FULL PUBLISHED RESEARCHEarnings and management commentary analysis
SAGE PRO | Buy-side equity research | 15 November 2026 | NSE: INFY | BSE: 500209 | Q2 FY27
To the Portfolio Management Team. Reference price ₹1,141.50 (13 November NSE close). Research stance: WATCH. Horizon: 3–12 months, supported by FY29 earnings underwriting. Unless specified, financial figures are consolidated ₹ crore. CC means constant currency; owner PAT excludes non-controlling interests. All scenarios are illustrative, not personalized recommendations or execution instructions.
1. The investment decision in one page
Key takeaways
- Our view: Infosys is worth spending time on again.
- The valuation is no longer demanding, the balance sheet is strong, and there is a credible FY28 growth catalyst.
- But the earnings recovery is not yet visible in the numbers.
We would keep the name on the active investment list rather than assume a sustained re-rating is imminent.
The most important feature of the Q2 call was the contradiction between the company's forward deal wins and its current revenue outlook. Management signed $5.2 billion of large deals, including a net-new financial-services contract worth more than $2 billion over seven years.
AI revenue reached 10.5% of group revenue. Yet FY27 constant-currency (CC) growth guidance was reduced to 1–2%. The business is winning work, but it is not converting that success into enough growth today.
[S1–S3, S11]
Our variant view is that FY28 could be better than the current earnings trajectory suggests. The new financial-services contract offers a visible contribution, while a stabilization of retail and high-tech demand would make the recovery broader. What we refuse to underwrite is a return to historical growth rates simply because the company has a growing AI business.
This is not a high-growth technology thesis. It is a recovery thesis in a profitable incumbent that may be closer to an earnings trough than the recent guidance revision suggests. The distinction matters. We do not need a large multiple expansion for the stock to work, but we do need earnings to improve.
What changed this quarter?
The guidance midpoint fell from 2.25% to 1.5%. We treat that as a negative earnings signal. Management has a better view of near-term client budgets than we do, and a downgrade in the same quarter as large bookings suggests the immediate pipeline is still not enough to compensate for volume and pricing pressures. The positive signal is the specific FY28 contribution expected from the mega deal. The investment opportunity is to identify whether those two trajectories converge.
The price at which we are asked to underwrite this is ₹1,141.50. On the existing trailing owner-earnings proxy, that is around 15×. That is attractive enough to justify serious work, not cheap enough to ignore the downside. We would rather buy evidence of an improving business than a persuasive presentation about artificial intelligence.
2. AI is growing. Is the business growing because of it?
The AI revenue bridge is the most revealing exhibit in this report
Key takeaways
- Dollar revenue moved from $5,082m in Q1 to $5,145m in Q2.
- Applying management's reported AI shares to those totals gives a rough estimate of the AI-related revenue category and what remains outside it.
[S2, S3, S7, S11]
The AI category increased by approximately $124m, nearly twice the increase in group revenue. The residual declined by about $61m. That is a much more useful starting point than the 29% AI growth headline.
We are not claiming that AI caused the entire decline in the remaining business. These estimates are derived from rounded revenue shares, not separately audited segments. Acquisitions, foreign exchange, reclassification, individual client losses and macro demand all matter. But the result tells us where to focus our diligence.
A vendor can win in AI in two very different ways. If the client buys additional work and the vendor reduces delivery costs without surrendering all the savings, earnings grow faster than revenue. If the same work is delivered with fewer billable hours and the client captures most of the productivity benefit, AI activity can grow while the vendor's economic opportunity shrinks.
Infosys has the skills to participate in both outcomes. Only the first is especially attractive to shareholders. Management did not disclose AI-specific margins in the call. Until that changes, we need to look at overall revenue growth, utilization, pricing, margins and cash rather than assign a premium to an AI-labelled revenue number.
Underwrite the new deal as a ramp, not a contract announcement
The $2bn-plus seven-year financial-services deal is the clearest forward catalyst. Management expects transition in the latter half of Q4 FY27 and more than one percentage point of FY28 revenue growth contribution. It described employee transfers without an asset purchase, and a margin profile broadly comparable with prior large contracts. [S2 pp2–4,10–11]
That is useful because it gives us a date and an observable outcome. It does not mean the entire contract is revenue today, or that dividing total value by seven produces a reliable annual forecast. The commercial terms, ramp schedule and recognition milestones are not fully disclosed. We would not build FY27 earnings around FY28 growth.
One percentage point of gross contribution can still be offset by one percentage point of legacy contraction. That is why the deal is a necessary but insufficient part of our recovery thesis. We want management to show the transition progressing, group revenue accelerating and cash conversion holding up.
The legacy book needs to stop getting worse
Financial services is the largest end-market and central to the new deal pipeline. Life sciences is growing rapidly, but from a much smaller base. Retail and high technology together represent 18.3% of revenue and are contracting. A stabilization in those two businesses could improve group growth without requiring a dramatic acceleration in AI.
North America contributes 54.6% of revenue and Europe 32.5%. That makes the next stage of the cycle heavily dependent on US and European technology budgets. Cost-takeout spending can remain resilient even while discretionary modernization remains weak. We need to understand which type of work is driving bookings: a cost-reduction mandate can be large and commercially valuable without generating the same incremental economics as a new transformation budget. [S3 p1]
There is another reason for caution. Management attributed about 0.4 percentage points of Q2's 1.2% sequential growth to acquisitions. That leaves roughly 0.8 percentage points ex-acquisition, on the rounded disclosure. It is not a full audited organic bridge, but it is a better starting point than assuming every percentage point of reported growth came from existing operations.
3. The franchise is strong; pricing power is the question
Key takeaways
- Infosys is not a commodity staffing firm.
- Its most durable advantage is the ability to deliver complex, mission-critical projects at scale.
- Replacing a trusted provider in a bank or a large enterprise comes with real operational risk.
That gives Infosys access to work that a small competitor cannot easily win.
The distinction we make is between franchise strength and pricing power. The client can remain with Infosys for years while demanding better economics at each renewal. Switching costs protect the relationship more reliably than the price. With AI reducing the time required to perform some tasks, that distinction becomes even more important.
The $1bn renewal reported in Q2 supports client relevance. It does not tell us what margin Infosys earns on the renewed work. We would rather see stable realization and contribution margins than a larger renewal headline.
Peer comparison: where Infosys actually stands
Infosys is profitable relative to much of the sector, but TCS has a higher reported margin and HCLTech is growing faster sequentially. We should not confuse this with a perfectly matched profitability ranking: the companies have different mixes of software, engineering and IT services, and Wipro's disclosed margin relates to its IT services segment. [P1–P3, S10–S11]
We are not looking for Infosys to lead every metric. We are looking for its growth and earnings trajectory to improve enough to justify a better relative valuation. Until it demonstrates that, there is no strong reason to apply a premium for AI.
Margin is the next important variable
Utilization excluding trainees rose to 85.4% from 84.9% in June. Trailing voluntary attrition declined to 12.7% from 13.0%, and offshore effort represented 77.4% of delivery. Headcount, however, increased from 328,062 to 332,542. [S3 p2]
Lower attrition and offshore delivery help the cost structure, but the increase in headcount while growth remains weak deserves scrutiny. It could reflect preparation for new work. It could also mean capacity is being built before revenue arrives. We want to see billable volumes catch up.
Management protected the margin with currency and productivity benefits despite support and acquisition costs. That is good execution, but it is not a new margin cycle. Wage increases, client pricing and transition costs can all make the next two quarters harder.
At Q2 revenue, a 100bp operating-margin move is roughly ₹492 crore of quarterly EBIT before tax. Against ₹7,823 crore of owner earnings, that is meaningful. Our central margin question is whether AI productivity savings offset those pressures without weakening revenue. If it does, the earnings recovery becomes much more credible. If not, margin discipline alone is not enough to support a re-rating.
4. Earnings and cash: what we are really buying
Key takeaways
- Cash generation is the main reason we are prepared to be patient with this company.
- Infosys can produce substantial shareholder value without the capital intensity of an industrial business.
- But that should make us more demanding about the quality of cash, not less.
FY26 is instructive. INR revenue rose 9.6%, but operating profit increased only 5.3%. The ₹1,289 crore labor-code expense explains a material part of the difference: the company reports a 21.0% adjusted FY26 operating margin compared with 20.3% reported. FY26 also included tax-related benefits. We cannot simply add back unusual costs and leave unusual income in our normalized earnings base. [S4, S6]
The September-quarter earnings are not showing meaningful sequential operating leverage. Owner profit increased only 0.7% on 2.1% INR revenue growth. The group PAT of ₹7,838 crore includes ₹15 crore attributable to non-controlling interests; the relevant shareholder profit is ₹7,823 crore. The operating profit of ₹10,330 crore is consistent with a 21% margin. [S1, S11]
Follow the cash, especially through the new deal transition
Operating cash rose ₹1,447 crore. Better receivables and unbilled-revenue movements helped materially, even as cash tax payments rose. The working-capital improvement is welcome, but it is not automatically recurring. We will watch how the cash cycle behaves as large contracts move into transition. [S1 p5]
The issuer reports H1 free cash flow of ₹17,949 crore under its IFRS definition, distinct from the audited Ind AS operating cash figure of ₹18,958 crore. A simple CFO-minus-net-PPE/intangible-spending calculation is ₹17,373 crore. This is a proxy, not the issuer's FCF measure or free cash to equity. Acquisition cash and lease-related spending still matter. The full cash-flow exhibit preserves the original definitions rather than combining them.
FY26 issuer FCF of ₹33,097 crore was lower than FY25's ₹34,549 crore even as owner earnings rose. That is a reminder that a strong historical FCF yield is not a guaranteed shareholder return.
The balance sheet gives us time, but acquisitions need to earn their keep
Current investments of ₹13,761 crore and non-current investments of ₹8,390 crore add further flexibility, but are not all cash equivalents. Lease liabilities total ₹8,525 crore. We would account for these obligations before claiming a precise enterprise value or owner-cash yield. [S1 pp1,5 and notes]
Goodwill and intangible assets increased alongside much higher acquisition cash spending. The acquisition programme may be a sensible way of buying specialized talent and client relationships. But we cannot award the company an organic growth premium for purchased revenue or assume the new assets will earn attractive returns.
Receivables declined and DSO improved to 64 days from 71 a year earlier. The current non-financial unbilled component rose; it is not total IFRS unbilled revenue. We would be concerned if DSO and unbilled balances rise together during the new contract ramp, particularly if operating margin also weakens.
The conclusion for the PM is that the cash and balance sheet reduce the immediate financial risk of underwriting a recovery. They do not remove the risk that the company has to spend more to maintain the same revenue and earnings base.
5. Management: trust the execution, verify the promises
Key takeaways
- We give the management team credit for operating discipline and less credit for the near-term revenue outlook.
- Those are separate judgments.
- The latest guide revision tells us not to assume management's enthusiasm about AI translates immediately into revenue.
The analyst Q&A was particularly useful. Questions on high-tech client weakness, AI productivity givebacks and the economics of the new contract went directly to the quality of the earnings story. Management described high-tech weakness as largely concentrated in a US client, said productivity pressure had not materially intensified and indicated the new deal's margin profile resembled earlier deals. It did not disclose AI margins. [S2 pp9–15]
That is not evidence that AI margins are poor. It is equally not evidence that they are superior. We should use group margins and cash conversion as our proof until management provides a more granular picture.
Capital allocation: a source of shareholder return, not a free lunch
The company completed its ₹18,000 crore buyback in FY26, repurchasing ten crore shares at ₹1,800 per share. Today's reference price is much lower. The difference does not prove that the board made the wrong decision — we have the benefit of hindsight, and circumstances differed — but it underscores that the price paid for buybacks matters. [S4 pp67–68]
A repurchase can lift EPS with no improvement in operating profit. Acquisitions can lift reported revenue without improving return on capital. Our future earnings revisions will separate these effects from actual organic growth.
The October ₹26 interim dividend reinforces the company's willingness to distribute cash. We will not automatically count it as future income for a purchaser after its relevant entitlement date. Similarly, the capital-return policy is a useful framework but not a promise of a fixed annual yield.
For the next two reporting periods, we would focus on four management tests: the next revenue guidance, the actual contract transition, margin after wage increases, and cash conversion through the ramp. A management team earns higher conviction by delivering on those specific items, not by repeating a strategy that we already understand.
6. Valuation: what must go right at ₹1,141.50?
Key takeaways
- At ₹1,141.50, the valuation case is worth examining.
- We start with owner earnings rather than an unverified consensus number.
FY26 owner PAT of ₹29,440 crore less H1 FY26 PAT of ₹14,285 crore plus H1 FY27 PAT of ₹15,592 crore gives ₹30,747 crore trailing owner profit. Using the existing rounded 405-crore-share base gives a ₹75.92 EPS proxy. The implied equity value is about ₹4.62 lakh crore and the trailing multiple approximately 15×. This is a rounded calculation; it is not formally reported diluted TTM EPS. [S1, S4, S10]
There are two ways to get paid: the earnings grow or the market pays a higher multiple for those earnings. We prefer a thesis that can work on the first. Multiple expansion without operating improvement is a much more fragile investment.
The scenarios: a recovery is not the same as a return to old growth
We retain the three-year valuation framework from the source-reviewed research. Starting from ₹75.92, assume annual owner-EPS growth of 0%, 5% and 10% in the bear, base and bull cases, and illustrative exit multiples of 12×, 16× and 20×. These are independent underwriting scenarios, not probability-weighted targets.
The bear case is not a dramatic collapse. It assumes the company wins work but legacy volume contraction and pricing pressure offset the growth, margins are difficult to protect, and the market assigns a mature-service-business multiple. That is a plausible downside even with a strong balance sheet. The 12× assumption is not a floor.
The base case assumes a gradual return to growth. The financial-services ramp helps, retail and high-tech stabilize, and operating discipline plus capital allocation support 5% annual owner-EPS growth. We do not need an AI-driven boom. At 16× the implied annualized price return is approximately 7.2%, before dividends. That is respectable rather than exceptional.
The bull case requires more. AI work becomes genuinely incremental, the traditional business improves, and the company retains enough productivity gains to support margins. Earnings and the valuation multiple both expand. The outcome is attractive, but we do not yet have evidence that it should dominate our underwriting.
What is the price asking us to believe?
For a hypothetical 10% annual price return over three years, excluding dividends, ₹1,141.50 must become about ₹1,519.34. At 16× the company needs terminal EPS of approximately ₹95.0, implying roughly 7.8% annual EPS growth from the trailing proxy.
This is the reason we are not yet more aggressive. A modest growth recovery does not necessarily deliver a double-digit price return at a 16× exit multiple. To clear that hurdle, we either need stronger earnings, distributions that improve total return, or a higher multiple. We do not want the third variable to be the main reason to own the shares.
The grid strips away the narrative. With 5% EPS growth and a 12× multiple there is still downside from the reference price. With no EPS growth and a 20× multiple there is upside, but it is entirely a re-rating. We should be comfortable with the operating assumptions before we become comfortable with the terminal multiple.
We also use the historical FCF yield as a check, rather than claiming a precise discounted equity value. Issuer FCF is not automatically free cash to equity; leases, acquisition and reinvestment requirements, cash taxes and working capital matter. We would need a consistently normalized after-lease and acquisition cash path before presenting a certified DCF fair value.
Relative valuation: a reason for further diligence, not a trade signal
These are matched-price-date, Q2 run-rate multiples, not forward consensus or trailing-year multiples. TCS is excluded from the multiple column to avoid mixing statutory and adjusted earnings. The relative economics and accounting definitions differ. [P1–P4, S10–S11]
HCLTech's stronger sequential CC growth provides a plausible reason to pay a higher multiple. Infosys's margin strength is a reason to keep it under review at a lower run-rate multiple. Neither observation proves a mispricing. We need to underwrite the incremental earnings before drawing a relative-value conclusion.
The stock has an interesting valuation for a high-quality franchise with a possible growth recovery. It does not have an obviously compelling valuation for a business that remains stuck at very low organic growth.
7. The portfolio decision and the next catalysts
The next three to six months are about evidence, not enthusiasm
Key takeaways
- These are research thresholds and expected windows, not issuer-announced result dates or instructions to trade.
- A management meeting alone is not a catalyst unless it produces new information.
The main downside we need to respect is structural deflation in traditional billable work. AI can make the customer more productive and Infosys less economically valuable at the same time. If clients capture most of the savings, rising AI mix and deal wins may coexist with stagnating group earnings. That is the scenario in which a low P/E is compensation for lower growth rather than a bargain.
The next risk is customer concentration. High-tech weakness was described as largely tied to one US client; retail also suffered from ramp-downs and changes in third-party revenue. A broad customer base does not prevent individual accounts from moving the group's results. We should follow client cohort changes and sector CC growth, not infer the identity or future behavior of unnamed clients.
Currency is another source of risk. Q2's 10.7% INR growth compares with 1.3% USD growth and 1.5% CC growth. The exact currency bridge is more complicated than subtracting the two reported rates, but the implication is straightforward: if the translation tailwind reverses, underlying revenue and margin growth will matter more.
Finally, cash conversion can weaken during acquisition integration and contract transitions. The balance sheet can absorb the spending. What we need to know is whether it generates returns. A simultaneous rise in DSO, unbilled balances and acquisition spend would be a much more concerning signal than an isolated cash-flow movement.
Market structure: a timing input, not the investment thesis
The 13 November reference price of ₹1,141.50 is above the 50-observation average of ₹1,089.42 and below the 200-observation average of ₹1,193.59 in the dated weekday EOD series. The latest 20-observation range is ₹1,038.05–₹1,198.80. These are historical market observations, not guaranteed support or resistance. [S10, Exhibit 13]
The short-term recovery is encouraging, but it has not established a durable longer-term uptrend. We would use this information alongside earnings revisions and relative performance versus IT peers, rather than allow the chart to substitute for the investment case. The PM should know why the earnings thesis is changing before interpreting a move in the stock as confirmation.
The investment decision
We would become more constructive if organic growth improves, the financial-services deal starts on time, and margins and cash confirm that Infosys is retaining the benefits of AI productivity. We would reconsider the thesis if another guidance reduction, delayed transition, or persistent margin and cash deterioration undermines the FY28 earnings case.
The important thing is the combination of metrics. Rising AI share with falling organic growth is not an attractive outcome. Growing revenue with weaker collections is not an attractive outcome. A new deal with a larger headline value and worsening margins is not an attractive outcome. We want the business to show that it can earn more from the work it is winning.
Bottom line for the portfolio team: Infosys is a good business whose valuation has become interesting because growth expectations have deteriorated. The balance sheet gives us time, and the FY28 contract ramp gives us a catalyst. The next leg of the case must come from earnings, not from the AI story. Until the evidence improves, we keep the name on the active investment list and remain disciplined about what we are willing to underwrite.
Visual evidence book — charts, tables and supporting analysis
Exhibit 01 · Revenue growth: currency versus underlying demand
Q2 FY27 versus Q2 FY26. The gap between INR and CC growth is a translation effect and not an exact currency bridge.
*Sources S1–S3. CC removes exchange-rate effects; USD reported and CC differ.*
Exhibit 02 · FY24–FY26 annual revenue
Revenue rose but operating margins did not expand consistently.
*S4, S6. Consolidated INR.*
Exhibit 03 · AI revenue bridge (estimated USD millions)
Key takeaways
- AI rose from $416.7m to $540.2m.
- Other revenue fell from $4665.3m to $4604.8m.
- The estimated incremental AI revenue of $123.5m exceeded the $63m group revenue increase; residual non-AI revenue fell $60.5m.
Classification and currency effects prevent a causal conclusion.
*S2, S3, S7. Rounded estimates, not separately audited segments.*
Exhibit 04 · Large deals and guidance
Q1 $3.6bn; Q2 $5.2bn; H1 $8.8bn. Net-new Q2 TCV is about $3.22bn; renewals/existing $1.98bn. Two mega deals exceed 57% of Q2 TCV. Q2 guidance midpoint fell from 2.25% to 1.5%.
*S2, S7, S11. TCV is not recognized revenue.*
Exhibit 05 · Vertical constant-currency growth
Life sciences leads; high tech and retail are contracting. Financial services contributes 28% of sales and is the key forward booking engine.
*S3 p1. YoY CC, selected segments.*
Exhibit 06 · Geographic revenue exposure
North America and Europe represent 87.1% of group revenue; the recovery depends heavily on US and European enterprise budgets.
*S3 p1.*
Exhibit 07 · Margin bridge (basis points)
The disclosed main items sum to −10bp, consistent with the sequential move from 21.1% to 21.0%.
*S2 pp4–5. Selected disclosed drivers; other items net out.*
Exhibit 08 · Cash flow and capital deployment
Operating cash improved, but acquisitions consumed more cash. Cash conversion must be distinguished from cash available after acquisitions.
*S1, S4, S11. Annual FCF and half-year OCF are different measures.*
Exhibit 09 · Peer operating benchmark
TCS is the margin leader; HCLTech has stronger sequential CC growth. Wipro's IT-services segment margin is not its consolidated EBIT margin. AI categories differ across issuers.
*S11; P1–P3. Q2 FY27 issuer-reported margin definitions.*
Exhibit 10 · Scenario terminal values (₹/share)
Three-year illustrative values from ₹75.92 starting owner EPS and 0%, 5%, 10% annual growth, with 12×, 16×, 20× exit multiples. No dividends included.
*Analyst model; not price targets.*
Exhibit 11 · Valuation sensitivity
At ₹1,141.50, approximate trailing owner EPS ₹75.92 implies 15.0×. The FY26 FCF yield is 7.2% on indicative equity value ₹4.62 lakh crore. A 10% return hurdle without dividends requires FY29 EPS ₹95 at 16× exit.
*S1, S4, S10. Analyst assumptions; no dividend guarantee.*
Exhibit 12 · Decision dashboard
Track the operating thesis against explicit thresholds; thresholds are research triggers, not guidance.
*S2, S3. AI mix alone is not sufficient.*
Catalyst timing and response
Exhibit 13 · Three-to-six-month market structure
The latest close is 4.78% above the 50-observation SMA and 4.36% below the 200-observation SMA. The 200 positive weekday closing observations exclude weekends and are indicative; special trading sessions and price adjustments may affect exact values.
A sustained break above ₹1,194–1,199 with improving sector-relative performance would strengthen recovery evidence. Consolidation above ₹1,089 supports a patient watch posture; a sustained break below ₹1,038 accompanied by further guidance deterioration would challenge the thesis.
These are monitoring references, not guaranteed support, personalized trade instructions or assured returns. Fundamental Q3 confirmation matters more than the price level alone. No sector relative-strength series is claimed.
Source: S10 Sage PRO NSE-labelled price history.
---
Source ledger and methodology
- S1 — issuer signed consolidated Ind AS interim financial statements, quarter and half-year ended 30 September 2026, signed 23 October 2026, physical PDF pp1–6 and notes 2.1–2.19:
- S2 — issuer earnings conference-call transcript, 23 October 2026, PDF pp2–15 and unscripted analyst Q&A; Sage PRO canonical accepted transcript ID 58849, exact quarter 2026-09-30:
- S3 — issuer Q2 FY27 fact sheet, business verticals, geographies, client concentration, utilization, attrition and cash, pp1–3:
- S4 — issuer audited integrated annual report FY2025–26, Board's report pp64–68, financial capital pp30–31 and pp50–51, consolidated financial statements and cash-flow notes:
- S5 — issuer FY2024–25 integrated annual report, financial and capital-allocation comparisons:
- S6 — issuer three-year consolidated IFRS-INR data sheet, FY24/FY25/FY26 P&L, balance sheet and ratios:
- S7 — issuer Q1 FY27 IFRS-INR earnings release, dated 23 July 2026, guidance and Q1 signed-financial extract pp1,7:
- S8 — issuer Q4 FY26 Indian exchange results, 23 April 2026, signed consolidated quarter/year financial results:
- S9 — issuer Q3 FY26 Indian exchange results, 14 January 2026, audited consolidated quarter/nine-month figures:
- S10 — dated NSE close, Sage PRO public.sage_nifty500_price_history, 13 November 2026, NSE_BHAVCOPY_FALLBACK, EOD. Infosys ₹1,141.50; TCS ₹2,407.50; HCLTech ₹1,378; Wipro ₹194.60. Indian market reference: https://www.nseindia.com/market-data/live-equity-market .
- S11 — issuer Q2 FY27 IFRS-INR press release, 23 October 2026, operating margin, free cash flow, earnings and full-year guidance:
- P1 — TCS Q2 FY27 issuer results:
- P2 — HCLTech Q2 FY27 issuer investor relations:
- P3 — Wipro Q2 FY27 issuer results:
- P4 — HCLTech Q2 exchange filing index and quarterly EPS:
Data cutoff: original issuer Q2 disclosures through 23 October 2026; market prices 13 November 2026; research review 15 November 2026. Originality: seven-chapter research synthesis with 13 analytical exhibits, not a quarterly recap.
No separate AI segment profitability is disclosed; the non-AI bridge is an estimate and cannot establish causality. The scenario model uses analyst assumptions, not consensus forecasts. Peer run-rate P/E is an illustrative comparison, not a certified forward multiple.
No PDF attachment is claimed.