Oracle Corporation: A Long-Term Business Owner's Perspective

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?6/10

  A

  The ceiling is high, but it has to be split in two: traditional enterprise software is about expanding and defending a mature existing pie, while AI/OCI cloud infrastructure is about pushing into a new, much larger pie already being divided by the three giants, where Oracle's share is only about 3%. The latter has a high ceiling and low certainty. That is the whole bet embedded in today's valuation.

  Start with the existing pie. Oracle's database, ERP, HCM, middleware, and maintenance support map to the steadiest but moderately growing demand in enterprise IT. The report characterizes this as a good company in a good industry and continued upgrading in a mature industry, not a high-slope growth pole. Its value lies in stickiness and cash flow, not ceiling elasticity.

  The real driver of any fivefold-in-10-years imagination is the second piece: cloud infrastructure and AI training/inference. This is a large, fast-expanding market. According to Synergy Research, the global cloud infrastructure market had already reached about $107 billion in a single quarter in Q3 2025, up nearly 60% from about $68 billion two years earlier. Baillie would like this kind of long runway: on the demand side, the ceiling is not the issue.

  But Oracle is not creating a new market here. It is taking a mix of existing and incremental share in a market that already has incumbents. The same Synergy data show AWS, Microsoft, and Google together at 63% share (29%/20%/13%, respectively), while Oracle is only about 3%, rising but not dominant. The report's characterization is accurate: Oracle is a fast follower, not an incumbent ruler. Starting from only 3% share, the theoretical room for upward revision is indeed large. The condition is that it can keep taking share from the three giants, rather than merely landing a handful of mega-deals.

  Putting the two pieces together, the quality of the ceiling matters. In Q3 FY2026, RPO surged to $553 billion (+325% YoY), which on the surface looks like the ceiling was lifted instantly. But one point the report only implies and does not state plainly must be made: this incremental RPO is highly concentrated. By multiple estimates, the single 5-year, $300 billion Stargate contract with OpenAI accounted for about 94.6% of that quarter's incremental RPO (about $317 billion). That means Oracle's newly opened ceiling today looks more tied to one customer's capital-spending capacity than to a broad, self-propelled market expansion.

  Conclusion: the existing pie (software) has a limited but reliable ceiling; the new pie (cloud/AI) has an extremely high ceiling, but Oracle is still a 3%-share challenger, and the current step-change depends heavily on one super-large customer. This is a combination of a high market ceiling and low certainty that Oracle can capture it. The share curve deserves long-term tracking, rather than assuming the ceiling already belongs to Oracle.

  Jun 10, 2026
• QCan its revenue at least double over the next five years? Will growth mainly be driven by volume, price, or new businesses?6/10

  A

  It probably can, and may even exceed that. But growth is almost entirely single-engine, driven by volume growth in new businesses (OCI/AI cloud infrastructure), rather than volume and price rising together in traditional software. This doubling is high-capital-intensity growth where contracts have been signed, capacity still has to be built, and cash flows out first. Its quality is far below old Oracle's cash-cow growth.

  Start with the base. Oracle's FY2025 total revenue was $57.4 billion (+8%), Q3 FY2026 quarterly revenue had reached $17.2 billion (+22% YoY), and TTM revenue had risen to about $64.1 billion. Under the report's framing, the FY2021-FY2025 4-year compound growth rate was about 9.1%. At that old pace, five-year doubling would not be enough. Structurally, then, doubling can only be carried by accelerating new businesses.

  And the new businesses are indeed accelerating. In Q3 FY2026, cloud revenue was $8.9 billion (+44%), including OCI infrastructure at $4.9 billion (+84%). Looking more closely, AI infrastructure revenue was +243% YoY and multicloud database revenue was +531% YoY. By contrast, cloud applications (SaaS) grew only +13%. The growth engine is heavily tilted: it is mainly driven by new businesses, then by volume; pure price increases are barely the source of this growth round. In OCI/AI, Oracle is instead the party aggressively cutting prices to gain share (see unit economics below).

  More importantly, the visibility of doubling comes from orders, not realized revenue. RPO was $553 billion, +325% YoY, theoretically providing contractual support for revenue doubling over the next several years. Management itself used this as the basis to raise FY2026 capex guidance to about $50 billion (a doubled upward revision from the initial roughly $25 billion). So for Oracle, five-year revenue doubling is not just a story; it is backed by signed contracts with relatively high visibility. That is what separates it from pure narrative companies.

  But three discounts must be applied to this doubling, which the report repeatedly stresses:

  • Concentration discount: about 94.6% of incremental RPO came from OpenAI's single $300 billion contract. The visibility of revenue doubling is, to a considerable extent, visibility into whether OpenAI can pay under the contract.
  • Cash-quality discount: the report notes operating cash flow of $17.36 billion in the first nine months of FY2026, but capex as high as $39.17 billion, implying free cash flow of about -$21.81 billion. Revenue is doubling while free cash flow has turned deeply negative. This is doubling by fronting money to build capacity for future revenue, not old Oracle's doubling through upfront cash collection and asset-light renewals.
  • Conversion discount: RPO is contract value, not revenue. The report lists whether RPO can convert into high-quality, high-margin revenue as the number one issue to prove. Order doubling is not profit doubling.

  Conclusion: revenue doubling over the next five years is a reasonable expectation, with the driver almost entirely in new-business volume growth (OCI/AI), supplemented by volume growth in multicloud databases. But this is a high-visibility, low-cash-quality, high-customer-concentration doubling. Its substance has to be validated by later capex decline and RPO conversion, and it should not be treated as equivalent to high-quality traditional software growth.

  Jun 10, 2026
• QAfter five years, what will take over as the next growth engine? Does this second curve exist today?6/10

  A

  The second curve clearly exists today, and it is already the current main engine: AI infrastructure (OCI) plus multicloud databases. The question is not whether there is a second curve. The issue is that it is too new, too capital-heavy, and too concentrated. The real test five years from now is whether it can move from spending heavily to win orders into a self-funding third-stage steady state.

  Baillie asks this question to see whether the seeds of a company's next growth leg are already in the ground today. Oracle is unusual because its second curve not only exists, but its growth slope has already overtaken the first curve. The first curve (database/ERP/maintenance) is the steady-state cash cow; the second curve (OCI/AI cloud) is the entire source of valuation elasticity. Q3 FY2026 OCI revenue grew +84%, AI infrastructure revenue +243%, and multicloud database +531%, while RPO had reached $553 billion. From the perspective of whether the seed exists, the answer is clearly yes, and it has already sprouted into the trunk.

  So for Oracle, the more meaningful question is the report's implicit follow-up: when this second curve, which has already taken the baton, eventually slows, where is the third leg? Two potential but unproven handoff points are visible here:

  The first is multicloud databases, putting Oracle's flagship database into AWS, Azure, and Google Cloud. The report lists this alongside OCI as a growth engine, and its +531% growth rate suggests it may be a higher-quality curve than raw GPU rental, closer to Oracle's moat in database switching costs. It has a chance to become the successor that continues after AI infrastructure slows and does so with healthier margins, because it sells software capability rather than compute utility.

  The second is vertical industry software, mainly the healthcare IT platform brought by Cerner / Oracle Health. But the report is explicitly reserved here: Oracle acquired Cerner for about $28.3 billion in 2022, and whether it ultimately created enough value still cannot be considered proven today. The report calls it potentially valuable, but with returns not yet validated. The evidence is still insufficient to treat it as the third curve.

  The risk side must be stated honestly, because a Baillie lens cannot avoid it: Oracle's current second curve has an endogenous vulnerability. It depends heavily on one super-large customer. About 94.6% of incremental RPO came from OpenAI's $300 billion contract. That means the sustainability of the second curve itself deserves a discount. If this super contract has problems in execution timing or the customer's own funding, Oracle would need not a third curve to take over, but an emergency substitute for the second curve. The report's downside scenario, OCI growth falling below 30% and RPO growth slowing materially, is exactly the signal that this curve is burning out early.

  Conclusion: the second curve not only exists today, it is already the current main engine. That is Oracle's relative advantage. But it is still in an early stage: fast-growing, cash-burning, and betting on a single customer. The true third curve, whether higher-quality continuation in multicloud databases or deeper vertical penetration through Oracle Health, is still nascent and unproven. The core investment question is not whether there is a second curve, but whether the curve that has already taken over can enter a self-funding steady state. That is exactly why the report rates it Watch.

  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

  The core moat is high switching costs plus a large installed base, concentrated in the old businesses of database/ERP/mission-critical software. It is wide and stable. Over the next three to five years, the judgment is that the moat is broadly stable, slightly wider in places, but increasingly diluted in purity: the old software moat has limited room to widen but remains very strong; the newer OCI/AI infrastructure moat is strengthening but is far from the unassailable position of AWS/Azure, and the lower-margin nature of this business is reducing the quality of Oracle's overall moat.

  First define the source of the moat. The report is right: Oracle's strongest moat is not network effects, but switching costs. Once databases, ERP, finance, HR, and supply-chain systems are deeply embedded in enterprise processes, migration risk is extremely high, implementation cycles are very long, and failure costs are huge. The second source is scale and installed base. Many customers already operate inside Oracle databases, Fusion, NetSuite, Java, industry suites, and support contracts. There is hard data behind this: management has long stressed that the vast majority of customers renew support contracts, and FY2025 cloud services and license support revenue of $44 billion (+12%) is the financial expression of that stickiness. This is a moat repeatedly validated and truly converted into high profits.

  Over the next three to five years, whether the moat widens or narrows needs to be judged in three parts:

  Old software (database/ERP): stable, hard to widen materially. The switching-cost moat is already close to saturation. The customers that can be locked in have mostly been locked in, and competition is more about defending renewal rates than conquering new territory. It will not narrow because migration costs are structural, but it is also hard for this alone to recreate fivefold-in-10-years elasticity. This is a wide and solid moat with limited elasticity.

  Multicloud databases: the widening direction most worth watching. Deploying Oracle Database into AWS/Azure/Google Cloud produced Q3 FY2026 growth of +531%. The attractive part is that it extends Oracle's database switching-cost moat onto competitors' clouds. Even if customers do not use OCI, they still cannot leave Oracle Database. If this keeps scaling, it is a genuine moat-widening piece, not a dilution.

  OCI / AI infrastructure: strengthening, but lower-purity and dilutive to overall quality. The report is restrained on this: the new OCI moat is strengthening, but still has not reached the unassailable level of Azure/AWS, and its share is only about 3%. More honestly, raw GPU compute rental is essentially a low-differentiation business close to a utility. Reportedly, Oracle's AI accelerator gross margin was guided by management in the 30%-40% range, and some older GPU rentals are even loss-making. Even if this business scales, its moat is far weaker than Oracle's traditional software with 70%+ gross margins. As OCI's share rises, it will expand revenue while lowering the average quality of the company's moat.

  Pricing power supports the same conclusion. The report notes that Oracle has some pricing power in database maintenance, ERP, and mission-critical applications, but in OCI, pricing power is materially weaker than in maintenance support. Where the moat is strong, Oracle has pricing power; where the moat is weak (OCI), it has to cut prices to win orders. That is direct evidence of a moat structure that is expanding outward while being diluted.

  Conclusion: the report's 4/5 moat score is fair. The moat is real, validated, and converted into profits, but it is not top-tier (not 5). Over the next 3-5 years, the direction is that the old moat holds, multicloud databases may truly widen it, and OCI strengthens while lowering purity. Overall it will not narrow, but the moat-widening narrative has to rest on high-quality extensions such as multicloud databases, not simply on OCI revenue growth.

  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

  It does have the DNA to reinvent itself, and history has proved it twice: from an on-premise licensing giant to cloud subscriptions, and now into an aggressive AI cloud infrastructure bet. This is Oracle's genuinely scarce trait. The cost is that its reinventions often come through aggressive borrowing and heavy capital bets, and in how it handles mistakes and bad news, a strong founder-led structure is both a source of decisiveness and a risk of slower correction.

  Baillie is asking about corporate regeneration: when the core business is disrupted, will the company self-disrupt, or cling to the old model and decay? Oracle's record here is quite strong.

  Reinvention DNA: yes, and repeatedly validated. Oracle began with on-premise database licenses. When SaaS/cloud pressured its core cash cow, it did not stay locked in the old model. It built OCI, moved Fusion/NetSuite into the cloud, and then placed its flagship database into competitors' clouds through multicloud deployments. Today it is undertaking a third, more aggressive reinvention: moving the overall model from high-margin, low-capital-intensity software toward what the report calls a more cloud-utility-like, capital-intensive model. The scale of Q3 FY2026 cloud revenue share, OCI +84%, and AI infrastructure +243% shows this is not a PowerPoint transformation. It is rebuilding the business model with real money. FY2026 capex guidance of about $50 billion is proof of the size of the bet. For an old-line company with a market value near $600 billion to still self-disrupt at this scale, the DNA is real.

  But the reinvention method carries structural risk. Each Oracle reinvention is pushed forward through leverage plus heavy capital. The report discloses that as of 2026-02-28, total borrowings were about $134.61 billion (up sharply from about $92.57 billion at FY2025 year-end), that net debt issuance in the first nine months of FY2026 was $44.54 billion, and that the company also issued about $4.95 billion of mandatory convertible preferred stock, while free cash flow turned deeply negative. In other words, its reinvention DNA is always tied to a thinner financial safety cushion. The DNA is strong, but each activation pushes the balance sheet tighter. This time especially so.

  Handling mistakes and bad news: mixed. On the positive side, the report notes that Oracle's governance framework is not loose, with clawback policies, anti-hedging policies, and executive shareholding requirements. The 2024 annual report front page disclosed no prior financial-statement corrections, and EY issued an internal-control audit opinion. At the institutional level, there are constraints around mistakes. One detail also shows the board's realism: the FY2025 compensation committee decided not to grant annual cash bonuses to Ellison and Catz, explicitly to preserve cash and support AI strategic projects. That suggests the board at least formally recognizes capital scarcity.

  But the negative risk is equally clear. The report stresses that Oracle is an obviously founder-led company: Larry Ellison owns about 40.6%, and all directors and executives together own about 40.9%. The risk of this double-edged sword in handling bad news is that when strategy is dominated by one deeply controlling figure, admitting that this AI bet was wrong may face greater and slower correction resistance. The report describes the Cerner acquisition as potentially valuable but with returns not yet validated. That is precisely an unresolved example of a mistake not yet openly settled.

  Conclusion: Oracle truly has reinvention DNA, and it is stronger than at most software giants of the same age. That is a substantive reason to be constructive on its long-term vitality. But its reinventions are always high-leverage, heavy-capital, founder-decided moves. The DNA is strong and correction mechanisms exist, but the willingness and speed of correction are constrained by the founder-led structure. For long-term owners, the issue to watch is not whether Oracle can change, but whether this change is too forceful and whether the escape route is sufficient.

  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 five to ten years out?7/10

  A

  Yes. Long-term view, aligned incentives, and willingness to sacrifice current profits for five to ten years out are almost textbook at Oracle: founder Larry Ellison owns about 40.6%, and the company is sacrificing current free cash flow and taking on more leverage to bet on AI for the long term. But a Baillie lens has to add one point: extreme alignment plus extreme control means that if the bet wins, everyone benefits, and if it loses, no one can stop it. Incentive alignment does not guarantee correct decisions.

  Baillie cares most about whether the founder has tied his own fate to the company and whether he is willing to sacrifice the near term for the long term. Oracle has very hard evidence on both.

  Interest alignment: extremely strong. According to the 2025 proxy statement, Larry Ellison held about 1.158 billion shares, or about 40.6% of the company; all current directors and executives together held about 40.9%. This is not symbolic ownership. It is true alignment, with most of personal wealth tied to the company. The report's judgment is on point: management and shareholders are highly aligned, and Larry Ellison's large shareholding creates strong incentive alignment. For long-term investors, this means management is unlikely to damage long-term value for one or two quarters of reported numbers. He himself is the largest long-term shareholder.

  Long-term view plus willingness to sacrifice current profits: the evidence is clear and unfolding now. This is Oracle's most distinct current feature. It did not restrain investment to preserve attractive free cash flow. Instead, it raised FY2026 capex guidance to about $50 billion (a doubled upward revision from the initial roughly $25 billion), at the cost of about -$21.81 billion of free cash flow in the first nine months of FY2026 under the report's figures and borrowings rising to about $134.6 billion. One detail makes the stance even clearer: the FY2025 compensation committee decided not to grant annual cash bonuses to Ellison and Catz, explicitly to preserve cash for the AI strategy. Management itself is sacrificing the near term for the long term. On the question of whether it is willing to sacrifice current profits for five to ten years out, Oracle's answer is a loud yes, perhaps even too willing.

  Management structure: stable, with a new variable. The report notes the September 2025 transition: Clay Magouyrk and Mike Sicilia were promoted to co-CEOs, Safra Catz became executive vice chair, and Ellison remained chairman and CTO. The report's reading is accurate: Ellison still deeply controls technology and strategic direction, while the new CEO pair represents the OCI and applications software lines respectively. This arrangement preserves strategic continuity, but the co-CEO model itself is a new variable that has not yet been tested through a full cycle. The report also separately lists management risk: the co-CEOs have just changed.

  The other side that a Baillie lens must add: the flip side of alignment is lack of counterbalance. The report is clear: highly concentrated ownership is a double-edged sword; the governance structure is not dispersed and democratic, but clearly founder-led; decision style may be more aggressive. This AI bet is the best footnote. One person has pushed the company toward high leverage, heavy capital, and a path that bets on one super-large customer (OpenAI accounts for about 94.6% of incremental RPO). If this path is right, Ellison's vision will be celebrated; if it is wrong, no outside shareholder force can brake it in time. The report therefore rates management and capital allocation at 3/5 and lists whether capital allocation is rational as uncertain. This is not doubt about sincerity or horizon, but doubt about the win rate of allocating capital in a higher-risk way this time.

  Conclusion: measured by Baillie's standard of whether the founder is long-term, aligned, and willing to sacrifice the present, Oracle is one of the rare companies that hits all three. This is the side of Oracle most like a great growth company steward, and a substantive reason for long-term ownership. But the same alignment and control make the company's fate highly dependent on one person's judgment. Incentive alignment solves the question of whether management wants to do the right thing; it does not solve whether the judgment is right. For long-term owners, this is a positive factor, but it must be used alongside an independent judgment of the odds of this specific bet.

  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 regulation?6/10

  A

  They would miss it a lot, but in two layers: traditional database/ERP customers would miss it with extreme pain, because migration is close to impossible, so this layer is highly indispensable; AI compute customers would mostly face the hassle of finding an alternative supplier, rather than losing something irreplaceable. The growth model generally does not depend on harming society or regulation, making it cleaner than most high-growth technology companies. But the new AI infrastructure path introduces new sustainability questions around energy/grid load and single-customer dependency.

  Baillie tests two things here: indispensability, meaning how much customers would suffer if it disappeared, and whether growth is sustainable without harming society and regulation. Oracle needs separate answers.

  Indispensability: extremely strong in the old business, moderate in the new business.

  For the old business layer, the answer is close to irreplaceable. The report repeatedly stresses that the moat comes from high switching costs: once databases, ERP, finance, and HR systems are deeply embedded in enterprise processes, migration risk is extremely high, implementation cycles are very long, and failure costs are huge. The customer base consists of governments, large enterprises, and institutions highly dependent on mission-critical systems. For these institutions, Oracle's disappearance would mean core business systems going down, data migration measured in years, and failure that could threaten operations. Management has long stressed that the vast majority of customers renew support contracts, and FY2025 cloud services and license support revenue of $44 billion (+12%) is the financial measure of that degree of being missed. This indispensability is real.

  The AI/OCI layer is much weaker. Raw GPU compute rental is a relatively standardized commodity, and customers can move workloads to AWS, Azure, CoreWeave, and others. The report notes that Oracle's OCI share is only about 3% and its pricing power is weak. For AI customers, Oracle is more a cost-effective supplier than an irreplaceable partner. It wins large orders through capacity and price, not indispensability. The sole exception is multicloud databases: even if customers use another cloud, they still cannot leave Oracle Database itself, so this area inherits the strength of the old business's indispensability.

  Growth sustainability plus not harming society/regulation: generally clean, with two new questions.

  On the positive side, Oracle's business is inherently quite clean. It sells enterprise productivity tools (databases, ERP, cloud compute), and does not rely on harvesting user attention, manipulative algorithms, or regulatory arbitrage. It is completely different from high-growth models in advertising, social media, or gambling. The report supports this: in economic downturns, Oracle's mission-critical software revenue is usually more resilient than advertising or consumer internet. This is a positive-sum business in which customers become more productive as they use it, with positive social externalities.

  But a Baillie lens has to flag two new sustainability questions honestly:

  First, energy and grid load. AI data centers are energy-intensive, and the report lists data-center capacity and GPU procurement as company-disclosed risks. Large-scale AI infrastructure expansion is drawing global regulatory attention to power supply, grid pressure, and energy consumption. This is not unique to Oracle, but it is a social-cost variable embedded in its new growth model, and it may become a regulatory or cost constraint in the future.

  Second, dependence of growth on a single customer. About 94.6% of incremental RPO came from OpenAI's $300 billion contract. This is not harming society, but it is another sustainability concern. Growth built on one customer's capital capacity and business prospects is inherently less robust than diversified demand. This is exactly why the report separately lists implicit customer concentration risk.

  On regulation, the report notes that Oracle itself lists government contracts, legal/regulatory matters, and complex/changing regulations as risks. But these are compliance costs and uncertainty, not growth through regulatory harm. Oracle does not treat violations as a growth engine, which is consistent with Baillie's red line.

  Conclusion: if Oracle disappeared tomorrow, governments and large enterprises deeply tied to it would miss it with extreme pain. This indispensability is its hardest asset. AI compute customers would miss it much less. The growth model generally does not rely on harming society or regulation and is cleaner than most high-growth technology companies. The energy load and single-customer dependency introduced by new growth are sustainability variables that need ongoing monitoring, not red lights that have already gone off.

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

  A

  Unit economics are splitting: the traditional software piece is a top-tier business (70%+ gross margins, asset-light, high incremental returns), but the AI/OCI piece is a low-quality business closer to a utility (gross margin guided at 30%-40%, some older GPU rentals even loss-making, and extremely capital-heavy). As scale grows, overall unit economics are moving in a worse direction because low-margin OCI is rising rapidly as a share of the business. Almost all the money earned is being poured into data centers and GPUs, and free cash flow has turned deeply negative.

  Baillie is asking about the intrinsic earning efficiency of the business: whether gross margins are high, how much output each reinvested dollar produces, and whether scale effects are tailwinds or headwinds. Oracle has to be split into two parts because it is moving from one set of unit economics to another.

  Traditional software: top-tier unit economics. Database, ERP, and maintenance support are textbook high-quality businesses: software replication has near-zero marginal cost, cash is collected upfront, and renewals are sticky. Under the report's framing, FY2025 operating margin was 30.8% (a consolidated figure already dragged down by heavy-capital OCI), while the pure software gross margin is far higher. This business gets better as it scales: incremental revenue requires almost no incremental capital, making it a genuinely high-incremental-return business.

  AI / OCI: low-quality unit economics, and already dragging the whole. This is the key supplement the report needs. Raw GPU compute rental is essentially a capital-intensive utility:

  • Low gross margin: reportedly, Oracle's AI accelerator gross margin was guided by management in the 30%-40% range, and some older Nvidia GPU rentals are even loss-making. This is a different world from software's 70%+ gross margins.
  • Extremely capital-heavy: the report discloses FY2024 capex of only $6.87 billion, surging to $21.22 billion in FY2025, and already reaching $39.17 billion in the first nine months of FY2026, with full-year guidance of about $50 billion. Every additional AI compute order requires Oracle to front huge GPU plus data-center investment.
  • Incremental returns are uncertain: GPUs carry depreciation and generational obsolescence risk. Whether each reinvested dollar can earn a return above the cost of capital is the report's number one unresolved question: whether incremental AI infrastructure investment can be recovered in the future at above the cost of capital.

  Do unit economics improve or deteriorate with scale? The direction is deterioration. The key is business mix: low-margin, heavy-capital OCI is growing much faster (+84%) than high-margin software (SaaS +13%), which means the revenue mix is tilting toward the lower-quality bucket. As scale grows, the company's average unit economics are being diluted. The report says it plainly: Oracle is partly migrating from a high-margin, low-capital-intensity software model to a more cloud-utility-like, capital-intensive model. This is the opposite of Baillie's preferred pattern of getting more profitable as scale grows, unless Oracle can materially lift OCI gross margins through the networking efficiency, hardware optimization, and power economics mentioned in the report, plus management's claim that 10%-20% of AI data-center spend comes from higher-margin adjacent services (storage, load balancing, identity, security).

  Where does the money go? Almost all into data centers. The answer is extremely clear, and it is the report's central warning: operating cash flow was $17.36 billion in the first nine months of FY2026, while capex was as high as $39.17 billion, implying free cash flow of about -$21.81 billion. In other words, the company not only invested all operating cash flow, but also issued additional debt (net debt issuance of $44.54 billion) to fill the capex gap. The report's owner earnings estimate points to the same conclusion: starting from TTM operating cash flow of about $23.5 billion, subtracting real stock-based compensation of about $4.9 billion and maintenance capex of about $9-10 billion gives a conservative owner earnings midpoint of about $9 billion, but reported free cash flow is negative. The money is not flowing back to shareholders (buybacks have slowed materially and dividends are modest). It is all being bet on AI capacity.

  Conclusion: Oracle is a mixed business of high-quality software plus low-quality compute. The software half has impeccable unit economics. But today's growth center, capital allocation, and cash-flow direction all point to low-margin, heavy-capital AI infrastructure. In the short term, larger scale is diluting unit economics rather than improving them. Whether the business can become better again depends on one unproven thing: whether AI investment can be recovered after FY2027 at returns above the cost of capital. That is the fundamental reason the report rates it Watch and marks stable free cash flow as not passed.

  Jun 10, 2026
• QFor it to rise fivefold in 10 years, what conditions need to hold at the same time? Are those conditions realistic? What expectations are embedded in today's share price?3/10

  A

  For the stock to rise fivefold in 10 years, four difficult things have to be true at the same time, and none is easy on its own: 1. AI infrastructure has to move from cash burn to high returns; 2. capex has to fall sharply after FY2027 and free cash flow has to turn positive; 3. the OpenAI super contract, which accounts for about 94.6% of incremental RPO, has to convert at high quality without renegotiation; 4. the valuation multiple must not compress materially. This combination is realistic but hard. And at today's share price of about $206, the market has already discounted a substantial amount of excellent execution on the AI mega-deals. The report, written when the stock was $181.46, already judged it near the top of the optimistic scenario. The higher price today leaves an even thinner margin of safety.

  Anchor the starting point first. As of the 2026-06-09 close, ORCL was about $205.81, with a market cap of about $592 billion and a TTM P/E of about 37x. At the time of the report (2026-05-20), it was $181.46, with a market cap of about $528.4 billion and a trailing PE of 32.6x. A fivefold rise in 10 years means market value rising from about $590 billion to roughly $3 trillion. That would require Oracle to become a super-mega-cap comparable to today's Microsoft and Nvidia within 10 years. It is not impossible, but the bar is extremely high.

  Conditions needed for fivefold in 10 years (Baillie-style decomposition):

  Condition one: AI infrastructure must be proved to be a high-return business, not a low-margin utility. This is the hardest gate. Current AI accelerator gross margin is guided at 30%-40%, and older GPU rentals are even loss-making. For a fivefold outcome, OCI must structurally lift gross margins through network/power/hardware efficiency plus high-margin adjacent services, so that reinvested capital earns above the cost of capital. The report lists this as the number one unproven issue.

  Condition two: capex must fall materially after FY2027 and free cash flow must turn positive. Current FY2026 capex guidance is about $50 billion, and free cash flow in the first nine months of FY2026 was about -$21.81 billion under the report's figures. The fivefold story requires this to be a short-term bridge, not the new normal. The report is explicit: if capex remains above 20%-25% of revenue for a long period after FY2027, the logic has to be reviewed.

  Condition three: the super contract must convert at high quality, without customer concentration risk detonating. About 94.6% of incremental RPO came from OpenAI's $300 billion, 5-year contract. The fivefold narrative depends heavily on this contract being paid as agreed, not renegotiated or canceled, and on OpenAI itself continuing to raise enough money to perform. This ties the company's fate to one customer's commercial outlook. It is opportunity, but also the concentration risk the report repeatedly warns about.

  Condition four: the valuation multiple must not collapse. Even if the first three conditions hold, if the market reprices Oracle from a growth cloud platform back to mature software plus a mid-tier cloud platform, multiple compression would consume much of the upside. The current forward P/E is about 26.5x, still about 43% above the software industry median of about 18.6x. The valuation already includes a premium, leaving limited room for further multiple expansion. Maintenance or convergence of the multiple is more likely than expansion.

  What expectations are embedded in today's share price? The report's own valuation framework is the clearest lens: it gave a fair intrinsic-value range of $85-$125, an optimistic scenario of $150-$190, and an obviously overvalued zone above $170, and noted that the $181.46 price at the time of writing was already near the top of the optimistic scenario. Today's roughly $205.81 has already exceeded the upper end of that optimistic range ($190). In other words, today's price does not embed a discounted starting point for a fivefold outcome. It embeds the assumption that the market is already pricing close to the optimistic scenario: AI mega-deals will execute excellently, capex will fall as scheduled, and returns will improve meaningfully. The report's judgment is sharp: buying Oracle today is essentially prepaying for several future years of endogenous growth and return improvement.

  Are these conditions realistic? Each condition individually is possible, but all four together form a low-probability joint event. Two of them, improved AI margins and stable conversion of a single super contract, depend on external variables Oracle cannot fully control. More importantly, even if the fivefold story ultimately proves true, today's entry price has already paid upfront for the first stage of good news. Under the report's current framework, even the optimistic scenario annualized expected return is only 9%-13%, far short of a fivefold 10-year outcome.

  Conclusion: a fivefold 10-year rise requires AI to become high-return, capex to fall, the super contract to convert steadily, and valuation not to compress, all at the same time. This is realistic but a low-probability joint event. The current share price around $206 already embeds expectations of the optimistic scenario and has moved above the report's optimistic valuation ceiling, leaving almost no margin of safety for a fivefold outcome. This is fully consistent with the report's conclusion of Watch and insufficient margin of safety at the current price: a good business, but today's price has already paid in advance for future success.

  Jun 10, 2026
• QWhy has the market not recognized all of this yet? Is it hard to understand, ignored, or too long-term? What will become the narrative inflection point?3/10

  A

  For Oracle, this question has to be inverted. The market has not failed to recognize the AI/OCI story; it has fully recognized it, and may even have over-recognized it. This is not an overlooked stock hidden because the market cannot understand it, looks down on it, or cannot look far enough. It is a star stock that has already been repriced by the market and has seen violent valuation swings over the past year. The real perception gap is not that the market has missed the upside, but whether it has discounted execution quality, customer concentration, and the cash-flow cost enough. The narrative inflection point is not the market finally discovering Oracle, but whether RPO can convert into cash at high quality and how the OpenAI super contract is executed.

  The classic Baillie setup for this question is a good growth stock temporarily misunderstood by the market, creating a perception gap. Applied to Oracle, one has to admit honestly: it does not fit that setup.

  The market has already seen the upside, and chased it aggressively. ORCL's past-year stock path is the hardest evidence: the 52-week range was about $134.57 to $345.72. The stock once surged to about $345, then fell back to about $205.81 on 2026-06-09. After the $300 billion OpenAI contract was announced, the stock jumped more than 42% in one day. How is that ignored or too long-term for the market? The market not only saw the AI narrative, it once priced it to the extreme of the optimistic scenario. The report's relative valuation confirms the same point: Oracle's TTM PE was 32.6x (now up to about 37x), and its forward PE of about 26.5x remains about 43% above the software industry median of about 18.6x. It is not the cheap peer. It is the premium one.

  So the real perception gap is whether downside risk is underestimated, not whether upside is ignored. If the market is not looking far enough, the direction may be the opposite: it may be overestimating the quality of AI mega-deals and underestimating capital intensity and financing costs. The report states the strongest bear case clearly: Oracle may not be a cheap legacy software leader in transition, but a case where the market has overestimated the quality of AI mega-deals and underestimated capital intensity and financing costs. The specific discount points the market may be underweighting are three: 1. about 94.6% of incremental RPO is concentrated in OpenAI as a single customer, and concentration risk has been obscured by optimism; 2. AI accelerator gross margin is only 30%-40%, and some older GPUs are even loss-making, which differs from the market's imagined high-quality cloud revenue; 3. reported free cash flow is deeply negative and borrowings have risen to about $134.6 billion, with financing constraints masked by a high valuation.

  What will become the narrative inflection point? Because the market has already priced in the upside narrative, the inflection point is no longer discovery of the story, but confirmation or falsification of the story. The report provides a two-way checklist, which can be summarized into several high-signal events:

  • Upside inflection point (confirmation): revenue continues to grow rapidly while the capex/revenue ratio falls quickly, free cash flow turns positive again and returns to double-digit growth, OCI growth remains above the industry and margins improve meaningfully, net debt/EBITDA declines clearly, and RPO converts into revenue quickly with little bad debt or renegotiation. If these appear together, the market will confirm that high capex was only a short-term bridge, and the optimistic scenario will truly hold. The report even says that if these happen, it would admit it had been too conservative.

  • Downside inflection point (falsification): OCI growth falls below 30% for several consecutive quarters; RPO growth remains high but revenue conversion is weak, or signs of contract renegotiation/cancellation emerge, especially any execution issue in the OpenAI contract; capex remains above 20%-25% of revenue for a long period after FY2027 and free cash flow does not recover; debt keeps rising or another large equity financing dilutes shareholders. If any of these is confirmed, the valuation base will fall back from growth cloud platform to mature software plus mid-tier cloud platform, and a large part of the optimism embedded in the stock price, as the report says, will be removed.

  The nearest observation window is the upcoming Q4 FY2026 results (market expected after the close on 2026-06-10). These nodes concentrate validation of RPO conversion, capex pace, and OpenAI execution signals, and are high-probability moments for the narrative to be pushed in one direction.

  Conclusion: Oracle is not an overlooked growth stock the market has not yet recognized. Quite the opposite: its AI upside narrative has been fully priced by the market, and the stock has moved violently over the past year. The real perception gap is whether downside risks (customer concentration, low gross margin, negative free cash flow, high leverage) have been underestimated by optimism. The narrative inflection point will not be the market finally understanding the story, but hard evidence from RPO and capex cash conversion that confirms or falsifies it. Until then, the report's choice of Watch rather than Buy is a sober response to a price that has already pulled forward optimistic expectations.

  Jun 10, 2026