The Thesis Delivered in Full; With the Inflection Now Priced and 2026 Set to Digest the Z Cycle, We Downgrade IBM to Hold

Results vs. Consensus

Full-Year 2025 vs. the Investor Day Model

Year-over-Year Comparison (Q4)

Assessment

Revenue. Nine percent constant-currency growth is IBM's fastest in over three years and caps a year in which every segment accelerated. The composition is now unambiguously healthy: Software double-digit with organic growth above 7%, Infrastructure driven by a record mainframe year, and even Consulting positive. The full-year 6% cc result beat the entering-year guide and, more importantly, was delivered with the right mix. This is the quarter that retroactively validates the entire re-rating.

Margins and cash. Full-year operating gross margin of 59.5% and operating pre-tax margin expansion of a point (140 bps ex-charges) mark the highest gross margin in IBM's reported history and the highest pre-tax margin in a decade. Free cash flow of $14.7B (up 16%) is the headline: the highest cash generation in over a decade and a record margin, and it beat the $13.5B the company set a year ago by more than a billion dollars. The free-cash-flow engine is doing exactly what the bull case requires, and it is the single strongest reason this remains a Hold rather than a sell.

The forward tension. The problem for the equity is not the quarter, it is what the quarter implies about 2026. Infrastructure, which contributed meaningfully to 2025's acceleration, is now guided down low-single-digits as the z17 cycle digests, and Red Hat, the highest-quality growth asset, is decelerating with a fresh federal headwind. The company can still grow 5%+ cc on the strength of the software portfolio, but the rate is flat-to-slightly-lower than 2025, not accelerating, and that matters when the stock has already re-rated to price the good news.

Segment Performance

Key portfolio KPIs

Software (+11% cc): the mainframe-software inflection lands

Software accelerated to 11% cc, its highest in fifteen years on top of an equally strong prior-year comp, with organic growth above 7% and three of four sub-segments double-digit. Data surged 19% cc on GenAI demand and strategic-partner momentum; Automation grew 14% cc on a record HashiCorp bookings quarter; and, most importantly, Transaction Processing inflected to +4% cc as the record z17 placement began converting to software. Red Hat was the laggard at 8% cc, decelerating on the previously-flagged consumption-services wrap plus a new drag: single-digit ACV bookings caused by government-shutdown delays in federal deal activity.

"As we expected last quarter, given the record z17 placement this year, transaction processing inflected back to growth of 4%. ... The flat this year in 2025 was off a plus 10. So you could see the compounding effect of what's happening to TP." — James Kavanaugh, CFO

Assessment: The TP inflection is the payoff on the platform-multiplier thesis and it landed on schedule, arguably a quarter early relative to the z16 analog. That is a genuine win. The offsetting concern is Red Hat: an 8% cc print with management now saying it only needs Red Hat at "double digit" to clear the software guide is a quiet retreat from the mid-teens ambition it carried a quarter ago.

Infrastructure (+17% cc): a record year, and a 2026 headwind

Infrastructure grew 17% cc as IBM Z posted its highest fourth-quarter revenue in more than two decades (up 61% cc); for the full year Z rose 48%, the highest annual mainframe revenue in about twenty years, outpacing the record z16 program over its first three quarters. Management then guided Infrastructure down low-single-digits in 2026 (about a half-point drag to IBM) as the cycle enters its digestion phase, with Q1 Z strength giving way to product-cycle dynamics through the rest of the year.

"We enter 2026 three quarters into the z17. We expect infrastructure revenue to be down low single digits ... with Z growth in the first quarter balanced by product cycle dynamics throughout the rest of the year." — James Kavanaugh, CFO

Assessment: A spectacular mainframe year, but the guide-down is the crux of the downgrade. Infrastructure was a real contributor to 2025's acceleration, and its reversal to a drag in 2026 is why total revenue growth flattens rather than builds. The offsetting benefit is the software tail: the installed MIPS keep monetizing through Transaction Processing over the next one to two years.

Consulting (+1% cc): stabilized, still sub-scale as a grower

Consulting grew 1% cc, decelerating slightly from Q3, with the GenAI consulting book surpassing $2B in the quarter (over a third of bookings, over 25% of backlog, a $3.6B GenAI ARR run-rate) and margins up nearly 200 bps for the year to a three-year high. Signings were down on a record year-ago comp but with improving mix. Management guided 2026 Consulting to low-to-mid single-digit growth.

"We returned the business back to durable sustainable growth in the second half ... we guided low single to mid-single which we think is gonna be pretty much where the market is at overall." — James Kavanaugh, CFO

Assessment: Consulting has done what we modeled, stabilized and stopped being a drag, but the 2026 guide of "roughly market growth" confirms it is a low-single-digit business, not an engine. The GenAI ARR run-rate is the encouraging signal, but at 1% cc this quarter it is not moving the consolidated needle.

Key Topics & Management Commentary

Overall Management Tone: Management was confident and valedictory about 2025, repeatedly framing the year as validation of a multi-year repositioning from a one-growth-engine company (Red Hat) to one with three double-digit software growth factors. The forward posture was deliberately measured: on Infrastructure and Red Hat, management leaned conservative, emphasizing it guides only to numbers it has high confidence of hitting or beating. The one place the tone turned promotional was the software portfolio's durability, where the CFO pointedly upgraded the framing from "approaching double digits" to "double digits, not approaching."

The 2026 software double-digit guide

The headline forward claim is Software growth of about 10% in 2026, which management built as roughly seven points organic plus three points from acquisitions, underpinned by a ~$24B ARR base (high-single-digit growth), Data in the high teens, Automation low-double-digit, Transaction Processing low-to-mid-single-digit, and Red Hat at double digit.

"We feel confident about software now at double digits, not approaching double digits. ... organic will be north of seven points this year and acquisitions about three points." — James Kavanaugh, CFO

Assessment: The software guide is credible and the organic majority is reassuring, but it leans on Data staying in the high teens and Red Hat holding double-digit despite an 8% cc exit. The margin for error is thinner than the confident tone implies, and it is the segment doing the heavy lifting to offset the Infrastructure guide-down.

The Confluent acquisition

IBM announced its intent to acquire Confluent, a real-time data-streaming platform, framed as unifying hybrid-cloud and automation through a "smart data platform" that feeds AI agents real-time governed data. Management expects the deal to close by midyear 2026, absorb about $600M of dilution this year (mostly stock-comp and interest), reach adjusted-EBITDA accretion within the first full year and free-cash-flow accretion in year two, with ~$500M of operational-spend synergies by 2027.

"Confluent has the most capable technology to unlock the real-time value of data across applications, clouds, APIs, and as AI agents enter the enterprise, they will need access to that data in real-time." — Arvind Krishna, CEO

Assessment: Strategically coherent and consistent with the data-and-automation build-out (Apptio, HashiCorp, DataStax). The HashiCorp precedent (EBITDA-accretive ahead of plan) earns management some benefit of the doubt on integration. The near-term cost is $600M of 2026 dilution and further balance-sheet use, which the raised $5.5B productivity target is explicitly designed to offset.

Free cash flow: a record year and a decelerating guide

Free cash flow of $14.7B (up 16%) was the year's marquee number, driven by adjusted EBITDA up $2.8B. Management guided 2026 to ~$15.7B, high-single-digit growth, framed against the same headwinds as 2025 (cash taxes, higher capex, net interest) plus Confluent dilution.

"Over the last three years, we've grown free cash flow five and a half billion dollars. ... we said we'd guide confidently up a billion dollars at $15.7 billion. ... our job is to beat it this year." — James Kavanaugh, CFO

Assessment: The cash engine is the bedrock and the multi-year track record (up $5.5B in three years, beating each year's initial guide) is genuinely impressive. The deceleration to high-single-digit growth is the honest read on a year with an Infrastructure drag and Confluent dilution, but the "our job is to beat it" framing, consistent with IBM's habit of clearing conservative guides, tempers the concern.

The productivity flywheel scales again

Management raised the productivity target to $5.5B of annual run-rate savings by 2026, an incremental $1B, explicitly to mitigate Confluent dilution and mainframe product-cycle mix. Project Bob now has more than 20,000 internal users (up from 8,000 a quarter earlier) at 45% average productivity gains.

"We took up our productivity target to $5.5 billion. We will get revenue scale leverage on G&A, which will mitigate dilution and mitigate product cycles overall." — James Kavanaugh, CFO

Assessment: The productivity program keeps outrunning its targets ($2B to $4.5B to $5.5B) and remains the most durable, macro-independent lever in the model, the reason IBM can guide a point of margin expansion into a year with Infrastructure declining and a dilutive deal. This pillar is fully intact.

Memory pricing and the server-refresh dynamic

Asked about DRAM spot prices running roughly 6x year-ago levels (as capacity shifts to HBM for AI servers), the CEO argued the dynamic is net-neutral-to-positive for IBM: AI-server demand pulls through conventional CPUs and servers, RHEL keeps taking share, and Red Hat AI and OpenShift AI benefit from the AI-server buildout, offsetting any pressure on the standard server refresh.

"There is no AI server without a bunch of CPUs right next to it. ... I don't expect that the overall server dynamic ... is actually gonna be any headwind to us on the hybrid cloud or the Linux side." — Arvind Krishna, CEO

Assessment: A reasonable framing, and IBM's software exposure to memory pricing is indirect. Still, elevated memory costs plus the federal-bookings drag are two live pressures on Red Hat that make its double-digit 2026 guide the portfolio's key show-me item.

Where enterprise AI runs: the "and" world

On the strategic question of where enterprise GenAI workloads will reside, the CEO laid out a three-to-five-year view: roughly 50% of enterprise AI usage in private cloud or on-premise and 50% public, with the on-premise share of actual inferencing potentially 80-90% if smaller efficient models dominate. This underpins IBM's "sovereign," hybrid positioning across the mainframe, Red Hat AI, and watsonx.

"If I look out three to five years, 50% of the enterprise usage of AI [is] going to be in either a private cloud or ... their own data centers. ... it's an and." — Arvind Krishna, CEO

Assessment: The hybrid/sovereign thesis is the strategic core of the IBM AI story and it is internally consistent. Whether the enterprise on-premise share lands at 50% is unknowable, but IBM does not need to win the public-cloud AI race; it needs regulated enterprises to keep a material share of AI on-premise, which its install base makes a reasonable bet.

Guidance & Outlook

The 2026 guide is a study in offsetting forces. Software accelerates to ~10% and Consulting inflects to low-to-mid-single-digit, but Infrastructure reverses to a low-single-digit decline as the record mainframe cycle digests, netting to total revenue growth of more than 5% cc, roughly flat versus 2025's 6% rather than the acceleration the 2025 trajectory might have implied. Below the line, the model is intact: a point of margin expansion (funded by the raised $5.5B productivity target absorbing ~$600M of Confluent dilution) and ~$15.7B of free cash flow.

Q1 2026 shape: Management guided Q1 cc revenue growth similar to the full year (~5%+), operating pre-tax margin expansion of about 100 bps, a mid-teens tax rate, and, by their own math, double-digit profit and EPS growth, building on Q4 momentum.

Street at: The >5% cc revenue guide sits modestly above the ~4.6% sales-growth consensus LSEG carried into the print, and the ~$15.7B free-cash-flow guide is a clean high-single-digit step up. Management characteristically framed both as numbers it intends to beat.

Guidance style: Conservative and consistent. The pattern of setting achievable guides and clearing them is intact (2025 beat its initial FCF guide by $1.2B), so the flat-ish revenue guide likely embeds cushion. But the honest read is that 2026 is a digestion-and-software year, not an acceleration year, and that is the crux of why the risk/reward has narrowed at this price.

Analyst Q&A Highlights

The components of the 2026 software acceleration

The opening question asked management to decompose the confidence behind double-digit software growth. The CEO walked the sub-segments (Automation on secular demand, Data on GenAI and partnerships, mainframe software following hardware, Red Hat led by OpenShift), and the CFO reframed the whole portfolio as three double-digit growth factors where three years ago there was only one.

Q: "Good to see the comment around software growth accelerating to double-digit growth this year. ... could you maybe dig in the components and why you're excited for that organic-led initiative?"
— Brent Thill, Jefferies

A: "We finished '25 with one of the highest growths we've ever had in software ... but it's pervasive with three of our four software categories growing double digits. You dial back only about three years ago, we only had one growth factor. And that was Red Hat." — James Kavanaugh, CFO

Assessment: The diversification argument is the strongest part of the bull case and it is genuinely true, the software portfolio is far more robust than three years ago. The residual question is whether ~10% software can carry the whole company while Infrastructure declines, which makes the software guide the load-bearing assumption for 2026.

Free-cash-flow deceleration to high-single-digit growth

A pointed question probed why the ~$15.7B free-cash-flow guide implies high-single-digit growth versus the 16% posted in 2025. The CFO attributed 2025's strength to the fundamentals (EBITDA up $2.8B, revenue acceleration, operating leverage) and framed the 2026 guide as an Investor-Day-consistent high-single-digit step-up he intends to beat.

Q: "The $15.7 billion free cash flow guide ... implies high single-digit free cash flow growth versus the 16% that you saw last year. ... Was it something unique that you saw in '25?"
— Amit Daryanani, Evercore ISI

A: "The underpinnings behind that cash flow were entirely driven by the fundamentals of our business ... adjusted EBITDA, at double-digit growth, we finished at 17%. ... we feel confident about that fifteen seven, and our job is to beat it this year." — James Kavanaugh, CFO

Assessment: Management did not claim a one-off drove 2025, which is reassuring, and the multi-year track record of beating the initial guide is real. Still, the deceleration to high-single-digit growth, with Confluent dilution and an Infrastructure headwind, is the honest 2026 base case, and the "beat it" framing is the upside option, not the plan.

Bridging Red Hat from 8% back toward the model

The most probing exchange asked directly how Red Hat gets from 8% cc back toward mid-teens, or whether mid-teens is no longer the target. The CFO's answer was telling: he defended double-digit software without recommitting to mid-teens Red Hat, attributed the Q4 softness to a federal-bookings disruption, and noted IBM needs Red Hat only at double-digit to clear the software guide.

Q: "I wanted to see how we're going to bridge Red Hat ... how do we get the eight to the mid-teens or is mid-teens no longer the growth rate there?"
— Ben Reitzes, Melius Research

A: "We only delivered single-digit ACV bookings in the fourth quarter because ... we were disrupted by the down of the federal government. ... We only need [Red Hat] at double digit to get software over the line of double digit. Upside will deliver upside to software overall." — James Kavanaugh, CFO

Assessment: This is the single most important exchange for the rating. Management effectively lowered the Red Hat bar from mid-teens to double-digit and leaned on a federal-shutdown excuse for the bookings miss. The crown-jewel asset decelerating, with the ambition quietly reset, is exactly the kind of deterioration that argues for stepping back to Hold at a full multiple.

Productivity cadence and Confluent dilution absorption

A question on the cadence of margin improvement, given productivity actions already taken in Q4 to pre-fund Confluent dilution, drew a detailed bridge: of the ~1 point of 2026 margin expansion, about a half-point from revenue scale, a half-point lost to mainframe mix, and a full point from productivity.

Q: "How should we think about the cadence of the progress of that [productivity], given that Confluent is to hit midyear, but you've already taken some productivity actions here in 4Q?"
— Wamsi Mohan, Bank of America

A: "We set about a point growth in margin. Gonna get about a half a point out of revenue scale ... We will lose a half a point on portfolio mix because we are gonna wrap on an unprecedented launch momentum on mainframe. ... The remaining full point basically is gonna be driven out of productivity." — James Kavanaugh, CFO

Assessment: The margin bridge is transparent and credible, and it confirms productivity is doing the heavy lifting to offset the mainframe mix reversal. It also underscores that 2026 margin expansion is manufactured, not volume-driven, which is durable but not the mark of an accelerating business.

Consulting conversion path through 2026

On Consulting's trajectory, the CFO laid out a low-single-digit Q1 building through the year to low-to-mid-single-digit, supported by a $32B backlog, a $3.6B GenAI ARR run-rate, 400+ new clients, and improving net-new mix, while cautioning on an aggressive pricing environment.

Q: "Could you maybe outline the path or trajectory you're expecting for the consulting business throughout the year? Signings were a little bit weaker ... but you noted the strong backlog in AI."
— Jim Schneider, Goldman Sachs

A: "GenAI now represents over a third of our bookings, over 25% of our backlog ... over 15% of our revenue on an exit run rate. We have a $3.6 billion ARR GenAI revenue run rate in consulting ... low single-digit first quarter, and then growing throughout the year." — James Kavanaugh, CFO

Assessment: The GenAI-led backlog gives the Consulting acceleration more credibility than a year ago, and the segment is clearly stabilized. But low-to-mid-single-digit growth in an aggressive pricing environment keeps Consulting a modest contributor, not a reason to own the stock.

Infrastructure conservatism versus Z sustainability

The final substantive exchange challenged the Infrastructure guide-down against the strength of the z17 cycle and the mainframe's emerging role as an AI-inference platform, asking whether guidance was simply conservative. The CEO reiterated IBM guides only to numbers it can hit or beat, then made the bull case for structural mainframe demand (sovereignty, AI-assisted modernization via watsonx Code Assistant for Z, and on-box inferencing via the Spyre cards).

Q: "You're guiding to a bit of infrastructure decline in 2026. ... could you just be a bit conservative as you look out ... could you see some more sustainability in the Z cycle that maybe isn't fully accounted for?"
— Eric Woodring, Morgan Stanley

A: "We want to give guidance ... where we have incredibly high confidence that we can hit or beat those numbers. ... a very large number of our clients actually told us they're interested and they kept space in the machines to put in those [Spyre] cards. ... As those cards go in, then the software stack to support that goes in." — Arvind Krishna, CEO

Assessment: The Spyre-card upside is a real optionality that could make the Infrastructure guide-down conservative, and it feeds the Transaction Processing software tail. But management is guiding to a decline, and prudent analysis takes the guide at face value: Infrastructure is a 2026 headwind, with Spyre-driven upside as a call option rather than a base case.

What They're NOT Saying

1. A recommitment to mid-teens Red Hat: Asked point-blank how Red Hat bridges from 8% to mid-teens, management defended only "double digit" and leaned on a federal-shutdown excuse. The mid-teens ambition that anchored the Red Hat narrative appears to have been quietly retired.
2. The standalone AI revenue run-rate: Q4 was the last quarter IBM will disclose the GenAI book of business separately, retiring the metric just as investors had learned to track it. The stated rationale (AI is now embedded everywhere) is plausible, but it removes the cleanest external gauge of AI monetization right at the point of peak scrutiny.
3. The all-in Confluent price and financing: Management detailed 2026 dilution (~$600M) and synergy targets but was light on the total consideration and how the balance sheet absorbs it alongside the existing $61B debt load and $67.7B of goodwill.
4. A quantified 2026 Infrastructure trajectory beyond "down low-single-digits": The Q1-strong, then-declining cadence was described qualitatively; the magnitude of the second-half mainframe drag, which shapes the whole-year revenue path, was left vague.
5. The organic-only 2026 revenue growth rate: With Confluent, HashiCorp, and DataStax in the mix and roughly three points of the software guide inorganic, management did not isolate the organic consolidated growth rate, the number that would show whether the underlying business is still accelerating or merely holding.

Market Reaction

• Pre-print setup: IBM entered at $294.16, having pulled back modestly (-0.7% year-to-date 2026, -3.8% over the trailing 30 days) from its late-2025 highs after a strong year (+30% over twelve months). The 52-week closing high was $314.98, so the stock was consolidating just below its record.
• Reaction session (January 29): Shares jumped 5.1% to $309.24 ($15.08), gapping up 8.1% at the open ($317.86) before settling back toward the mid-$300s. Volume of 10.1M was 2.5x the 30-day average. This was the first clearly positive print reaction in our coverage.
• Context: The S&P 500 was roughly flat (-0.1%) on the session, so the rally was entirely IBM-specific.

The market rewarded exactly what the bulls were waiting for: the Transaction Processing inflection, double-digit software, a record free-cash-flow year, and a 2026 free-cash-flow guide that stepped up cleanly. That the stock could not hold its 8% opening gap and settled at +5.1% is a mild tell that the enthusiasm has limits at this valuation. Our read is that this print completes the thesis rather than extends it: the inflection we upgraded to own in October is now delivered, disclosed, and largely reflected in a stock trading in its all-time-high zone. From here, the incremental buyer is paying a full multiple for a 2026 that decelerates as the mainframe cycle digests, which is a different risk/reward than the one that made this an Outperform.

Street Perspective

Debate: Does 2026 accelerate or digest?

Bull view: The bull case argues software reaccelerates to ~10%, Consulting inflects higher, the Transaction Processing tail compounds, Confluent adds a growth lever, and management's conservative guides (which it beat all through 2025) leave upside; the Infrastructure dip is a known, temporary cycle trough.

Bear view: The bears counter that total revenue growth is guided flat-to-lower (5%+ cc vs. 6%), the highest-quality asset (Red Hat) is decelerating with the mid-teens goal abandoned, Infrastructure swings from tailwind to headwind, and Confluent adds dilution, so 2026 is a digestion year dressed up as a software-acceleration story.

Our take: We now lean bear on the near-term setup, having been bullish a quarter ago. The software acceleration is real but it is offsetting an Infrastructure reversal rather than compounding on top of it, so the consolidated growth rate flattens. That is not a broken thesis, it is a maturing one, and at a full multiple a maturing thesis is a Hold, not an Outperform.

Debate: Is the valuation still reasonable after the run?

Bull view: At roughly 25x forward operating earnings with ~$15.7B of free cash flow (~5% yield), a rising dividend, a fortress recurring-revenue base, and a management team that keeps beating its guides, IBM is fairly valued for a durable software-led compounder and the multiple can hold as the software mix keeps shifting.

Bear view: The stock has re-rated from the low-$260s to the low-$300s in six months and now sits at an all-time-high zone near 25x forward for a business guiding roughly-flat revenue growth; the re-rating is complete and further upside requires multiple expansion that is hard to justify on decelerating growth.

Our take: The valuation is the deciding factor. We do not think IBM is expensive enough to short (the cash flow and dividend anchor it), but at an all-time-high multiple into a digestion year, the asymmetry that made it an Outperform in October is gone. Taking the gain and reverting to Hold is the disciplined response to a thesis that has been paid out.

Debate: Is the acquisition cadence value-additive or empire-building?

Bull view: IBM's M&A machine (Apptio, Software AG, HashiCorp, DataStax, now Confluent) systematically buys category leaders in structurally growing markets and accelerates them through IBM's distribution, with HashiCorp already EBITDA-accretive ahead of plan, an underappreciated organic-growth amplifier.

Bear view: A steady stream of dilutive deals (Confluent alone is ~$600M of 2026 dilution) alongside a $61B debt load and $67.7B of goodwill raises the question of how much of the "growth" is bought, and each deal defers the moment investors can judge clean organic performance.

Our take: The track record earns management the benefit of the doubt on execution, and the synergy discipline (three hard criteria, HashiCorp's early accretion) is real. But the reliance on M&A to sustain the software growth rate is a genuine watch item, and the retirement of the standalone AI-book disclosure makes the organic-versus-inorganic picture harder to audit precisely when it matters most.

Model Update Needed

Valuation impact: At the post-print $309.24, IBM trades at roughly 25x our ~$12.6 FY26 operating EPS and a ~5% free-cash-flow yield, near its 52-week high. A fair-value framework at ~23-25x forward operating earnings supports a range around $300-320, which brackets the current price. That symmetry, cash-flow floor under the downside, full multiple capping the upside, is the arithmetic of the downgrade: we no longer see the asymmetric upside that justified an Outperform in October.

Thesis Scorecard Post-Earnings

Overall: Thesis delivered, and largely paid out. Every operational milestone we set landed (TP inflection, double-digit software, record free cash flow), which is precisely why the risk/reward has narrowed: the good news is now realized and in the price. The two emerging concerns, an Infrastructure guide-down and a decelerating Red Hat with its ambition reset, tip the forward setup from asymmetric to balanced.

Action: Downgrade to Hold from Outperform, and take the gain from the October upgrade. We would return to Outperform on a meaningful pullback that restores a margin of safety, on Red Hat reaccelerating back toward double-digit-plus, or on Spyre-driven mainframe upside that makes the Infrastructure guide-down conservative. We would move to Underperform only on evidence the software acceleration is stalling or the free-cash-flow guide is at risk, neither of which is present today.