Another Beat and Record Q1 Cash Flow, Yet the Stock Falls Again: the AI-Disruption De-Rating Has Overshot

This recap updates our coverage initiated with the Q4 2025 report (Hold, January 29, 2026).

Results vs. Consensus

Assessment: Revenue

The +6% constant-currency print is IBM's strongest first quarter in over a decade and beat the Street by ~2%, with the upside concentrated in infrastructure (+12% cc, well ahead of the $3.16B StreetAccount bar) and software (+8% cc). Reported growth of +9% now carries a currency tailwind, a reversal from 2025's headwind, and management flagged currency as a modest full-year positive. The composition is healthy: the beat came from the high-quality software and mainframe franchises, not from a low-margin surge.

Assessment: Margins

Operating pre-tax margin expanded 140 bps to 13.4%, "ahead of expectations," with operating gross margin up 110 bps and adjusted EBITDA up 17% to $4.0B. Infrastructure segment margin jumped 720 bps on z17 mix and volume. The one soft spot was consulting, where segment margin declined modestly on currency mix and reinvestment. The margin story remains the productivity flywheel doing exactly what management said it would, and it absorbed early Confluent dilution without derailing the full-year +1-point margin guide.

Assessment: EPS & Cash

Operating EPS of $1.91 grew 19% year-over-year, roughly double the revenue growth rate, which is the operating-leverage-to-cash story in miniature. GAAP EPS from continuing operations of $1.28 (up from $1.12) had no unusual tax distortion this quarter, unlike Q4. The headline remains free cash flow: $2.2B is a Q1 decade record and keeps the company firmly on track for its ~$15.7B full-year guide, which funds the newly raised dividend and the Confluent integration.

Segment Performance

Software: the optical downtick that isn't a deceleration

Software grew 8% cc, and the headline invites a "deceleration from 11%" narrative that the mix does not support. Management explained that fourth quarter is ~30% transactional revenue versus ~10% in the first quarter, and that the entire three-point sequential swing is that transactional mix; the underlying recurring annuity (ARR $24.6B, +10%) actually accelerated. Critically, Red Hat re-accelerated to 10% cc, precisely the "bridge from 8%" question we flagged in January.

"The entire 3-point drop quarter-to-quarter is the fundamentals of the mix of the portfolio. In fourth quarter, we have about 30% of our business in the fourth quarter is transactional. In the first quarter, that's about 10%. When you look at the underpinnings of the core annuity by itself, we're actually accelerating." — Jim Kavanaugh, CFO

Assessment: This directly resolves the biggest question from our initiation. Red Hat's re-acceleration to 10% and the annuity's acceleration de-risk the full-year software target, which management raised to 10%+ (helped by Confluent's early close adding to the data line). The optical downtick is a seasonality artifact, not a demand problem.

Infrastructure: a "strongest-ever cycle" that keeps outrunning the guide

Infrastructure grew 12% cc with IBM Z up 48% cc, distributed infrastructure up double digits (Power11 and new AI-enabled flash storage), and segment margin up 720 bps. Management reiterated the full-year infrastructure-down-low-single-digits guide but layered on a striking set of proof points: z17 has shipped 100%+ new-MIPS growth for four consecutive quarters, first-year placement value exceeded the prior (record) z16 cycle by more than $1B, and the platform's AI-inferencing capacity opens a genuinely new monetization vector.

"We monetize that value based on how many MIPS are shipped in the market and for 4 quarters in a row, on z17, we've shipped over 100% growth of new MIPS in the market, including first quarter. Higher capacity is higher monetization opportunity." — Jim Kavanaugh, CFO

Assessment: The Q4 bear point was that IBM would have to lap a peak. One quarter in, the cycle is running ahead of the prior record and adding an AI-inferencing capacity layer (450B inferences/day) that did not exist in past cycles. We now see the "down LSD" full-year guide as conservative, with the software/TP attach (3-4x stack multiplier) providing a multi-quarter tail.

Consulting: still the soft spot, and the eye of the AI storm

Consulting grew just 1% cc, the same sluggish rate as Q4, though signings returned to growth (+6%) and GenAI now represents ~40% of signings, ~30% of backlog, and a >$4B ARR run-rate. This is the segment at the center of the market's fear: that AI coding tools compress demand for IBM's labor-based modernization work. Management leaned on backlog quality and the services-as-software model.

"Signings, we returned to growth. Great quarter overall, large transformational deals around GenAI... 400 new clients captured in the first quarter... So that positions our confidence in the year of us accelerating our revenue growth around low single digits and if things go well, can we do better than that? Obviously, yes." — Jim Kavanaugh, CFO

Assessment: Consulting is the legitimate weak link and the reason we are not more aggressive than Outperform. But at ~33% of revenue, the lowest-margin third, and with signings inflecting, the market is treating a segment-level risk as a company-level one. The AI-services threat is real over years, not quarters, and IBM's asset-led delivery model is a partial hedge.

Financing: small, and quietly better

Financing grew 10% cc to $0.22B on a receivables book that is ~80% investment grade. Immaterial to the thesis; noted only for completeness.

Key KPIs

Key Topics & Management Commentary

Overall Management Tone: Confident and, if anything, more emphatic than in January, with management repeatedly pushing back on the AI-disruption narrative rather than dodging it. The posture shifted from Q4's victory-lap framing to a more assertive "the market misunderstands where we sit in the AI stack" defense, anchored by Krishna's claim that only ~4% of the software portfolio is application software. The one place the tone stayed deliberately restrained was guidance, where management chose to reiterate rather than raise despite a clear beat.

1. Where IBM Sits in the AI Stack

The defining exchange of the call was Krishna's framing of IBM's software as enabling/middleware rather than application software, which is the crux of the bull case against the disruption fear. He argued AI adoption at scale drives more consumption of the data, automation, and mainframe layers IBM owns.

"If I think of infrastructure versus applications, I think if I count right, 4% of our portfolio, if I'm to be generous, could be called an application... As people get serious about AI... many parts of our portfolio, be it Red Hat, be it Confluent, will come to be consumed more and more." — Arvind Krishna, CEO

Assessment: This is the single most important rebuttal to the sell-off. If IBM's software sits below the application layer, the Anthropic/Claude-Code threat hits a smaller, specific slice (consulting modernization) rather than the software franchise. The market is not yet crediting this distinction, which is the source of the opportunity.

2. The Deliberate Decision Not to Raise Guidance

With a clear beat and record cash flow, the most-pressed question was why guidance was only maintained. Management framed it as philosophy and prudence, not a demand warning, noting IBM essentially never raises after Q1 and pointing to the 2025 pattern of raising through the year and beating.

"I've been in this role now 9 years, Arvind's been in the role 6, 7 years. I don't think we've ever raised guidance in the first quarter... within our lens of what we're looking at, we're executing extremely well... our view is we should be prudent." — Jim Kavanaugh, CFO

Assessment: The reiterated guide is a sentiment negative but a fundamental non-event. Given IBM guided FCF to $13.5B in early 2025 and delivered $14.7B, the maintained ~$15.7B guide should again be read as a floor. The market's "sell the reiterate" reaction is exactly the kind of setup we like on a de-rated, cash-generative name.

3. The Mainframe as an AI Inferencing Platform

Management devoted significant airtime to a genuinely new monetization vector: running AI inferencing in line with transactions on IBM Z via the Spyre accelerator, enabling fraud scoring on 100% of transactions rather than a sample.

"What happens if you could run a 20 billion, 30 billion parameter model right on the mainframe... you can do that to every single transaction. So if you can take your fraud rate down from 50 basis points to 40, you can now do the math on what that is." — Arvind Krishna, CEO

Assessment: This reframes the mainframe from a mature, cyclical hardware line into a platform with an AI-driven capacity expansion. It is early, but it directly attacks the "Z is a melting ice cube" bear case and supports management's "strongest cycle ever" claim.

4. Red Hat Re-Acceleration

Red Hat grew 10% cc, up two points sequentially, on stabilization of consumption-based services, with OpenShift crossing $2B ARR. Management did flag that RHEL specifically decelerated on federal-signing softness and a dislocated hardware supply chain.

"Red Hat growth accelerated 2 points sequentially to 10%, largely driven by the stabilization of consumption-based services revenue growth that we expected. OpenShift is now $2 billion ARR business with strong growth." — Jim Kavanaugh, CFO

Assessment: This answers the biggest open question from our initiation. Red Hat back at 10% is the linchpin of the software-10%+ target; the RHEL/hardware-supply-chain caveat is a watch item but not yet a growth risk.

5. Confluent Closes Early

Confluent closed in mid-March, roughly two months ahead of the assumed timing, adding incremental near-term dilution but accelerating the data-revenue contribution. Management raised the full-year data growth outlook to the low-20s%, with Confluent contributing more than 15 points.

"We now see data up low 20-plus percent range. That's going to deliver 5 points of software growth... M&A contribution from Confluent... should be about a little bit north of 15 points of that 20% to 25% growth overall for the year." — Jim Kavanaugh, CFO

Assessment: The early close pulls dilution forward but also pulls synergy capture forward, and management held the +1-point margin guide by accelerating cost synergies. Execution here mirrors the HashiCorp playbook that earned management credibility.

6. Productivity and Margin Durability

The productivity engine delivered again: $4.5B of run-rate savings since 2023 with $1B more expected in 2026, driving the 140 bps of operating pre-tax margin expansion "ahead of expectations." IBM's use of its own AI tooling (Project Bob, now GA with ~45% developer productivity gains) is the "client zero" proof point.

"IBM Bob, our AI-based software development system, is now generally available. Our entire developer workforce is using Bob with average productivity gains of 45%." — Arvind Krishna, CEO

Assessment: The margin story remains the most reliable pillar and is now visibly AI-assisted, which is both a cost lever and a credibility proof for IBM's external AI-transformation pitch.

7. Capital Allocation and M&A Appetite

IBM raised the dividend for a 31st straight year to $1.69 and signaled a possible re-opening of the M&A pipeline in the second half, given attractive valuations across a de-rated software sector, once Confluent is fully integrated.

"The values that are out there right now are very attractive... So second half, if things stay where they are and if the values are where they are, maybe we can do something in the second half as we build up our cash balances." — Arvind Krishna, CEO

Assessment: A disciplined acquirer signaling appetite into a software de-rating is a potential value-creating optionality. The raised dividend, at a ~2.9% yield on the depressed price, also puts a firmer floor under the stock.

Guidance & Outlook

Management maintained the headline framework (>5% cc revenue, +$1B FCF, +1-point margin) while nudging the composition in a more favorable direction: software raised to 10%+, data to the low-20s% on Confluent's early close, and currency now a modest tailwind rather than a drag. The decision not to raise the total-revenue and FCF envelopes is a function of it being early in the year and IBM's stated preference to guide to numbers it can beat.

Implied ramp: With software raised and infrastructure held at down-LSD, the total-revenue guide has more cushion than in January; consulting inflecting to low-single-digits from +1% cc remains the swing factor. Q2 shape: revenue growth similar to the full-year rate, operating pre-tax margin +~50 bps (lighter than Q1's +140 bps as early Confluent dilution weighs). Street at: ~5.1% 2026 revenue growth, roughly in line with the maintained guide. Guidance style: consistent with IBM's now well-established pattern of conservative Q1 guidance followed by through-year raises.

Analyst Q&A Highlights

Where value accrues in the AI stack

The opening question probed IBM's software mix (infrastructure versus application, subscription versus consumption) and where incremental AI value accrues. Krishna gave the most complete articulation yet of why IBM's portfolio is an AI beneficiary rather than a victim.

Q: "How would you characterize your mix between infrastructure versus applications, consumption versus subscription... as AI adoption really scales, where in that stack do you see the most incremental value accruing to IBM versus the ecosystem?"
— Amit Daryanani, Evercore ISI

A: "If I count right, 4% of our portfolio, if I'm to be generous, could be called an application... Value is going to decrease in that interaction layer because as agents replace people... the agents are going to be interacting much more with the underlying data and the business logic. And we sort of saw that coming 6, 7 years ago, and that is why we picked the portfolio we did."
— Arvind Krishna, CEO

Assessment: The most thesis-relevant answer of the call. If the durable value migrates to the data and business-logic layers IBM owns, the AI-disruption fear is misdirected. This is the argument the de-rated stock is not pricing.

Why not raise guidance after a clear beat

A recurring line of questioning pushed on the maintained guide, asking whether management was seeing something soft in Europe or simply being cautious. Krishna addressed the macro directly; the CFO addressed the philosophy.

Q: "There's just some concern out there as to are you seeing something in Europe that keeps you at bay right now?... the free cash flow was better than expected in the quarter. Why not raise it?"
— Ben Reitzes, Melius Research

A: "We had the strongest growth we have seen in decades, not years, decades in the Middle East... In the first quarter, Europe was also strong... So just based on only 3 months of the year have gone by is why we're making the prudent comment."
— Arvind Krishna, CEO

Assessment: Management explicitly denied any demand softening and pointed to record Middle East and strong Europe results. The maintained guide is conservatism, not a warning, which makes the guidance-driven sell-off look like a sentiment overreaction.

Mainframe AI-inferencing monetization

Questioning turned to how much of the mainframe base is shifting toward AI-inferencing workloads and how that flows through to transaction-processing revenue. Management quantified the new capacity vector and the stack multiplier.

Q: "What type of a workload mix are you seeing today that you would consider conventionally mainframe? And what is that velocity of potential mix shift?... how should that transpire and be expressed in the business in terms of the growth cadence?"
— Fatima Boolani, Citigroup

A: "AI is adding a third kind of compute capacity into the mainframe... I believe we have a fully populated system we can do about 450 billion inferences a day on the mainframe... We just anniversaried our first full year of z17... We've increased hardware placement value by over $1 billion."
— Arvind Krishna, CEO / Jim Kavanaugh, CFO

Assessment: The AI-inferencing capacity is a structural addition to mainframe monetization, not a one-cycle bump. Combined with the 3-4x stack multiplier, it lengthens the tail of the current cycle and undercuts the cyclicality bear.

The software constant-currency downtick

An analyst pressed the optical deceleration in software from 11% cc in Q4 to 8% in Q1. The CFO said he "fully expected this one" and attributed the entire move to transactional seasonality.

Q: "Just on the constant currency for software... last year, 9% growth in Q1, I think it was 11% in Q4. So investors are asking, you're seeing a little bit of a downtick. Is that due to seasonality...?"
— Brent Thill, Jefferies

A: "The entire 3-point drop quarter-to-quarter is the fundamentals of the mix of the portfolio... When you look at the underpinnings of the core annuity by itself, we're actually accelerating fourth quarter to first quarter... coming out of first quarter, I feel pretty good about 8% growth."
— Jim Kavanaugh, CFO

Assessment: A credible, quantified explanation. The recurring base accelerating while transactional mix falls off seasonally is exactly what you would want to see; the headline downtick is not a demand signal.

Memory and supply-chain headwinds

An analyst asked how IBM is managing higher memory costs and supply constraints and how material memory is to the infrastructure base. Management called the impact "de minimis" for the mainframe while flagging storage and RHEL as the watch areas.

Q: "Can you maybe just detail how IBM is broadly managing and/or mitigating some of these supply chain headwinds, whether that's higher memory costs or supply challenges... how does it impact your outlook on growth and margins?"
— Erik Woodring, Morgan Stanley

A: "Underneath it around the supply chain dislocation... in particular around memory, it has a de minimis impact to us overall... The area we're watching it is in the software area around RHEL... But look at our hardware performance, we accelerated growth at 15%."
— Jim Kavanaugh, CFO

Assessment: Reassuring on the high-margin mainframe, honest on the storage/RHEL exposure. Memory inflation is an industry watch item, but IBM's mix (75%+ human-capital/IP, hardware ~25%) limits the direct hit.

Consulting outlook and the AI-bookings trajectory

Questioning returned to whether AI bookings accelerated and whether macro would dent the consulting growth outlook. The CFO reframed the retired GenAI-book metric into revenue-based disclosure and defended the low-single-digit acceleration.

Q: "I was wondering if you could maybe comment on the AI bookings... Did that accelerate or decelerate in the quarter? And then... any update you see for the consulting business this year, either given more macro uncertainty?"
— James Schneider, Goldman Sachs

A: "Consulting is about 40% of our signings, 30% of our backlog is GenAI now, over 20% of our revenue... in the first quarter, we eclipsed $4 billion ARR... signings, we return to growth... 400 new clients captured in the first quarter."
— Jim Kavanaugh, CFO

Assessment: The GenAI-book retirement we flagged in January was replaced with revenue/ARR disclosure, which is arguably more useful. The +6% signings and $4B GenAI ARR support the low-single-digit consulting inflection, though revenue still has to follow.

Winning regardless of the GenAI application-layer winner

The closing question asked how IBM is positioned to win whichever foundation-model or application vendor prevails. Krishna reiterated the neutral, "Switzerland" model strategy.

Q: "You showed how low of an exposure you have to the application space... how are you setting IBM up to win kind of regardless of what ends up being the winner of the GenAI application layer? And what kind of investments does that take?"
— Matthew Swanson, RBC Capital Markets

A: "We made the decision about 3 years ago that we were going to be neutral and Switzerland like also on our usage of frontier models... We actually do not want to even predict which of them will be the eventual winners. We want to work with all of them... we think that's a good half of the world is interested in that paradigm."
— Arvind Krishna, CEO

Assessment: The model-neutral positioning means IBM does not need to pick the AI winner to benefit from AI adoption. It is a coherent hedge that the current de-rating gives investors for free.

What They're NOT Saying

1. No hard consulting demand acceleration: Management repeatedly cited signings (+6%), backlog quality, and GenAI mix, but consulting revenue still grew only 1% cc. The answer was again about realization and mix, not a demand inflection an investor can see in the revenue line yet.
2. The Anthropic/COBOL threat, unaddressed by name: The single biggest driver of the YTD de-rating was the fear that AI coding tools disrupt IBM's modernization work. Management made the structural "we sit below the app layer" case but never directly engaged the specific competitive threat that spooked the market.
3. RHEL deceleration understated: RHEL specifically decelerated on federal softness and hardware supply-chain dislocation; this surfaced only in Q&A and was quickly bracketed by the stronger OpenShift and consumption stabilization story.
4. No raise despite a clear beat: Reiterating revenue and FCF after beating both, with currency now a tailwind, leaves conspicuous room to the upside that management declined to formalize. Prudent, but it withheld the catalyst the beat arguably warranted.
5. Confluent dilution phasing still vague: The early close pulled dilution forward and pressured Q2 margin (+~50 bps versus Q1's +140 bps), but the precise quarterly EPS drag was left qualitative.

Market Reaction

• Pre-print setup: IBM closed at $251.86 on April 22, already down ~15% YTD (versus the S&P 500 +4%) after a punishing first quarter that included a ~13% single-session drop in February on the Anthropic/Claude-Code COBOL-modernization fear. Over the trailing twelve months the stock was up just ~5% versus the S&P's ~35%, a stark reversal from the market-doubling twelve-month return it carried into the Q4 print.
• Reaction (April 23, T+1 session): Shares gapped down to open at $232.27 (−7.8%), touched an intraday low of $221.73 (−11.96%), and closed at $231.08, down 8.25%. Volume of 22.3M shares ran ~4.4x the 30-day average, the heaviest of the period. The S&P 500 was roughly flat (−0.4%), so the move was almost entirely IBM-specific.

The reaction is a textbook "sell the news into fear." IBM beat on revenue, EPS, software, and infrastructure, set a Q1 cash record, and raised its software target, yet the market fixated on the reiterated (not raised) guide and the unresolved consulting-disruption narrative. When a company beats and the stock falls 8% on top of a 15% year-to-date decline, the burden shifts to whether the fear is proportionate to the fundamentals. On the evidence of this print, it is not: the software franchise is enabling/middleware that benefits from AI scaling, the mainframe cycle is running ahead of its record predecessor, and cash generation is at a decade high. That gap between price action and fundamentals is the basis for our upgrade.

Street Perspective

Debate: Is the consulting-disruption fear a company-level or segment-level risk?

Bull view: The bull case holds that the AI-coding threat hits only IBM's modernization consulting, roughly a slice of a ~33%-of-revenue segment, while the software franchise (Red Hat, data, automation, mainframe software) is enabling/middleware that consumes more as AI scales.

Bear view: The bear camp argues consulting is a structural share donor to AI tools, that its +1% cc growth already shows the pressure, and that the overhang will cap the multiple regardless of software strength until consulting demonstrably inflects.

Our take: The bull has the stronger argument. Krishna's ~4%-application framing is credible and consistent with the segment data; the market is pricing a whole-company disruption on a segment-specific risk. We would monitor consulting signings-to-revenue conversion but do not see it as thesis-breaking.

Debate: Should investors trust the "strongest cycle ever" mainframe claim?

Bull view: A growing view is that z17 is structurally different: 100%+ new-MIPS growth for four quarters, >$1B more placement value than the record z16 cycle, and a new AI-inferencing capacity layer that adds monetization the guide does not fully capture.

Bear view: Skeptics note management still guides infrastructure down for the full year, that mainframe is inherently lumpy, and that a strong Q1 is the easy part of any cycle.

Our take: We lean bull here, a shift from our January caution. The proof points are hard to dismiss, and the AI-inferencing vector is genuinely new. We treat the "down LSD" full-year guide as conservative and any beat as upside.

Debate: Has the de-rating created value or is it a value trap?

Bull view: At ~18x forward operating EPS, a ~7% free-cash-flow yield, and a ~2.9% dividend raised for a 31st year, IBM is priced for stagnation despite accelerating cash flow and a raised software target.

Bear view: The bear contends a slow-growth, consulting-exposed IBM deserves a low multiple, and that the stock can stay cheap while the AI-disruption debate lingers.

Our take: This is where we land the upgrade. The valuation risk that made us cautious in January has resolved; the price now embeds the bear case, and the fundamentals argue against it. We would rather own the disconnect than wait for the debate to resolve on the tape.

Model Update Needed

Valuation impact: We trim our fair-value range to ~$270–285 (from ~$290–300 in January) to reflect the genuine consulting/AI-disruption overhang and a more cautious market multiple. Even after that haircut, the range implies roughly 17–23% price upside from the $231.08 reaction close, plus a ~2.9% dividend. At ~18x forward operating EPS and a ~7% FCF yield, the risk/reward is now clearly favorable, which flips the rating to Outperform. The downside case (~$205) is a consulting stall plus a prolonged AI-disruption discount; the upside case (~$300+) is consulting inflecting while the mainframe cycle beats the conservative guide.

Thesis Scorecard Post-Earnings

Overall: Thesis strengthened versus January. The three bull pillars all advanced (software durability confirmed via the Red Hat re-acceleration, record cash flow, early Confluent close), the mainframe-cyclicality bear de-escalated on hard proof points, and the valuation bear that anchored our Hold resolved sharply in our favor. The lone escalation is the consulting/AI-disruption risk, which the market has over-extrapolated to the whole company.

Action: Upgrade to Outperform from Hold. The de-rating has overshot a business that is executing and generating record cash; we would buy the disconnect. The key risk to monitor is consulting revenue conversion, and the key catalyst is the mid-year guide raise IBM's own pattern implies.