Platform IP + disclosure = pricing power.
Accenture, IBM and Capgemini are the only firms putting AI on the income statement at scale. Disclosure is now a moat.
We ranked the world’s 50 largest IT/ITeS firms by what they have built, hired, signed and disclosed — not by what they say in keynotes.
Between ChatGPT’s launch and May 2026, the AI services market split in two. Two firms industrialised the work and disclosed the numbers. The rest sold press releases.
cumulative Advanced AI bookings across 11,000 projects. $4.8B converted to revenue inception-to-date as of Q1 FY26.
GenAI book of business — $10.5B Consulting plus $2B Software, as of Q4 2025.
No other firm globally reports AI work at this scale. Accenture has told the market that Q1 FY26 will be its last quarter of separate Advanced AI disclosure — AI is now “embedded across nearly everything” it does.
Composite score out of 40. Tap any row for the strength, the gap and the signature evidence behind the number.
Tap a row to expand
Each firm is graded 1–5 on eight sub-parameters across two pillars. CD = Capability Depth — the sum of the four Pillar 1 axes (IP, Partners, Talent, M&A), out of 20. Add the four Pillar 2 axes (Deals, Disclosure, Vertical, Speed) and you get the composite total out of 40.
| # | Company | 1.1 IP | 1.2 Partners | 1.3 Talent | 1.4 M&A | 2.1 Deals | 2.2 Disclosure | 2.3 Vertical | 2.4 Speed | CD | Total |
|---|---|---|---|---|---|---|---|---|---|---|---|
| 01 | Accenture | 5 | 5 | 5 | 5 | 5 | 5 | 5 | 5 | 20 | 40 |
| 02 | IBM | 5 | 5 | 5 | 4 | 5 | 5 | 4 | 5 | 19 | 38 |
| 03 | TCS | 5 | 5 | 5 | 3 | 5 | 4 | 5 | 4 | 18 | 36 |
| 04 | Infosys | 5 | 5 | 5 | 3 | 5 | 4 | 5 | 4 | 18 | 36 |
| 05 | Capgemini + WNS | 4 | 5 | 4 | 5 | 4 | 4 | 4 | 4 | 18 | 34 |
| 06 | Deloitte | 4 | 5 | 4 | 4 | 4 | 3 | 5 | 4 | 17 | 33 |
| 07 | Genpact† | 5 | 4 | 4 | 3 | 5 | 5 | 4 | 3 | 16 | 33 |
| 08 | EXL† | 5 | 4 | 4 | 3 | 4 | 4 | 5 | 4 | 16 | 33 |
| 09 | EPAM† | 4 | 4 | 5 | 4 | 4 | 4 | 4 | 4 | 17 | 33 |
| 10 | HCLTech | 4 | 4 | 5 | 3 | 4 | 4 | 4 | 4 | 16 | 32 |
| 11 | Wipro | 4 | 5 | 5 | 3 | 4 | 4 | 3 | 4 | 17 | 32 |
| 12 | Cognizant† | 4 | 5 | 5 | 4 | 5 | 3 | 3 | 4 | 18 | 32 |
| 13 | NTT DATA | 4 | 4 | 4 | 3 | 4 | 3 | 4 | 4 | 15 | 30 |
| 14 | PwC | 3 | 5 | 4 | 3 | 4 | 3 | 4 | 4 | 15 | 30 |
| 15 | EY | 3 | 5 | 4 | 3 | 4 | 3 | 4 | 4 | 15 | 30 |
| 16 | Booz Allen Hamilton | 4 | 4 | 4 | 3 | 4 | 4 | 4 | 3 | 15 | 30 |
Source: AIM Research · Top 16 of 57 ranked entities · Cut-off 22 May 2026

All 57 ranked entities — 50 in the universe plus 7 just outside top-50 by revenue — grouped by score band.

The full universe of IT/ITeS firms scored this edition. Composite score shown where AIM Research disclosed it.
Source: AIM Research · Cut-off 22 May 2026

Four patterns the top-20 view misses entirely.
EPAM, Globant, Persistent and Coforge are growing 3–5× faster than the Indian Big Four — but on a much smaller base. The strongest of them (EPAM) now belongs in the leadership conversation on capability per dollar of revenue.
Genpact and EXL have positioned aggressively enough — and disclosed AI revenue transparently enough — to lead BPM outright. Concentrix, Conduent and TaskUs trail. Capgemini's $3.3B WNS deal signals that scaled IT firms now view BPM as the AI deployment surface, not a separate industry.
Leidos, Booz Allen, SAIC and CGI are running aggressive sovereign and defence-AI plays largely invisible to commercial-AI watchers — but materially shaping US, Canadian and allied government AI procurement.
Kyndryl has built the most credible AI play in this cohort — Kyndryl Bridge plus the Agentic AI Framework (Jul 2025), with 25% of signings now AI-related and record $18.2B FY25 signings (+46%). DXC and Atos face structural pressure as AIOps replaces traditional managed services. Kyndryl is the surprise of the universe.
The AI-services product launches that built this leaderboard.
The 57-firm universe splits into nine commercially distinct segments. Each tells a different AI-readiness story.
Accenture, IBM and Capgemini are the only firms putting AI on the income statement at scale. Disclosure is now a moat.
TCS, Infosys, HCLTech and Wipro hold the talent base — but acquisition inactivity and AI-deflation pressure cap the upside.
Genpact and EXL elevated into Leaders on transparency. Capgemini's $3.3B WNS deal absorbs BPM into the AI delivery surface.
Deloitte leads on trust. PwC, EY and KPMG combined committed >$4.4B to OpenAI/Microsoft — but none discloses AI revenue at Accenture's resolution.
Booz Allen, Leidos, SAIC and CGI are running production AI inside defence, civil and allied agencies — largely invisible to commercial AI watchers.
EPAM, Globant, Persistent and Coforge grow 3–5× faster than the Indian Big Four. AI-native engineering culture is the differentiator.
AIOps replaces managed services. Kyndryl Bridge + Agentic AI Framework = 25% of signings AI-related. DXC and Atos face structural pressure.
Teleperformance, Concentrix, TTEC and Foundever face per-seat compression as voice-AI lands. The €100M Sanas-class bets are defensive.
NTT DATA, NEC, Fujitsu, Hitachi DS and NRI lead Asia-Pacific AI but trail on global brand outside home markets.
Tap any prediction for the underlying logic.
One concrete move per buyer persona — plus the anti-pattern to refuse.
Insist on disclosed AI bookings before signing a frame contract.
Accenture, IBM and Genpact publish AI revenue by quarter. Make this table-stakes in your next RFP — it is the cheapest fraud filter you have.
Negotiate outcome-linked pricing while AI deflation is still a vendor concession.
HCLTech, Infosys and TCS have publicly conceded 3–5% revenue compression from AI productivity. Capture it now — the window narrows once vendors re-price.
Demand an AI-revenue disclosure line in the next vendor scorecard.
If your top-three providers cannot quote a GenAI book, you are over-indexed on capability theatre. The Leaders all can.
Go long the disclosers; underweight the unmeasured.
ACN, IBM, GENP and EXL all publish AI bookings or revenue. Pure managed-services names without an AI signal are most exposed to AIOps re-rating.
BPM scale-up windows are closing fast.
Genpact and EXL are the next $5B+ targets after WNS. Coforge–Cigniti and EPAM–NEORIS prove mid-tier roll-up math still works.
Six honest limits on the scoring, the universe and the cut-off.
Each 1–5 mark is an analyst grade against public disclosure, briefings and primary research. Two reasonable readers might differ by a point on any axis.
Firms with strong capability but weak public disclosure (Cognizant is the canonical case) score lower than their internal reality. The index measures what is visible.
57 firms ranked. Excludes hyperscaler-owned services arms (AWS Professional Services, Azure Industry Solutions, Google Public Sector), captive GCCs, and regional boutiques below the revenue threshold.
Revenue figures mix FY24, FY25, CY24 and Q1 FY26 depending on the most recent audited filing. Compare scores, not raw revenue, across the table.
Booz Allen, Leidos and SAIC's classified AI workloads are under-counted by design. Their commercial-AI scores should be read as a floor, not a ceiling.
22 May 2026. Post-cut-off deals (e.g. any Q2 FY26 disclosures from Accenture, IBM or the Indian majors) are not reflected. The next edition refreshes in November 2026.
It is about who has put GenAI on the income statement.
A repeatable, evidence-based ranking of the firms shaping enterprise AI delivery worldwide.
Discussion
What's your take on this story?
No comments yet — be the first to share your view.