IBM Long-Term Owner's Perspective Research

Hunting ten-year five-baggers among great growth stocks — pressing the upside question: "Can it get much bigger?"

• QHow large is its market ceiling? Is it expanding an existing pie, or creating an entirely new market?5/10

  A

  Bottom line: IBM is mainly expanding and defending a large existing pie, not creating a wholly new market. Its ceiling is set by the mature, enormous, but moderately growing pool of enterprise IT spending. The upside comes from gaining deeper penetration and higher pricing in existing lanes such as generative AI, hybrid cloud, and mainframe modernization, rather than opening up untouched territory.

  Start with the pie itself. IBM's 2025 total revenue was $67.5 billion, spread across three businesses with very different economics: software at about $30 billion, consulting at about $21.1 billion, and infrastructure at about $15.7 billion. All three map to markets that have existed for decades: enterprise software and middleware, IT consulting delivery, and mission-critical infrastructure. The report gives industry attractiveness only 3/5 precisely because "long-term demand exists, but competition is exceptionally intense, customers have strong bargaining power, and technology substitution is fast." In other words, the pie is large enough, but it is not a new pie that has yet to be divided.

  How high the ceiling is depends on how much deeper IBM can penetrate the existing pool. Here IBM does have real room to expand the existing pie: by 2026 Q1, software ARR had reached $24.6 billion, up about 10% year over year, and remaining performance obligations were about $69 billion, with about 69% to convert into revenue within two years. These figures show that it can still keep deepening share within existing customers, rather than having hit a revenue ceiling. But this penetration means taking more share and selling more modules in an existing market. In substance, it is thickening the existing pie, not enlarging the pie into a new category.

  The only line with a "new market" flavor is generative AI. IBM has disclosed that its generative AI book of business has cumulatively exceeded $12.5 billion. But the honest version is this: IBM did not create this new market from nothing. It is capturing orders in a wave defined and ignited by others such as OpenAI, Microsoft, and NVIDIA, using its enterprise customer relationships and consulting delivery capability. It is one participant sharing in the new pie, and it is not the definer. The report's bear case puts it bluntly: IBM may only be an established hybrid technology vendor growing in the low to mid-single digits, repackaged by the AI narrative.

  Measured by Baillie Gifford's "fivefold in ten years" yardstick, a stock that can truly support a fivefold outcome usually either sits in a total pool that is structurally expanding, such as early EVs or cloud computing penetration rising from single digits, or is itself the definer and largest beneficiary of a new market. IBM does not have much of either. Its total pool grows moderately; its own revenue CAGR was only about 4.2% from 2021 to 2025. In the new AI wave, it is a participant rather than a definer. So on this dimension, IBM is expanding a large existing pie and taking a slice of a new pie. Its ceiling is jointly determined by moderate growth in total enterprise IT spending and improvement in its own penetration rate. It is a high and thick ceiling, but not a steep one, and it lacks the "entirely new market" attribute that can drive a fivefold imagination.

  Jun 10, 2026
• QCan its revenue at least double over the next five years? Will growth be driven mainly by volume, price, or new businesses?3/10

  A

  Bottom line: almost certainly not. For IBM to double revenue over the next five years, it would have to grow from $67.5 billion in 2025 to about $135 billion, implying roughly 15% compound growth for five consecutive years. That is far above the company's own growth centerline and clearly above the top end of its guidance. The growth mix leans more toward software price increases and renewals, mainframe cycle volume, and acquisition pieces, rather than an explosive ramp from one curve.

  Anchor on the facts first. IBM's revenue rose from $57.35 billion in 2021 to $67.535 billion in 2025, a CAGR of about 4.2%. 2025 was its best recent year, yet year-over-year growth was only about 8% (6% at constant currency). By 2026 Q1, the company's full-year guidance remained "more than 5% revenue growth at constant currency". For a company that treats "more than 5%" as a good-year target, doubling in five years, which requires 15% per year, does not work mathematically unless there is an enormous acquisition or a reconstruction of the financial perimeter.

  Break down the sources of growth to see its quality:

  Volume. This mainly comes from the mainframe cycle and expansion in software seats. In 2026 Q1, IBM Z grew about 51% year over year, and infrastructure overall grew 15%. That is real volume, but mainframes are a strongly cyclical business. A new z-series generation pulls demand forward, then the cycle gives some of it back as it rolls over. It cannot be linearly extrapolated for five years. Software seats rising as enterprises adopt cloud, automation, and AI are the more sustainable volume source.

  Price. Software's high renewal and pricing power is IBM's revenue stabilizer. Software ARR was $24.6 billion, up about 10% year over year, supported by sticky pricing in critical middleware, Red Hat subscriptions, and mainframe software. But this price contribution is moderate compounding, not a doubling engine.

  New businesses. This is the most external piece of IBM's growth, relying on serial acquisitions to add revenue. HashiCorp, with enterprise value of $6.4 billion, was folded in during 2025, followed by Confluent, at about $11 billion and $31 per share, in 2026. Generative AI's book of business has exceeded $12.5 billion. These will contribute incremental revenue, but the report repeatedly warns that acquisitions which turn cash into goodwill and add revenue do not necessarily create per-share intrinsic value, and this growth is not organic.

  Add the three together: organic volume plus price is roughly mid-single digits, while external growth from acquisitions can add a few more points but requires continuing cash outlays and flawless integration. In combination, IBM revenue is more likely to rise by 30% to 50% over five years than to double. Against Baillie Gifford's hard threshold of "revenue doubling in five years," IBM clearly fails. It is a steady medium-speed compounding machine, not a high-growth stock whose revenue can double. The report's neutral scenario also assumes only about 4% annual growth in Owner Earnings over the next ten years, far away from "revenue doubling."

  Jun 10, 2026
• QAfter five years, what will take over as the next growth engine? Does this "second curve" exist today?5/10

  A

  Bottom line: IBM's "next engine" already exists today and is running: software, meaning Red Hat/hybrid cloud plus data plus automation, layered with generative AI and continuously fed by serial acquisitions. But it looks more like an extension and thickening of the main curve than an independent second curve that can recreate IBM outside its existing business. The handoff is real. Whether it can become a new growth pole at the "fivefold in five years" level is much less clear.

  First confirm that the engine is actually turning. Software revenue in 2025 was $30 billion, up about 11% year over year. In 2026 Q1, software grew another 11%, with ARR reaching $24.6 billion. The data segment grew about 16%, helped by generative AI and the inclusion of Confluent; Red Hat grew about 10%; OpenShift ARR passed $2 billion. Generative AI's book of business has cumulatively exceeded $12.5 billion. These are visible and accelerating sources of growth today, not long-dated stories in a slide deck. The report also confirms that software is the "most important profit engine" and that the moat "has a chance to widen."

  But it is important to distinguish extension from recreation. When Baillie Gifford asks about a second curve, it is really looking for something that can independently take over after the core business tops out, and perhaps become larger than the core itself, as AWS did for Amazon's e-commerce business. IBM's software plus AI is not that kind of independent new pole. It is heavily attached to existing enterprise customer relationships, the mainframe ecosystem, and consulting delivery channels. AI orders are largely sold to big enterprises that already use IBM and are implemented by IBM consulting teams. It deepens the wallet of existing customers, rather than opening a wholly new customer set and business model. The report's bear case says this directly: IBM may simply be "repackaged by the AI narrative," while software growth is "increasingly dependent on acquisition pieces."

  That leads to the biggest concern about the second curve: the fuel is bought, not grown. IBM made 9 acquisitions in 2023, 11 acquisitions in 2024, then HashiCorp in 2025 ($6.4 billion EV) and Confluent in 2026 (about $11 billion). Goodwill had accumulated to about $67.7 billion at the end of 2025 by the report's measure. A second curve that must be sustained by constantly spending cash to buy capabilities requires far greater execution than an organic curve. If integration fails, the engine stalls and leaves impairment risk behind.

  As for whether the two traditional pieces can take over: consulting was nearly flat for full-year 2025, with signings down 13.3% year over year at one point by the report's measure and only recovering in 2026 Q1. It does not have the ability to take over. Infrastructure depends on the mainframe cycle; it is an added boost, not a sustainable engine.

  Overall judgment: IBM's second curve does exist today and is the brightest part of its current growth. The direction, software plus AI plus hybrid cloud, is right. But it is a vertical extension of the core business rather than a horizontal recreation, and the fuel for growth is quite dependent on external purchases. Viewed through the question of whether this second curve can independently support high growth over the next five to ten years, it looks more like a medium-speed engine that must be continuously refueled than a new species that can remake IBM.

  Jun 10, 2026
• QWhat is its core competitive advantage? Will this moat widen or narrow over the next three to five years?6/10

  A

  Bottom line: IBM's core competitive advantage is enterprise-grade switching costs plus mission-critical trust. This includes mainframes and transaction-processing software, Red Hat/OpenShift's position in hybrid cloud, and processes and compliance systems embedded deeply in customers over many years. This is a real but uneven moat. The software and mainframe-related portion will probably widen over the next three to five years, while consulting and traditional infrastructure are closer to "defending the base, with no deterioration already a decent outcome." Overall, this is stable and internally differentiated, not a wide moat expanding uniformly across the whole company. The report's 3/5 score for the overall moat is consistent with this.

  The hardest part is switching cost, and the evidence is solid. By 2026 Q1, software ARR was $24.6 billion, remaining performance obligations were about $69 billion, with about 69% to be recognized as revenue within two years, and consulting backlog was about $31.3 billion by the report's measure. For a large enterprise, core transaction processing running on mainframes, critical middleware, and years of compliance and personnel habits are not assets that can simply be "moved tomorrow." Competitors would have to replicate years of customer code, processes, and audit systems, not a single product. The report lists this as one of IBM's most important moats.

  Trust is the second hard barrier. This is not a consumer brand; it is a mission-critical trust brand. In 2026 Q1, IBM Z grew about 51% year over year, reflecting customer emphasis on resilience, security, and compliance when modernizing mission-critical workloads. Transaction Processing still contributed about $8.6 billion of revenue in 2025 by the report's measure. Financial and government customers buying mainframes and core middleware do not switch vendors the way they change an Office plug-in.

  Why this portion should widen over the next three to five years: software momentum is visible and accelerating. Software grew about 11% in 2025 and another 11% in 2026 Q1. Red Hat grew about 10%, OpenShift ARR passed $2 billion, and HashiCorp plus Confluent fills out the hybrid-cloud and real-time data puzzle. The more modules customers stack on IBM's platform, the higher their migration costs become, deepening the moat.

  But the narrowing or merely defensive parts need to be marked honestly. Consulting is highly substitutable and is naturally not a great business. Full-year 2025 revenue grew only about 2%, while signings were down 13.3% year over year at one point by the report's measure. Traditional infrastructure has product-cycle volatility, and a mainframe surge itself implies the risk of future cyclical fallback. The report states clearly that software growth is "increasingly dependent on acquisition pieces," meaning part of the moat widening is bought rather than purely organically deepened, which raises the execution requirement. At the same time, IBM does not have consumer-internet-style network effects and is not the low-cost provider, so those two moat types are basically absent.

  Net view over three to five years: the direction is wider, but modestly and unevenly. The software and mainframe line should widen the moat at the margin, while consulting and traditional infrastructure are drags and at best hold the line. This is an old-school but real enterprise moat: customers cannot lightly switch, processes are deeply embedded, and long-term contracts have accumulated. It is enough to support steady cash flow, but it is not yet powerful enough to drive the kind of compounding expansion Baillie Gifford seeks, where the moat strengthens and widens faster as usage grows.

  Jun 10, 2026
• QIf its core business is disrupted, does it have the DNA to reinvent itself? How does it handle mistakes and bad news?6/10

  A

  Bottom line: IBM is one of the rare companies that has repeatedly proven its reinvention DNA over more than a century, from tabulating machines to mainframes, from hardware to services, and now to software plus hybrid cloud plus AI. It has repeatedly replaced its own core business. This is its most underappreciated long-term asset. But the honest caveat is that this reinvention looks like a large institution turning slowly: it is slow, relies on acquisition pieces, and depends on drastic portfolio moves such as divestitures. It is not the founder-led, intense, agile self-disruption that Baillie Gifford prefers. IBM's disclosure around mistakes and bad news is relatively candid, but its response to growth anxiety through continued large acquisitions instead of buybacks is itself a signal to watch.

  Start with the hard evidence for reinvention DNA. IBM was founded in 1911, and its uninterrupted quarterly dividend history traces back to 1916. Surviving for more than a century itself shows that it has lived through multiple waves of technological disruption. The most important recent act of deliberate self-pruning was the 2021 spin-off of the traditional managed infrastructure business into Kyndryl, concentrating resources on higher-quality software and hybrid cloud. The report marks 2021 as the first year after the Kyndryl spin-off. Cutting away an old large block and betting on a new direction is exactly what reinvention DNA looks like. The result today: software accounts for about 45% of total revenue, software revenue in 2025 was $30 billion, and the company's center of gravity has moved from selling hardware to selling software and platforms.

  When the core is disrupted, what does it use to save itself? A time window created by switching costs plus continuous purchases of new capabilities. Even if a generation of technology is disrupted, IBM's RPO of about $69 billion and software ARR of $24.6 billion give it several years of buffer to adjust. Facing the AI wave, it responded quickly by using acquisitions to fill out its data and hybrid-cloud stack: HashiCorp (2025, $6.4 billion EV) and Confluent (2026, about $11 billion), while bringing generative AI to a book of business above $12.5 billion. This shows both willingness and funding to change engines.

  But the reinvention method has obvious weaknesses. The self-disruption Baillie Gifford really wants is founder-level, agile, and intense, like Amazon or Tesla actively disrupting themselves. IBM's reinvention is slower, acquisition-dependent, and led by professional managers. The report says plainly that software growth is "increasingly dependent on acquisition pieces" and may be "not entirely organic." Reinvention through purchases carries high integration risk and leaves large goodwill deposits, about $67.7 billion at the end of 2025 by the report's measure. IBM's history also includes lessons of reacting slowly to cloud and to Watson commercialization, starting early but arriving late.

  How it handles mistakes and bad news, to fill in the implicit premise: relatively candidly, but with a response mechanism that deserves scrutiny. On the positive side, the report says IBM disclosed noise factors such as pension settlement charges, the close of tax audits, and currency effects relatively fully, without deliberately packaging unusual financial statements as pure operating miracles. The unattractive items were stated explicitly: the $3.1 billion pension settlement charge in 2024 and the negative effective tax rate in 2025. That adds to management credibility. On the other hand, when faced with the bad news that organic growth is not fast enough, management's response has been to suspend open-market buybacks since 2019 and instead pursue large acquisitions year after year, by the report's measure, using purchased revenue to ease growth anxiety. Whether this is rational capital allocation or a way to mask organic weakness will be tested by whether acquisitions can translate into per-share free cash flow growth over the next few years.

  Overall judgment: reinvention DNA exists, has been validated over a century, and is one of IBM's genuinely scarce qualities. On this dimension, it is stronger than most legacy technology companies. Its treatment of bad news is candid in disclosure, which is a plus, but the response of replacing buybacks with acquisitions leaves a question mark. IBM can survive and can turn, but it does not turn fast enough or organically enough to match Baillie Gifford's ideal agile self-disruptor.

  Jun 10, 2026
• QDoes management, especially the founder, have a long-term view and deep alignment with the company? Is it willing to sacrifice current profits for the next five to ten years?4/10

  A

  Bottom line: IBM's management is a rational, focused professional-manager team with long-term strategic continuity, but it is not founder-led, and its alignment with the company falls into the "institutional arrangements are in place, but personal net-worth exposure is limited" category. CEO Arvind Krishna has a clear strategic direction, is willing to make large acquisitions for the long term, and has highly at-risk compensation. But he is not the founder and owns a very small percentage of the company, so the alignment is not comparable to the founder profile Baillie Gifford values most, where personal wealth is effectively all-in. IBM is willing to sacrifice some current profit for the next five to ten years, as shown by continued acquisitions and continued R&D spending, but that sacrifice is more about spending future money to buy the future than a founder-style all-in tradeoff.

  Start with long-term view and strategic continuity. Krishna has been CEO since 2020, and the report notes that company materials describe him as one of the principal architects of the Red Hat acquisition. In other words, he personally helped design IBM's current shift toward software and hybrid cloud. Since he became CEO, the strategic direction has been highly consistent. That continuity is real: from spinning off Kyndryl, to serial acquisitions including HashiCorp (2025, $6.4 billion EV) and Confluent (2026, about $11 billion), to building generative AI to a book of business above $12.5 billion, this is a clear multi-year line rather than frequent strategic reversals.

  Next, alignment needs two layers and an honest assessment. At the institutional level, alignment is in place: the report discloses that IBM's CEO share-ownership requirement is 10 times salary, above the 7 times peer median shown in proxy materials; about 95% of the CEO's 2025 target compensation was at-risk compensation; and 12 of 13 nominated directors were independent. The governance form is standard and robust. But personal net-worth alignment is limited: the report discloses that Krishna directly held about 418,689 common shares, plus options/RSUs. At the current share price of about $272, those directly held shares are worth about $110 million, a tiny percentage of IBM's approximately $256 billion market value, far less than 0.1%. This is a typical high-paid professional-manager ownership structure, fundamentally different from the founder-heavy ownership Baillie Gifford prefers, where the founder may own several percentage points and personal wealth rises and falls deeply with the company. The report gives "management and capital allocation" only 3/5.

  Is management willing to sacrifice current profit for the next five to ten years? To fill in the implicit premise: yes, but the method leans toward "spend money to buy the future" rather than "painfully cut the present." Positive evidence includes continued R&D investment, with 2025 R&D expense of about $8.3 billion by the report's measure, and large cash outlays to acquire pieces of the future software stack, including about $11 billion for Confluent in 2026. These reduce the distributable portion of near-term free cash flow and represent long-term bets. But one key negative must be noted: IBM has suspended open-market buybacks since 2019 and has not resumed them by the report's measure, while making acquisitions year after year. As a result, share count has risen rather than fallen, from about 898 million shares in 2021 to about 940 million in 2026 Q1 by the report's measure. In other words, what management sacrifices for the long-term growth narrative is current share count reduction and per-share value compounding, while directing capital to acquisitions whose returns still need verification. Whether this is a five-to-ten-year setup or an external mask for weak organic growth depends on whether the acquisitions convert into per-share free cash flow growth.

  Dividend discipline is another mirror of alignment and integrity. In 2026, the company raised its quarterly dividend to $1.69 per share, marking the 31st consecutive year of dividend increases, and it has paid quarterly dividends continuously since 1916. This shows a long-term commitment to shareholder returns and is positive. But dividends are a certain current return and sit somewhat in tension with sacrificing the present for the next five to ten years. They are more a marker of a steady income company than the high-growth playbook of cutting dividends and investing everything into the future.

  Overall judgment: management is trustworthy, long-term, strategically consistent, and candid in disclosure, which are real positives. But IBM does not satisfy Baillie Gifford's core condition of founder-level deep alignment. The person at the helm is an excellent professional manager, not a founder with personal wealth all-in, and his way of sacrificing the present for the long term leans toward continuously spending cash to buy the future rather than founder-style extreme tradeoffs. On this dimension, IBM is above average and steady, but not exceptional.

  Jun 10, 2026
• QIf it disappeared tomorrow, how much would customers miss it? Is its growth model sustainable and not dependent on harming society or exploiting regulation?6/10

  A

  Bottom line: if IBM disappeared tomorrow, the large enterprises, banks, and governments running on its mainframes and core transaction systems would miss it badly. That portion is critical infrastructure that cannot be replaced in the short term and whose failure would create serious consequences. But its consulting and some standardized IT services would be missed much less, with many substitutes available in the market. So IBM's indispensability is extremely strong locally but medium overall. On social and regulatory sustainability, IBM is a rare clean case: it does not make money by harming users, regulatory arbitrage, or addiction-based monetization. It is instead an "order provider" that helps customers meet compliance, security, and audit requirements, which gives it high long-term sustainability.

  Start with how much customers would miss it, and answer in layers. The most-missed part is mission-critical workloads. Customers buying IBM mainframes and transaction processing run core banking ledgers, payment clearing, key government systems, and similar workloads where one second of downtime is unacceptable. Evidence: in 2026 Q1, IBM Z grew about 51% year over year, as customers actively invested during mission-critical modernization, valuing platform resilience, security, and compliance by the report's measure. Transaction Processing still contributed about $8.6 billion of revenue in 2025 by the report's measure. These customers' years of code, processes, and compliance systems are rooted in IBM platforms. If IBM disappeared, they would have nowhere to move in the short term and would feel severe pain. This matches its hardest moat: contract stickiness reflected in RPO of about $69 billion and software ARR of $24.6 billion.

  But the degree of customer pain is uneven. Consulting is highly substitutable; Accenture, Deloitte, and peers can take over, and the report explicitly says consulting has relatively high substitutability and is not a great business. Standardized IT services and parts of infrastructure also have many alternative suppliers. Those customers would find someone else and feel limited pain. IBM is therefore not indispensable as a whole company. Its core mainframe plus critical software is indispensable, while peripheral businesses are substitutable. The report's 3/5 overall moat score reflects this differentiation.

  Now add the implicit premise of social and regulatory sustainability, which is one of IBM's relative strengths. Baillie Gifford would ask whether a company's growth is built on harming society, over-extracting from users, or walking the edge of regulation, such as addictive design, data abuse, or regulatory arbitrage. IBM's answer is clean:

  Its growth model does not depend on harming users. IBM sells enterprise software, consulting delivery, and critical infrastructure to enterprises and governments rather than to consumer users, so there is no addiction-based monetization or attention harvesting controversy. The report states that IBM has no single customer accounting for more than 10% of total revenue and that revenue is geographically diversified, so it is not fragile because it depends on holding one large customer or one country's policy hostage.

  It stands on the side of compliance and order. Customers choose IBM Z and critical middleware precisely for security, compliance, auditability, and governance by the report's measure. In other words, IBM's business helps financial and government customers meet regulatory requirements, so tighter regulation is often a tailwind rather than a headwind. That is the opposite of companies whose business models are impaired when regulation tightens.

  The sustainability risk that needs balancing is not social or regulatory; it is technological and execution-related. The real force that could erode IBM's indispensability over time is technology substitution. If enterprises gradually move mainframe and traditional transaction processing to cloud-native architectures, IBM's most stable cash flow block could be slowly eroded, which the report lists as a core risk. But this is a business-evolution risk, not a sustainability problem where growth depends on harming society.

  Overall judgment: indispensability is extremely high in the core business and medium overall, because losing IBM would hurt core customers badly while peripheral customers can switch. Social and regulatory sustainability is high and clean. IBM is an order and compliance provider; its growth does not rely on social harm, and tighter regulation is more tailwind than headwind. Under Baillie Gifford's dual standard, IBM scores very well on sustainable and non-harmful growth, while its overall indispensability is only moderately above average.

  Jun 10, 2026
• QWhat are the unit economics of this business, including gross margin and incremental returns? Do they improve or deteriorate with scale? Where does the money it earns go?6/10

  A

  Bottom line: IBM's unit economics are a combination of three businesses with completely different economics. Software is close to top-tier SaaS levels, with gross margin above 80%; consulting is a labor-intensive business with gross margin below 30%; and infrastructure sits in between. Overall gross margin is about 58% and improves moderately as software mix rises, showing that unit economics slowly get better as scale increases. But incremental return is pulled down by one key fact: the money earned mainly goes to dividends and serial acquisitions, rather than high-return organic reinvestment or undervalued buybacks, so per-share compounding efficiency is discounted. The report rates capital returns as a pass, but not top-tier.

  First break down the internal split in unit economics, which is the key to understanding IBM. In 2026 Q1, the three segment gross margins were: software 82.8%, consulting 27.5%, and infrastructure 56.9% by the report's measure. This means:

  The software unit economics are first-class. An 82.8% gross margin combined with subscription and renewal models means scale makes each incremental unit more profitable; selling another software copy adds little cost. In 2026 Q1, software ARR was $24.6 billion, up about 10% year over year. This is IBM's most valuable engine.

  Consulting has naturally mediocre unit economics. With 27.5% gross margin and revenue built on headcount, taking on another project requires more people. Scale does not create obvious margin improvement. The report says directly that consulting is "naturally not a super-good industry."

  The overall trend is "rising software mix leads to slowly improving blended unit economics." In 2025, total gross margin was 58.2%, up about 150 basis points year over year. The operating, non-GAAP pretax margin that the company prefers investors to focus on rose from 17.9% in 2024 to 18.8% in 2025 by the report's measure. As software rises to about 45% of total revenue, the mix is migrating toward higher gross margin. So the answer to whether scale makes the business better or worse is: moderately better, driven by structure, meaning software mix, rather than pure volume.

  Light capital expenditure is another positive for unit economics. IBM is not asset-heavy: 2025 PP&E spending was about $1.091 billion, internal software investment about $647 million, and total capitalized investment about $1.738 billion by the report's measure, extremely low relative to $67.5 billion of revenue. This allows more profit to convert into cash. 2025 free cash flow reached a record $14.7 billion, up $2 billion.

  But incremental return must be discounted because of the financing business and large goodwill. The report acknowledges that IBM has a financing business, large goodwill, and acquisition amortization, making precise ROIC "easy to calculate as either pretty or meaningless." A conservative estimate places ROIC roughly near the low double digits: not mediocre, but also not at the level of the best software platforms, where returns are clearly above the cost of capital and expansion can continue without acquisitions. This is where Baillie Gifford would deduct points. The quality of unit economics is good, but incremental reinvestment returns are not top-tier.

  Most important: where does the money earned go? Dividends plus serial acquisitions, not high-return organic reinvestment or undervalued buybacks. This is the core feature of IBM's capital allocation and the reason its unit economics do not fully translate into per-share compounding:

  • Dividends: in 2026, IBM raised the quarterly dividend to $1.69 per share, the 31st consecutive year of increases, showing very strong discipline;
  • Acquisitions: HashiCorp ($6.4 billion EV) in 2025 and Confluent (about $11 billion) in 2026, plus dozens of acquisitions in prior years, lifted goodwill to about $67.7 billion at the end of 2025 by the report's measure;
  • Buybacks: open-market buybacks have been suspended since 2019 and have not resumed by the report's measure, so share count has risen rather than fallen.

  This playbook of turning cash into goodwill means that even if the standalone software business has beautiful unit economics, per-share free cash flow growth still depends on successful acquisition integration. The report repeatedly stresses that this requires higher execution quality and may not create per-share value.

  Overall judgment: unit economics are first-class in software, mediocre in consulting, and upper-middle overall, improving slowly as software mix rises; light capex is a plus. Incremental return is low double-digit ROIC, not mediocre but not top-tier. The money earned mainly goes to dividends and acquisitions, not high-return organic reinvestment or undervalued buybacks, so excellent unit economics have not translated efficiently into per-share compounding. From a Baillie Gifford perspective, IBM's business quality, in gross margin and asset-lightness, is genuinely good, but capital-allocation choices lower the final efficiency of making money from money.

  Jun 10, 2026
• QWhat conditions would need to be true for it to rise fivefold in ten years? Are those conditions realistic? What expectations are implied by today's share price?2/10

  A

  Bottom line: for IBM to rise fivefold in ten years, a difficult set of conditions would almost all have to hold at the same time, and the realistic probability of that full set coming together is low. More importantly, today's share price, about $272, market value about $256 billion, and PE about 24.6 times, already prices in a significant amount of optimism that IBM has transformed into a high-quality AI/software platform company. The margin of safety is not sufficient. The report's final rating is exactly "Watch, wait for a better price first."

  First translate "fivefold in ten years" into hard metrics. Fivefold is roughly a 17.5% annualized total return, including dividends. IBM's current dividend yield is about 2.5% ($1.69 per quarter x 4 / $272), so the share price would need to compound at roughly 15% annually. For a mature giant already worth $256 billion, that implies a market value above $1.2 trillion ten years from now. A fivefold outcome at that scale is historically extremely rare.

  The conditions that must come together:

  1. Revenue growth must accelerate from "more than 5%" toward something close to doubling. The company's current full-year guidance is only "more than 5% revenue growth at constant currency", and its own 2021-2025 CAGR was about 4.2%. A fivefold outcome requires a structural step-up in revenue growth, which would require software plus AI to sustain high growth and become a much larger mix. Realism: low.

  2. Software ARR must keep expanding at least in the mid to high single digits, and organically. Current ARR is $24.6 billion, up about 10% year over year, but the report warns that software growth is "increasingly dependent on acquisition pieces." If growth slips to the low single digits, the current PE itself would struggle to hold, let alone support a fivefold return. Realism: medium-low.

  3. Serial acquisitions must truly convert into per-share intrinsic value, not merely add revenue and goodwill. HashiCorp and Confluent (about $11 billion) must integrate successfully, and goodwill, about $67.7 billion at the end of 2025 by the report's measure, must avoid impairment. Realism: high execution risk.

  4. The valuation multiple must hold or even expand. A fivefold outcome must come either from earnings rising fivefold, which is extremely hard, or from earnings growth plus multiple expansion. But IBM already trades at 24.6 times PE, leaving limited further expansion room. Realism: low.

  5. Free cash flow must rise from $14.7 billion to a level far above $15-16 billion and then compound. The report lists "FCF steadily pushing above $15 billion" as a precondition for the investment case; fivefold requires more. Realism: medium.

  Multiply these five conditions. Each one alone is not fantasy, but asking all of them to hold over ten years produces a low joint probability. That is where Baillie Gifford's fivefold yardstick is sharpest: it forces the question of whether this is a high-probability event. For IBM, the answer is no. It is a steady medium-speed compounding machine, but it does not have the condition set for a fivefold return. The report's bullish scenario produces only 10%-12% annualized returns over ten years, far below the about 17.5% needed for fivefold.

  Now add the key premise of what today's share price implies, because this is the main risk to watch:

  The current price already embeds the optimistic assumption of a successful transformation. The report explicitly maps the current valuation to the middle of fair value, reasonable but not cheap, and identifies one major uncertainty as whether the current valuation has already moved ahead of fundamentals by pricing IBM as a high-quality AI/software platform company. In other words, the market has already labeled IBM as an AI beneficiary and attached a premium to it.

  Peer comparison makes the implied expectation clearer. IBM's current PE is about 24.6 times, while Microsoft is about 24.1 times and Oracle about 23.3 times. Note that Oracle's about 34.5 times quoted when the report was written has now fallen to about 23 times. IBM is no longer obviously cheaper than these two stronger-growth platforms. The market's quality discount for IBM has narrowed to almost nothing. Buying IBM today therefore implies the expectation that it can keep delivering high-quality growth like Microsoft, while its moat, organic growth, and growth profile are clearly weaker than Microsoft's. The report's bear case is precise: IBM may simply be a legacy hybrid technology vendor growing in the low to mid-single digits, repackaged by the AI narrative.

  There is also timing risk: the share price is already higher than the report's anchor. The report judged at about $253.84 that there was no obvious margin of safety, and IBM has risen to about $272, up about 7% from the report's anchor and not far from the 52-week high of $332. That means the report's conclusion that the stock was not cheap is even more valid at the higher price, and the margin of safety has been further compressed.

  Overall judgment: the five sets of conditions required for a fivefold return in ten years are difficult to satisfy simultaneously, and their joint probability is low. IBM is not a fivefold candidate. At about $272 and 24.6 times PE, today's share price already prices in optimism that IBM has become a high-quality AI platform, and the quality discount versus Microsoft has largely disappeared. From a Baillie Gifford perspective, this is a good business with a price that has already pulled forward much of the upside, leaving insufficient downside tolerance.

  Jun 10, 2026
• QWhy has the market not realized all this yet? Is it because the market cannot understand it, looks down on it, or cannot look far enough? What would become the narrative inflection point?3/10

  A

  Bottom line: for IBM, this Baillie Gifford question needs to be reversed first. The main risk is not that the market has not yet recognized its strengths, but that the market may already have recognized them and even overpaid for them. Over the past year-plus, IBM's share price has risen from around the report's anchor to about $272, approaching the 52-week high of $332, with PE rising to about 24.6 times. The market has already re-rated it from cheap old tech to an AI beneficiary. So the real perception gap is not "it is cheap and nobody has noticed," but whether the market may discover that it is not as much like a high-quality growth stock as priced. That is a downside-skewed perception gap, not the upside-skewed gap Baillie Gifford usually seeks.

  First clarify why the market is pricing it this way, testing the three lenses one by one:

  "Cannot understand it" was once valid, but now the risk is the reverse. IBM's accounts are indeed complex: financing, pension settlements, tax-audit closure, acquisition amortization, and mixed software and services reporting make GAAP earnings very noisy; the 2025 effective tax rate for continuing operations was even negative by the report's measure. In the past, this complexity made the market unable to understand it and apply a discount. But since 2025, with software growth of about 11%, generative AI book of business above $12.5 billion, and free cash flow reaching a new high of $14.7 billion, the market has understood the transformation story and assigned a corresponding valuation. The discount has largely been filled.

  "Looks down on it" was the real historical source of the discount, but it has been substantially repaired. IBM was long labeled as old tech that fell behind and missed cloud and mobile, so the market was unwilling to give it a growth premium. That bias once made it cheap. But IBM's PE is now about 24.6 times, roughly in line with Microsoft at about 24.1 times and Oracle at about 23.3 times. The "looked down on" discount has disappeared. The report specifically notes that IBM's quality discount versus Microsoft is not large, while Microsoft's moat and growth are clearly stronger. This shows that the market is no longer looking down on IBM; if anything, it is being generous.

  "Cannot look far enough" cuts both ways. Bulls may argue that the market is not looking far enough because it underestimates the long-term cash flow visibility from RPO, about $69 billion, and ARR, $24.6 billion, plus the durability of the mainframe and AI resonance. But the bear-side version of "cannot look far enough" matters more: the market may not be looking far enough to separate how much of this growth is the large mainframe cycle, with IBM Z up 51% in Q1 and inherently cyclical, and how much is revenue added by acquisitions rather than organically generated. The report's bear case says directly that IBM may just be a legacy hybrid technology vendor growing in the low to mid-single digits, repackaged by the AI narrative. If that is right, the market is looking too far ahead and mistaking cyclical and external growth for sustainable structural growth.

  Now add the implicit premise of what would become a narrative inflection point, in both directions:

  Upside inflection points that would strengthen the "IBM is a real growth platform" narrative: software ARR sustaining organic growth above 10% and proving it can run without acquisitions; Confluent and HashiCorp clearly lifting per-share free cash flow after integration; free cash flow steadily pushing above the $15-16 billion level; and management restarting buybacks and shrinking the share count. If these materialize, the market would be more confident in assigning an AI/software platform premium.

  Downside inflection points that puncture the current optimistic narrative, which need more attention: the mainframe cycle rolls over and infrastructure growth reverses from +12%; software ARR growth slips below 5%, revealing dependence on acquisitions; consulting book-to-bill remains below 1, after 2025 signings were down 13.3% year over year at one point by the report's measure; a large goodwill impairment appears, proving that acquisitions in recent years did not create real value; or the market recategorizes IBM as a low-growth consulting/infrastructure company rather than an AI beneficiary, with PE moving from 24 times back to the mid-teens. The report estimates that if such a re-rating occurs, today's price could imply about 30%-45% permanent capital loss, driven mainly by having bought too expensively rather than by business collapse.

  Overall judgment: IBM is not a typical Baillie Gifford case where the market has not yet recognized the company's strengths. The opposite is true: the market has recognized them and re-rated IBM from cheap old tech to an AI beneficiary. The discount has largely closed and PE is in line with Microsoft. What is underpriced is instead the downside: cyclicality and acquisition dependence inside this growth cycle. The most likely narrative inflection point is therefore negative: a mainframe-cycle rollover or failed proof of acquisition quality triggering valuation mean reversion. From a Baillie Gifford perspective, this is a setup with a downside-skewed perception gap and insufficient margin of safety, the opposite of the great growth stock the market does not yet understand.

  Jun 10, 2026