AI productivity is the single binary — eClerx either captures it as new revenue or cedes it as price.

• Flat output-per-head. Revenue per employee has sat at ₹20.5 lakh for three years against a 91% time-and-materials book. In a productized-services story this is the metric that would prove the model — and it has not moved.
• The peer is talking; eClerx is not. Genpact separately discloses $1.2B (23.7% of revenue) as Advanced Tech Solutions. eClerx discloses no AI revenue line. The disclosure gap is the disclosure.
• Resolution arrives Aug 6. Q1 FY27 is the first call where the large-scale agentic-AI contract closed in Q4 must print as a quantified revenue line — or the productized-IP premium baked into 18× is at risk versus Genpact's 10×.

FY26 was the strongest year on record — revenue, margin, cash, and returns all inflected up.

Revenue grew 22%, the operating-margin band re-stabilised at 25.5%, and free cash flow cleared 100% of PAT for the second straight year on a 34.8% ROCE balance sheet. Debtor days collapsed from 88 in FY24 back to 59, and ₹622 cr was returned via buyback — roughly the full year of FCF. The 18× multiple is below the company's own five-year average and below every India BPM peer except Datamatics; whether that is cheap or fair depends on whether the AI premium consensus is paying for actually shows up in disclosed revenue.

A 1:1 bonus issue mechanically halved the quote — the real de-rate is roughly 46%, not 72%.

• Bonus mechanics. Shares outstanding doubled on 16 March 2026 from 4.70 cr to 9.40 cr; pre-bonus close ₹3,153 became the post-bonus equivalent ₹1,576 overnight. The headline −72% conflates this with a real correction.
• True drawdown. From the post-bonus 52-week high of ₹2,498 to ₹1,360 is −46% — severe, but on top of a record FY26 print. The valuation reset is the unwind of a Feb-2026 parabola that briefly touched 76× trailing P/E, not a fundamental break.
• Sell-side still constructive. Consensus reads 8 Buy / 0 Hold / 1 Sell at an average target of ₹1,859; Nomura raised FY27 EPS by 2.4% after the Q4 print, Emkay upgraded to Buy on 1 April.

Two Wharton/Booth founders own 54.5%, have not sold a share in 26 years, and skip every buyback.

• Mundhra and Malik. Co-founded eClerx in 2000, took it public in 2007 at ₹315; combined personal stake is ~₹6,930 cr today, pledge 0.00% every quarter, never trimmed. The whole-time founder draws ₹1.7 cr salary with no bonus and no ESOPs.
• The buyback compound. Founders sit out every buyback — promoter holding has drifted from 53.61% to 54.53% across 11 quarters at zero insider cost while the company returned roughly ₹2,930 cr to public shareholders over nine years.
• Professional CEO, kept band. Kapil Jain (ex-Genpact) took the keys in May 2023, cut the margin guide from 28–32% to 24–28%, and held the band through two wage cycles. 10 of 14 valuation-relevant promises kept; FY26 is the strongest year of his tenure.

Two new threats hit the less-protected two-thirds of the revenue base.

• FCC NPRM 26-16. The US FCC adopted it on 26 March 2026 (released 27 March 2026), proposing offshore-call caps and onshore-transfer rights. Customer Operations is ~45% of revenue with no regulatory tailwind (unlike BFSI's KYC/FCC refresh). Cairo, Lima and Fayetteville are partial hedges — not at scale.
• Capgemini-WNS inside the niche. Capgemini's July 2025 acquisition of WNS placed a Tier-1 integrator with bundled consulting + tech + ops directly inside eClerx's BFSI specialist niche. CEO Kapil Jain declined to engage on the competitive risk on the Q1 FY26 call — the non-answer is itself a disclosure.
• The moat has been broken before. A single telco-client renegotiation compressed operating margin from 37% (FY16) to 22% (FY19) — 15 points lost in three years; took four years to rebuild. Top-10 client concentration is still 59% today and one re-negotiation away from the same setup.