grew 8% this quarter as we continue to benefit from our high-value annual recurring revenue base, of which grew to $22.7 billion, up 10% since last year. Red Hat growth accelerated 1 point sequentially to 14%." — James Kavanaugh, CFO

The $22.7B ARR base (management rounds to "$23 billion") growing double digits is the structural support for the "approaching double digits" full-year software guide. But the Street's $7.43B estimate was set against a re-rated stock, and the ~$40M shortfall, with organic growth decelerating to a low-single-digit rate once HashiCorp and DataStax are stripped out, is what investors fixated on.

Assessment: The quality of the software franchise (recurring, high-margin, share-gaining in containers and Linux) is not in question; the rate of organic acceleration is. Until Transaction Processing turns back up on the z17 stack multiplier and the M&A contribution anniversaries, reported software growth will keep running ahead of organic, and the market has shown it will punish even a small miss here.

Infrastructure — a z17 launch for the record books

Infrastructure revenue of $4.142B grew 14% (11% at constant currency) at a 23.3% segment margin, up 540bps year over year. The driver was IBM Z, up 67% at constant currency, on what management called the strongest first quarter of any mainframe launch in IBM's history, with more than 100% of a generation's MIPS capacity shipped into the market in the opening quarter of z17. Distributed Infrastructure fell 17% on a Power product-cycle trough ahead of the July Power11 launch, and Infrastructure Support declined 3%.

"We're off to the strongest-ever first quarter start in the history of IBM around mainframe in a launch quarter. We shipped over 100% MIPS capacity into the marketplace on z17... we run that mainframe platform as a stack economic play, $3 to $4, a platform multiplier over the cycle." — James Kavanaugh, CFO

Assessment: This is the most underappreciated line in the quarter. The mainframe is not a growth business in isolation, but a z17 launch this strong drags Transaction Processing software, storage, and services with it over the following two years. Management explicitly raised the full-year infrastructure contribution rather than calling it a pull-forward, which is the right tell that they see durable stack pull-through, not a one-quarter spike.

Consulting — stabilizing, not recovering

Consulting revenue of $5.314B was up 3% reported but flat at constant currency, with segment profit of $562M at a 10.6% margin. Backlog stood at $32B (up ~8% at spot rates), trailing-twelve-month book-to-bill was 1.14, and signings fell 18% on tough prior-year renewal comparisons. The bright spot is GenAI: the consulting GenAI book of business crossed $6B inception-to-date, now ~17% of backlog and over 20% of bookings, at a 3-point margin premium to the segment.

"Our signings in the quarter... we were down 18%. That was entirely driven by last year's large early renewals... our net new business penetration was up 13% year-to-year... read that, 200-plus new clients we acquired already in our consulting business year-over-year." — James Kavanaugh, CFO

Assessment: Consulting is the part of IBM most exposed to discretionary deferral and to GenAI cannibalizing legacy bodies-on-seats work. Management's "green shoots" framing (net-new client penetration, shorter backlog duration improving revenue realization) is plausible but unproven; flat constant-currency revenue with down signings is a show-me story. We model Consulting as a low-single-digit grower at best for the year.

Key KPIs

Key Topics & Management Commentary

Overall Management Tone: Confident and notably more upbeat on the macro than a quarter earlier, with the CEO explicitly upgrading his framing from "cautious optimism" to "optimism." Management leaned into the productivity and free-cash-flow story and pre-empted the software-deceleration question in prepared remarks, which reads as awareness that the segment was the quarter's vulnerability. Where they were least convincing was on why a stronger macro and an ahead-of-pace first half did not warrant a revenue guidance raise.

1. The Software Deceleration Question

The first analyst question went straight at the soft spot: organic software growth appears to have decelerated to a 3-4% range for a third consecutive quarter, which sits uncomfortably against the "approaching double digits" full-year guide. Management's answer was a category-by-category bridge: Red Hat contributing ~3.5 points, Automation ~3.5 points and accelerating on HashiCorp synergies, Data delivering to model, and Transaction Processing the swing factor that should return to growth in the second half as the z17 stack multiplier kicks in.

"Between accelerating growth continued on Red Hat, the Automation synergistic value on Hashi, Transaction Processing capitalizing on that multiplier effect of a very strong start on mainframe, we see that organic growth moving forward. And oh, by the way, we're comfortable with third quarter guidance, and third quarter guidance already accelerates organic growth." — James Kavanaugh, CFO

Assessment: The bridge is internally coherent and the TP-recovery logic has a precedent in the z16 cycle. But it asks investors to underwrite a second-half organic re-acceleration that has not yet shown up in the numbers. This is the single most important thing to track next quarter.

2. The z17 Stack Multiplier

Management repeatedly returned to the idea that the mainframe is not a hardware line item but the anchor of a $3-to-$4 revenue multiplier across software (Transaction Processing), storage, and services over a cycle. The unusually strong z17 launch therefore matters less for the infrastructure line this quarter than for what it pulls through over the next two years.

"While storage was impacted by the new IBM Z cycle as clients prioritized hardware spend, our early strength in z17 and the growth in the installed MIPS capacity drives a long-term benefit given the 3 to 4x Z stack multiplier." — James Kavanaugh, CFO

Assessment: Credible and, if anything, under-modeled by the Street, which keeps treating each mainframe cycle as a pull-forward. The risk is timing: the multiplier is real but lumpy, and a weak z17 attach rate would undercut the whole second-half software re-acceleration thesis.

3. Red Hat, OpenShift, and the Virtualization Land-Grab

Red Hat is the cleanest growth story in the portfolio, and management framed the post-Broadcom/VMware virtualization disruption as a structural share-gain opportunity. Clients wanting one platform across containers and virtualization default to OpenShift; IBM cited more than $300M of virtualization bookings through three quarters and an accelerating pipeline.

"If people are wanting a container platform that goes across public cloud and on-premise, I will tell you that we are the leading answer for that... there is a set of clients who would prefer a common answer across containerization and virtualization. And then by default, we tend to win those." — Arvind Krishna, CEO

Assessment: This is the most durable secular tailwind in the software stack and the part of the thesis we have the highest confidence in. VMware-account disruption is a multi-year migration cycle, and IBM is positioned as the default consolidation target.

4. The Productivity Flywheel and the Margin Raise

The clearest beat-and-raise of the quarter was on profitability, not revenue. IBM raised its annualized productivity-savings target to ~$4.5B (from $3.5B exiting 2024) by embedding its own AI across more than 70 internal workflows, which funds margin expansion and reinvestment simultaneously. The company raised its full-year operating pre-tax margin expansion to ~1 point.

"We are seeing extreme penetration around that, and that's given us guidance and confidence to raise that to $4.5 billion. That flows to operating margin... that adjusted EBITDA is now going to be low teens growth." — James Kavanaugh, CFO

Assessment: This is the most reliable lever in the IBM story and the foundation of the free-cash-flow case. Self-applied AI ("client zero") is both a margin driver and a reference-sell for the software portfolio, a genuine two-for-one.

5. GenAI Book of Business Crosses $7.5B

The cumulative GenAI book of business reached more than $7.5B inception-to-date in roughly six quarters, split about $6B Consulting and $1.5B Software, with management characterizing momentum as accelerating quarter over quarter. Notably, consulting GenAI now exceeds 10% of segment revenue at a 3-point margin premium.

"Our GenAI book of business now stands at over $7.5 billion inception-to-date with momentum accelerating quarter-over-quarter." — Arvind Krishna, CEO

Assessment: The headline number is impressive, but "book of business" blends software transactional revenue, SaaS ACV, and consulting signings, so it is a momentum indicator rather than a clean revenue figure. Still, the directional acceleration is the best evidence that IBM is monetizing AI rather than just talking about it.

6. AI: Incremental or Cannibalistic?

Asked whether AI spend is incremental or cannibalizing the rest of IBM's wallet, the CEO drew a sharp line: hardware and enabling-software layers are purely incremental, while the cannibalization shows up in labor, both clients' internal labor and third-party consulting bodies, as customers redirect budget toward AI-enabled transformation.

"When you get to the enabling software layers, that is purely incremental. There is no cannibalization there... But some other parts of customer application development do tend to get cannibalized towards AI." — Arvind Krishna, CEO

Assessment: An honest answer that quietly explains the Consulting pressure: AI is a tailwind for IBM's software and transformation work but a headwind for traditional staff-augmentation consulting. The net depends on whether IBM's "transformative project" mix outruns the cannibalized legacy work, and that is still unproven.

7. The Guidance-Raise That Wasn't (Revenue)

A pointed question challenged why IBM raised free cash flow, margin, and productivity but not constant-currency revenue, given the more optimistic macro and an ahead-of-pace first half. Management's answer leaned on conservatism: a half-year of revenue still to sell, ~$40B of revenue to go, and a preference to let the year prove out.

"We still got $40 billion worth of revenue to go... We feel confident about the position we're in... And yes, we have upside and conservatism? Absolutely, but that's what you would expect of us." — James Kavanaugh, CFO

Assessment: Defensible discipline, but it cuts against the bullish narrative. If the macro and software were genuinely re-accelerating, a revenue raise would have followed the cash-flow raise. The decision to hold revenue flat is the most honest signal in the quarter that the software re-acceleration is a forecast, not yet a fact.

8. Federal and the M&A Environment

Management was constructive on two adjacencies. On federal, the CEO sees spending priorities shifting from first-half cost-cutting toward modernization and AI adoption, with IBM a credible incumbent. On M&A, he characterized the regulatory environment as newly "rational," signaling appetite for more deals after the HashiCorp and DataStax closings.

"What we've seen over the last 4 months has made us optimistic that we are now in a rational regulation environment where M&A that makes sense will get approved in reasonable time frames." — Arvind Krishna, CEO

Assessment: Both are optionality rather than base-case drivers. The M&A signal is double-edged for shareholders: it supports software growth synergies but also explains why reported software growth keeps outrunning organic, and why the cash going to deals is not going to buybacks.

Guidance & Outlook

The shape of the guide tells the story: everything cash- and margin-related went up, the revenue line held, and the one revenue sub-line that moved (Transaction Processing) moved down. Management framed Infrastructure as contributing ~1.5 points to full-year IBM growth on z17 strength and reiterated software "approaching double digits," but the refusal to raise the top-line number despite a $1B+ upward revision to Street revenue estimates since January is the constraint on the bull case.

Implied second-half ramp: To hit "approaching double digits" in software for the year against a first half that ran high-single-digit organic, IBM needs Transaction Processing to inflect positively and the HashiCorp/Automation pipeline (cited at 3x last year) to convert. Street at: consensus had already moved up over $1B on revenue since January; management's "comfortable with Q3 consensus" caps near-term upside. Guidance style: characteristically conservative on revenue, more forthcoming on cash and margin, consistent with the pattern of letting the top line prove out before raising.

Analyst Q&A Highlights

Organic software growth and the path to the full-year guide

The dominant line of questioning challenged whether organic software growth had quietly decelerated for a third straight quarter, and how that reconciles with a full-year guide approaching double digits. Management did not dispute the deceleration; it reframed the second half around Transaction Processing recovery and Automation acceleration.

Q: "It looks like organic growth decelerated in the quarter to maybe between 3% to 4%. Could you confirm that? ... what are some of the other puts and takes that we should be expecting in the business?"
— Wamsi Mohan, Bank of America

A: "Halfway through the year right now, we actually feel even more confident about the year of approaching double digits... we're comfortable with third quarter guidance, and third quarter guidance already accelerates organic growth."
— James Kavanaugh, CFO

Assessment: Management engaged the question directly rather than dodging it, which is to their credit, but the answer is a forecast of re-acceleration, not evidence of it. The Q3 "already accelerates" claim is now a checkable commitment.

Why raise cash and margin but not revenue

A second pointed exchange pressed on the asymmetry of the guidance raise, with the optimistic macro and ahead-of-pace first half seemingly arguing for a revenue raise too. Management defended holding the line as conservatism with half a year still to sell.

Q: "What is causing you not to raise guidance if you feel better about the economy? ... is it conservative to say that you're comfortable with The Street for the 3Q, when all that is looking actually better and you're already at your 5%-plus and already ahead of pace on free cash flow?"
— Ben Reitzes, Melius Research

A: "We still got $40 billion worth of revenue to go. We feel confident about the position we're in... And yes, we have upside and conservatism? Absolutely, but that's what you would expect of us."
— James Kavanaugh, CFO

Assessment: The most revealing exchange of the call. The decision to raise everything except revenue is the clearest signal that the software re-acceleration remains a projection. Investors looking for confirmation of the growth thesis did not get it here.

Software growth durability into 2026

Looking past the current year, an analyst asked whether software growth could improve from the ~10% 2025 level into 2026. The CEO expressed unusually high conviction on Red Hat, Automation, and Data sustaining or improving, with Transaction Processing normalizing back to its long-term model.

Q: "Heading into 2026... do you think there is potential for improving both organic and overall software growth heading into 2026 from the current 10% levels this year?"
— Jim Schneider, Goldman Sachs Asset Management

A: "All those 3 parts of the portfolio, I would give you equal or higher growth rates going into 2026. Now you come to TP... I would fully expect TP to return to its long-term model, which is in between low and mid-single digits."
— Arvind Krishna, CEO

Assessment: A forward-leaning answer that frames 2025 as a trough year for organic software, not a ceiling. If TP normalizes and the M&A contribution anniversaries cleanly, the 2026 organic comparison could look materially better, which is the crux of any upgrade case from here.

AI: incremental spend versus cannibalization

A nuanced question probed whether the GenAI book of business is incremental to client wallets or is pulling budget from IBM's other lines. The CEO's layered answer was candid about where cannibalization lands.

Q: "How are customers prioritizing AI over non-AI? Is there cannibalization in other areas of client spend or is it incremental to client spend?"
— Erik Woodring, Morgan Stanley

A: "When you get to the enabling software layers, that is purely incremental... I will tell you that the cannibalization is going to come from the fourth part. People are looking at their own internal labor expenses, and... their third-party labor expenses."
— Arvind Krishna, CEO

Assessment: This quietly explains the Consulting overhang. AI lifts IBM's software and transformation work while pressuring traditional staff-heavy consulting; the segment outcome hinges on mix shifting fast enough toward transformative projects.

watsonx Orchestrate and IBM's right to win in agentic AI

The closing question focused on the crowded agent-orchestration space and what differentiates IBM's Orchestrate. The CEO grounded the answer in heterogeneity, IBM integrating both its own ~70-80 agents and ~70 third-party agents rather than locking clients into a single vendor's agents.

Q: "On the Orchestrate layer, that seems to be a space that a lot of companies have started to talk about... If you could just talk a little more about your right to win in that space and how you're positioned."
— Matt Swanson, RBC Capital Markets

A: "It cannot be about just our own agents... We also integrate in about 70 agents from third parties... That is the client where we are going to win in Orchestrate, and that is what distinguishes us from those who are kind of focused on primarily their own agents."
— Arvind Krishna, CEO

Assessment: A coherent positioning that plays to IBM's historical strength in heterogeneous, open environments. Whether "open orchestration" translates into durable revenue against hyperscaler-native agent stacks is the open question, but the strategy is differentiated rather than me-too.

What They're NOT Saying

1. A clean organic software growth number: Management bridges software growth by category and contribution points but never states a single organic-growth figure, leaving the Street to back into the ~3-4% the first question implied. The reluctance to put the number on the table is itself a signal.
2. A revenue guidance raise: With cash, margin, and productivity all raised, the conspicuous omission is the top line. Holding constant-currency revenue at "at least 5%" despite a stronger macro is the loudest thing unsaid.
3. Buyback commentary: With ~$7.8B spent on acquisitions in the first half and the share count rising year over year, capital is going to M&A and dividends, not repurchases. Management did not address whether buybacks return once the M&A pipeline clears.
4. Consulting margin trajectory: Plenty of "green shoots" on backlog and net-new clients, but no commitment on when flat constant-currency revenue turns into actual growth, or where segment margin (10.6%) goes from here.

Market Reaction

• Pre-print setup: IBM closed at $282.01 on 2025-07-23, up 28.3% year to date and 53.2% over the trailing twelve months, near a multi-year-high valuation. The stock had been comprehensively re-rated as a software-and-AI growth story rather than a value/dividend name.
• Reaction session (2025-07-24): Shares gapped down 7.4% at the open and closed at $260.51, down 7.6% ($21.50), on volume of 22.6M (6.1x the 30-day average). It was IBM's largest single-day decline in roughly 15 months. The S&P 500 was flat (+0.1%) that session, so the move was entirely idiosyncratic.

The reaction is a textbook expectations reset. A stock priced for software acceleration met a quarter where software decelerated organically and missed the segment estimate by ~$40M, even as everything cash- and margin-related got better. The market chose to weight the secular line (software) over the cyclical beat (mainframe) and the cash raise, and de-rated the multiple accordingly. The selloff improves the entry multiple but does not, by itself, resolve the question the quarter raised.

Street Perspective

Debate: Is IBM a software-growth story or a cash-return story?

Bull view: The mix shift is real, software is ~45% of revenue with a $22.7B recurring base growing double digits, Red Hat is gaining share, and AI is monetizing; the multiple re-rating is justified by a structurally faster-growing, higher-margin business.

Bear view: Strip out M&A and organic software is only growing low-single-digits; this is still a low-growth, cash-generative incumbent whose ~26x pre-print multiple priced in an acceleration that the numbers do not yet support.

Our take: The truth is in between, and that is precisely why it is a Hold. The franchise quality and cash generation are excellent and the z17/Red Hat tailwinds are durable, but the valuation now requires the organic software acceleration to be real, and this quarter the company itself declined to raise revenue. We need the proof before the multiple.

Debate: Does the z17 cycle de-risk or distract from the software story?

Bull view: A record mainframe launch with a $3-4 stack multiplier pulls software and services through for two years, and the Street is under-modeling it as a one-quarter pull-forward.

Bear view: Leaning on the mainframe for the beat is exactly the "old IBM" story investors thought they had moved past; cyclical hardware should not anchor a growth re-rating.

Our take: The bulls are closer to right on the mechanics, the multiplier is genuine and under-appreciated, but the bears are right about the optics. A quarter where the beat comes from mainframe and the miss comes from software will, fairly or not, keep the re-rating capped.

Debate: Is Consulting stabilizing or structurally challenged by AI?

Bull view: Backlog is up 8%, net-new client penetration is up 13%, GenAI consulting is over 10% of revenue at premium margins; the segment is bottoming and AI is a tailwind.

Bear view: Flat constant-currency revenue and signings down 18% in a business directly exposed to AI replacing billable labor; "green shoots" have been promised before.

Our take: The bear case has the burden of proof on its side until constant-currency revenue turns positive. We treat Consulting as a flat-to-low-single-digit grower and would not underwrite a recovery we cannot yet see.

Model Update Needed

Valuation impact: The selloff to $260.51 takes the forward operating-earnings multiple from roughly 26x to roughly 24x on consensus FY2025 operating EPS near $10.95, still a multi-year high and well above IBM's historical mid-teens range. At that multiple, the stock already capitalizes the cash-flow and margin strength; further upside requires the organic software acceleration to materialize. We see fair value broadly around the post-print level and would want either a cheaper multiple or hard evidence of software re-acceleration to turn constructive.

Thesis Scorecard Post-Earnings

As this is our initiation of coverage, the scorecard below establishes the standing thesis we will grade against in subsequent quarters.

Overall: The bull pillars on cash, margin, mainframe, and AI monetization are confirmed; the swing pillar (organic software acceleration) is unproven, and all three bear points are live. The franchise is high-quality but fully priced.

Action: Initiate at Hold. We would turn constructive on (1) a quarter of confirmed organic software re-acceleration with Transaction Processing back to growth, or (2) a more attractive entry multiple. We would turn cautious on a Consulting deterioration or evidence the z17 stack multiplier is failing to pull software through.