L&T Finance is a retail-focused NBFC whose economics are improving but are not yet at the Lakshya 2031 destination. Q2 FY27 showed a consolidated book of ₹1,36,635 Cr, retail book of ₹1,34,600 Cr, PAT of ₹975 Cr, NIMs+fees of 10.48%, credit cost of 2.46%, RoA of 2.60% and RoE of 13.56%. The investment question is whether the AI underwriting and distribution build can push credit cost to 2-2.2% by Q4 FY27 and then to 2% or less, while keeping NIMs+fees in the 10-10.5% corridor. If that happens, the 2.8-3% Q4 FY27 RoA corridor and the FY31 3-3.2% target become plausible. If it does not, the stock remains a 2.4-2.6% RoA lender with a longer wait for the next leg of returns.
The business economics are a mix of high-yield secured and unsecured retail lending. Gold Finance is the new secured growth engine, with a ₹5,043 Cr book at Q2 FY27, up 242% YoY, and 502 branches. Personal Loans is the high-yield digital unsecured engine, with a ₹18,664 Cr book, up 72% YoY, and large partnerships at 40% of origination. Two-Wheeler and SME provide prime and near-prime retail assets, while Rural Business Finance is the large rural franchise that is being deliberately restrained because of monsoon uncertainty. Mortgages are the lower-yield secured anchor. The mix matters because management uses Gold and Micro LAP to offset the lower-yield mortgage book and the slower-growing rural book. NIMs+fees are being held at 10-10.5% through this mix, not through a single product yield increase.
The AI stack is the central operating mechanism. Project Cyclops is the underwriting engine. In Two-Wheeler it covers almost 90% of the base and the 90+ rate is around 1.2%; the earlier 10-month 30+ DPD comparison was 2.8% versus 7.1% for the industry. In SME, management says bounce rates on the Cyclops-covered book are about half those of the legacy book. Project Nostradamus is the portfolio-monitoring layer, being implemented in RBF after Two-Wheeler and Personal Loans. Project Hercules is the cross-sell layer; Phase 1 is live, gold-loan cross-sell lead conversion was above 14% in early October, and the full agentic workflow is targeted for early December 2026. The economic case is straightforward: better selection lowers credit cost, earlier warnings lower collection cost, and cross-sell lowers acquisition cost. The evidence is strongest in Two-Wheeler, thinner in Personal Loans and RBF, and that is why the Q4 FY27 credit-cost target remains a test rather than a proven outcome.
Capital allocation is now focused on distribution and software rather than acquisitions. Gold Finance added 172 branches in H1 FY27 and management expects about 800 branches by FY27-end, with a plan to scale to 2,000 branches. That is a company-funded rollout, and the economic test is branch productivity and cross-sell, not just branch count. Management targets Gold Loan AUM at 14-15% of total AUM, or ₹40,000-45,000 Cr, by FY31. The technology spend is modest in absolute terms: about ₹102 Cr was capitalised across Cyclops, Nostradamus and related projects by Q1 FY27. The private cloud is a longer-dated opex lever, with a 70% five-year cost-saving estimate versus hyperscale cloud. The payments build is strategic but has no disclosed capex or revenue target; the UPI stack is in closed-user testing and prepaid products are targeted for FY28 subject to a PPI licence.
The earnings path is guide-dependent. Management has explicitly guided to a 2-2.2% credit-cost corridor by Q4 FY27 and a 2.8-3% RoA corridor at Q4 FY27 exit. It has also maintained the Lakshya 2031 targets of 20%+ book CAGR, credit cost of 2% or less, RoA of 3-3.2% and RoE of 16-18% by FY31. The bridge management has described is about 20 bps of RoA from the ARC drag disappearing and 30-40 bps from credit and collections efficiency over the plan. These are guidance, not achieved results. The analyst inference is that if credit cost reaches 2-2.2% and NIMs+fees hold at 10-10.5%, RoA can approach 2.8%, but the timing depends on the unsecured book seasoning, rural collections and gold-loan branch productivity.
Management delivery is mixed. It delivered Karnataka normalisation, the gold-branch build and most technology rollout milestones. But FY26 credit cost came in at 2.64% against an earlier 2.3-2.5% target, and Q4 FY26 RoA was 2.40% against an earlier 'around 2.5%' checkpoint. Q2 FY27 credit cost of 2.46% is progress, but still 26-46 bps above the Q4 FY27 corridor. The company has not yet proven that the AI stack lowers credit cost across a full cycle in Personal Loans and RBF, and the IRDAI insurance-commission consultation is an unquantified fee-income risk.
The falsifiers are specific. Credit cost staying above 2.2% in Q4 FY27 would delay the RoA bridge. NIMs+fees falling below 10% on funding costs would squeeze the spread. A weak monsoon spreading beyond the 3.3% drought-affected RBF exposure would force further rural restraint. Gold branch expansion failing to produce productivity, or new RBI gold-loan rules compressing yields, would weaken the secured mix shift. An adverse IRDAI commission outcome before payments and Hercules scale would hit fee income. A PPI licence delay would defer the payments fee path. Conversely, delivery on the Q4 FY27 credit-cost and RoA corridors would validate the compounder case and make the FY31 targets more credible. On current evidence the business is improving, but the thesis is still a show-me compounder rather than a proven one.
companyname: L&T Finance Limited ticker: LTF sector: Non-Banking Financial Company (NBFC) - Retail Finance L&T Finance is a Mumbai-headquartered retail finance NBFC within the Larsen & Toubro Group, which holds a 65.99% stake. Incorporated in 1994 and listed on both exchanges, it is classified as an Upper Layer NBFC by the RBI. The company operates through 2,841 branches and 14,000+ distribution touchpoints across ~2 lakh villages and 450+ cities, serving a base of 2.8 Cr+ customers. The compa...
Read the full report →geographic expansion
Management reiterated a 2-2.2% credit-cost corridor by Q4 FY27 and a 2.8-3% RoA corridor at Q4 FY27 exit, with NIMs+fees maintained in the 10-10.5% range. Gold Finance is expected to reach about 800 branches by FY27-end, with a plan to scale to about 2,000 branches and Gold Loan AUM targeted at 14-15% of total AUM, or ₹40,000-45,000 Cr, by FY31. Project Hercules is targeted for full implementation by early December 2026, prepaid payment products are targeted for FY28 subject to a PPI licence, and Lakshya 2031 targets remain 20%+ book CAGR, credit cost of 2% or less, RoA of 3-3.2% and RoE of 16-18% by FY31.
Guidance maintainedmixed
Get valuation models, detailed research reports, thematic primers, one-pagers, risk analysis, growth triggers, bear case, capex tracker, walk the talk, and more for L&T Finance Ltd and 4,900+ companies.
5-day free pass. No card required.