Research Findings TrackerLab Products About Kaal →
Indian IT majors Q2 FY27 AI services and deal wins test
Lab call · Markets and macro · Window 01 October to 15 November 2026 · Reconciliation within 7 days of the second earnings release

Indian IT Q2 FY27: the forward metrics carry the test.

Two Lab calls on Indian IT died on trailing revenue thresholds in July while the forward-looking lines inside the same results were the ones moving. This call tests those lines instead, on an either-or structure, and states the disclosure problem that could leave one leg ungradeable.

Tempora's call. For the September 2026 quarter, either TCS and Infosys combined AI-services revenue grows above 10 percent year on year or their combined deal-wins TCV grows above 20 percent year on year, on the figures each company discloses in its Q2 FY27 results. Either leg clearing fires MET. Both legs falling short fires FAILED. Reference: TCS quarterly fact sheet and press release, and the Infosys Q2 FY27 press release and SEC Form 6-K.

Falsifier, verbatim: “TCS + Infosys Q2 FY27 combined AI-services revenue growth >10% YoY OR combined deal-wins TCV growth >20% YoY fires MET”.

Stated probability: 0.55. Base rates quoted below: combined TCV was about flat year on year in Q1 FY27, and the +20 percent leg needs roughly 15.7 billion dollars against the 13.1 billion dollars combined in Q2 FY26. Calibration tier: Lab, signature under test. Reconciliation within 7 days of the second of the two earnings releases.
Under test in the Lab. This is an experimental signature, published on a dated window with a fixed check-back while it earns its place. It lives in the Lab, our surface for signatures still proving themselves. A call graduates to the flagship Tracker once it clears calibration. Track its window in the Lab →

The base rates, both legs

Deal wins first, because both companies disclose them. Q1 FY27, reported in July 2026: TCS total deal TCV 9.5 billion dollars, Infosys large-deal TCV 3.6 billion dollars with 61 percent net new, a combined 13.1 billion dollars against the 13.2 billion dollars the same pair reported in Q1 FY26 (TCS 9.4 billion, Infosys 3.8 billion). That is about flat year on year. The relevant comparison quarter for this call is Q2 FY26, October 2025: TCS 10.0 billion dollars, up 16 percent year on year, and Infosys 3.1 billion dollars with 67 percent net new, a combined 13.1 billion dollars. Above 20 percent therefore asks for roughly 15.7 billion dollars combined.

On the observed year-on-year pairs at this granularity, neither pair delivered a combined 20 percent step, so the base rate for the TCV leg reads zero of two. Two observations is a sample of two and the article does not dress it up as more.

AI services second. TCS put its AI-services revenue at a 2.6 billion dollar annualised run rate in Q1 FY27 and disclosed growth of close to 14 percent, and the basis differs across the record: Forbes India and Upstox describe the step as sequential, while Tempora's own July reconciliation of TCSQ1FY27 recorded it as 13.6 percent year on year. Infosys reported AI revenue at 8.2 percent of its 5,082 million dollar quarterly revenue, roughly 417 million dollars, with no comparable prior-year figure disclosed in the same form. If TCS holds anything near its recent pace, the 10 percent leg clears; the live uncertainty on that leg is disclosure rather than performance.

The trailing picture is quoted for contrast: TCS grew 3.2 percent year on year in constant currency in Q1 FY27, Infosys 2.4 percent, and Infosys narrowed FY27 guidance to 1.5 to 3.0 percent constant currency by cutting the top end. The forward lines and the trailing lines are telling different stories, which is the whole reason this call exists.

Method: lesson applied

The lesson applied here is forward-looking metrics, logged as M5 after TCSQ1FY27 and INFYQ1FY27. TCS graded FAILED against a 5.0 percent constant-currency threshold at 3.2 percent, and Infosys graded FAILED against 3.0 percent at 2.4 percent. Both reconciliations recorded the same finding: the readings were about structural continuation while the thresholds were set for acceleration, and the movement inside those results sat in AI services, deal wins and operating leverage rather than in trailing quarterly revenue. This call moves the test onto the lines that were moving.

The limitations of that move are three, and they are named before the window opens. Forward metrics are company-defined, so their definitions can shift between quarters. Disclosure is voluntary, so a leg can become ungradeable. And the either-or structure makes the call easier to hit than a single-metric test would be, which is why the stated probability of 0.55 sits close to a coin instead of high: two shots at one bar deserves a discount for the structure, not credit for it.

Test condition, grading rule and falsifier

MET: either leg clears. FAILED: both legs fall short. The falsifier reads “TCS + Infosys Q2 FY27 combined AI-services revenue growth >10% YoY OR combined deal-wins TCV growth >20% YoY fires MET”.

AI-services leg. Construction: TCS disclosed AI-services revenue (annualised run rate as the company reports it) plus Infosys disclosed AI revenue (its published AI share of revenue applied to quarterly revenue), for Q2 FY27 against the same construction for Q2 FY26. If either company omits a comparable figure for both quarters, this leg is ungradeable and the call grades on the TCV leg alone, with the omission recorded.

Deal-wins leg. Construction: TCS total deal TCV plus Infosys large-deal TCV, Q2 FY27 against the 13.1 billion dollars combined in Q2 FY26. Above 20 percent growth, roughly 15.7 billion dollars, clears. The two series are not like for like, since TCS publishes total TCV while Infosys publishes large-deal TCV only, and the reconciliation will quote both components separately as well as the sum.

If both legs are ungradeable the call grades VOID. The live failure branches: enterprise deal cycles slipping past the quarter end, a strong Q2 FY26 comparison base on the TCS side at 10.0 billion dollars, definitional resets on AI reporting, and the FY27 guidance cut already on the record as a company view of the demand environment.

Chart-side note

This instance asserts no transit or period mechanism. The dated fact sheet that governs astronomical claims for this batch was unavailable at publication, and Tempora's rule is that chart-side dates come from the dated sheet or they do not appear. The call ships on its measurable side alone.

That fits a Lab entry. Lab signatures are under test, the surface exists so experimental readings grade in public before they reach the flagship Tracker, and the July cohort on Indian IT is a live reminder that a mechanism story does not rescue a badly chosen metric.

Limitations, stated

Sources

Frequently asked

What is the Indian IT Q2 FY27 call?

For the September 2026 quarter, either TCS and Infosys combined AI-services revenue grows above 10 percent year on year, or their combined deal-wins TCV grows above 20 percent year on year. Either leg clearing fires MET; both falling short fires FAILED. Stated probability 0.55.

Why forward metrics rather than revenue?

Lesson M5, applied after TCSQ1FY27 and INFYQ1FY27. TCS graded FAILED at 3.2 percent constant currency against a 5.0 percent threshold and Infosys at 2.4 percent against 3.0 percent, while AI services, deal wins and operating leverage were the lines moving inside the same results. Both reconciliations recorded the metric-selection cause.

What are the base rates?

Combined deal-wins TCV was about flat year on year in Q1 FY27 at 13.1 billion dollars against 13.2 billion, and the Q2 FY26 comparison base is 13.1 billion, so above 20 percent needs roughly 15.7 billion. Neither observed pair delivered a 20 percent step, a base rate of zero of two on a sample of two. On AI services, TCS disclosed a 2.6 billion dollar run rate growing close to 14 percent, which would clear the 10 percent leg if the pace holds and the disclosure continues.

What happens if a company stops disclosing?

The AI-services leg is ungradeable if either company omits a comparable figure for both quarters, and the call then grades on the deal-wins leg alone with the omission recorded. If both legs are ungradeable the call grades VOID.

When does Tempora reconcile?

Within 7 days of the second of the two Q2 FY27 earnings releases, expected in October 2026, on the public Lab surface with both components quoted separately as well as the combined figure.

Lab-tier forward call published by Tempora Research. Lab entries are experimental signatures under test on dated windows. Methodology reproducible against the public engine using Swiss Ephemeris with True Pushya Paksha ayanamsa (PVRN Rao). Internal audit log maintained. This article does not constitute investment, financial, legal, medical or professional advice. First published 30 July 2026 by Tempora Research.

Methods & Data

Tempora's calibration runs on the Swiss Ephemeris with the True Pushya Paksha ayanamsa by PVRN Rao. Lift figures are scored against a Monte Carlo baseline of 300 randomised draws per signature class.

Methodology: Calibrated lift · Falsifier discipline · Forward-call tracker