India Thematic: Three Currents We Are Tracking
AI, Financials, CDMO
Sowilo Research | 21-Jul-26
We flag three currents crafted by conviction. First, India’s CDMO base is re-rating from cyclical contract manufacturing to a recurring ‘Pharma as a Service’ model. Second, the financial sector sits at a crossroads: regulatory tailwinds now counter macro headwinds even as the microfinance cycle turns up. Third, the global AI trade has become the most concentrated position markets have ever held, leaving India lightly owned but building fast. We frame each below and flag what we are watching.
1. CDMO upgrades to ‘Pharma as a Service’: a $21 bn India opportunity by FY30
Global CDMO markets are shifting toward Asia even as China’s share saturates after roughly a decade of high double-digit growth. On a bottoms-up read of global biotech pipelines and the disease funnel, we see Indian biopharma pivoting from contract manufacturing to a software-style services model, a Pharma-as-a-Service (PaaS) layer incubated separately over the short to medium term, which could unlock about $21 bn for Indian biopharma by FY30, set against an increasingly complex mix of modalities and a higher regulatory bar on quality, environment and CRDMO scrutiny beyond the US FDA.
| $21 bn | PAAS OPPORTUNITY |
|---|---|
| Indian biopharma by FY30 |
| Post-2028 | BIOSECURE EFFECTS |
|---|---|
| lucrative despite dilution |
| 3-4 yrs | INDIA FERMENTATION |
|---|---|
| capacity doubling cycle |
| 2028-29 | CHINA+1 ACCELERATION |
|---|---|
| India a natural magnet |
| Driver | What is happening | Our read |
|---|---|---|
| Pool shift | CRDMO markets shifting to India and Korea as China saturates; medium-term acceleration expected | A multi-year, structural reallocation, not a cyclical swing |
| BIOSECURE | Passed the US Senate early 2026; effects visible post-2028 and still lucrative despite a watered-down text | A slow-burn tailwind that de-risks the FY29+ contract pipeline |
| India’s edge | Fermentation doubling every 3-4 years, a conducive Bioshakti policy, and Middle East geopolitics levelling price versus Korea | Cost and capacity are now competitive enough to win discovery-pipeline partnerships in FY27-28 |
| New tokens of service | High-potency APIs and next-gen GLP-1 as new categories; organised patient pools and faster trial approvals aid discovery work | The service menu broadens from molecules to full development capability |
Source: Bernstein ‘Pharma as a Service’; Sowilo estimates.
Our view: we view CDMO as the most durable and longest-dated of the three themes: the PaaS framing re-rates the model from cyclical contract manufacturing toward a recurring, capability-led service layer. The forcing functions line up, BIOSECURE post-2028, China+1 from 2028-29, fermentation doubling and a supportive Bioshakti policy, and India’s cost structure is now level with Korea. We would position ahead of the FY27-28 discovery-partnership wins that are the leading indicator of the $21 bn pool converting, with cell and gene therapies and ADCs as un-underwritten upside beyond 2030.
2. Financials at a crossroads: macro headwinds, regulatory tailwinds, MFI recovery
FY26 saw a widening divergence inside the financial sector: public-sector banks, small and mid-cap private banks and second-line NBFCs outperformed, while large private banks and large NBFCs lagged on growth, margin and management niggles. Into FY27, West Asia conflict and El Niño warrant caution, but recent FCNR (B) and ECLGS measures provide countercyclical support to growth, margins and asset quality. We prefer banks and within banks, private over public, better placed to capture the regulatory tailwinds and endure the coming ECL transition while still delivering superior RoAs.
| 14-15% | SYSTEM CREDIT GROWTH FY27E |
|---|---|
| PVBs to take the lead |
| 1.8-2.1% | LARGE PVB ROA |
|---|---|
| as margins, credit costs ease |
| ₹2.55L Cr | ECLGS 5.0 ENVISAGED |
|---|---|
| ~1.2% of outstanding credit |
| Decadal best | FY26 ASSET QUALITY |
|---|---|
| delinquencies stabilising |
Source: Nuvama Institutional Equities; RBI; Sowilo.
| Segment | FY26 read | FY27E stance |
|---|---|---|
| Large private banks | Lagged on leadership, margin and management overhangs; EBLR books hit by rate cuts | Set to bounce back: above-system growth, share gains from PSBs, RoA 1.8-2.1%; FCNR relief aids deposit-strained lenders |
| SMID private banks | Re-rating underway on capital augmentation and leadership strengthening | Further legs: easing unsecured/MFI credit costs and operating leverage lift RoAs above 1-1.5% |
| Public-sector banks | Notable outperformers on share gains and ~1% RoA delivery | Supportive near term, but ECL impact, elevated G-Sec yields and consolidation could stall the rally |
| NBFCs / fintechs | NBFCs outpaced banks on growth and turnarounds; fintechs pivoting to monetisation | NBFC easy returns behind as funding-cost tailwind fades; prefer proven second-line turnarounds and profitable, diversified fintechs |
Source: Nuvama Institutional Equities; Sowilo.
The MFI drag of the past two years, driven by MFIN guardrails from Aug-24, is largely behind. CRISIL expects NBFC-MFI AUM to surge about 20% this fiscal against 4% last, though the pure microfinance book grows a more measured ~13%, with diversification into secured and non-microfinance lending doing the heavy lifting and portfolio at risk over 90 days now below 1%. The recovery is quality-led, not volume-led, and the industry is unlikely to revisit its FY24 peak AUM in the near term. Separately, ECLGS 5.0 envisages at least ₹2.55L Cr of working-capital disbursement (versus ₹3.7L Cr cumulatively over versions 1.0 to 3.0), a credit booster and asset-quality buffer mainly for banks.
| Metric | Trend | Detail |
|---|---|---|
| NBFC-MFI AUM growth | ~20% this fiscal vs 4% last | Micro-finance book ~13%; non-MFI (secured/retail) the primary driver |
| Portfolio mix | Non-MFI book 6% to 14% YoY | Projected ~18% of AUM by end of this fiscal |
| Borrower quality | Repeat borrowers 53% to ~66% (Mar-26) | Average ticket up ~15% to ~₹59,000; wider CGFMU cover on new loans |
| Regulatory | Qualifying-assets floor cut to 60% from 75% (Jun-25) | Growth led by NBFCs as banks lean to individual over JLG loans |
Source: CRISIL Ratings press release (Jul-26); Nuvama Institutional Equities; Sowilo estimates.
Our view: The MFI recovery is real but is a normalisation and diversification story: the durable growth is coming from the secured and individual-loan pivot, not a return to the JLG peak, and we prefer lenders that carried collection and underwriting discipline through the down-cycle.
3. AI: the most concentrated bet in market history, and India is barely in it
The AI trade abroad has become the largest and most concentrated position global markets have ever held, and it is priced for perfection: the Magnificent 7 now account for roughly a third of the entire S&P 500, so a passive world index is in effect a leveraged bet on a handful of US AI names. The four largest hyperscalers will spend about $725 bn on AI this year, up around 77%, even as the frontier itself accelerates: models now land every few weeks not once a year, agents run whole jobs, and the makers now call the frontier too powerful to release freely. India is lightly owned but building the compute, power and data-center layer fast.
| $725 bn | HYPERSCALER AI CAPEX |
|---|---|
| Up ~77% this year |
| ~1/3 | MAG 7 SHARE OF S&P 500 |
|---|---|
| world index = US AI bet |
| ~$5.0 tn | LARGEST CHIPMAKER |
|---|---|
| 8% of the S&P 500 alone |
| 45,000+ | GPUS IN INDIA POOL |
|---|---|
| public AI compute |
Source: Company filings; BIS; Micron; NSE / IndiaAI; CNBC and Reuters (as at Jun-26); Sowilo.
| Signal | What the data shows | Our read |
|---|---|---|
| Concentration | A third of the S&P 500 sits in seven names; the largest chipmaker alone is 8% of the index; largest-ever private raises now rival the biggest listings | Index-level AI exposure is now a factor bet, not diversification |
| The first doubts | On 28-Jun-26 the BIS called the boom ‘increasingly leveraged’ and circular, a 2008-like risk; prominent investors invoke 1999; memory prices have tripled | The debate has shifted from adoption to financing and returns on capital |
| From talking to doing | Four frontier models in Apr-Jun 26; ~75% of one hyperscaler’s new code is AI-written (70-90% at a leading lab); the most capable models now gated on safety grounds | The productivity claim is measurable output now, a tailwind for services demand |
| India is building | A domestic conglomerate leased 168 MW of data-centre capacity, a global cloud major added $13 bn, and corporates plus the government keep committing tens of billions | The domestic expression is compute, power, and data centers, not the crowded US mega-caps |
Source: Company filings; BIS; NSE / IndiaAI; Anthropic; OpenAI; Google; CNBC and Reuters; Sowilo.
Our view: we view AI as a structural theme where the risk and the opportunity sit in different places. The leadership abroad is priced for perfection and is where the BIS leverage warning bites; the Indian expression, still lightly owned, is the physical and services layer, data centres, power, compute and the IT services franchises that monetise adoption. We would rather own the rotation into an underowned market than the crowded trade that must keep beating a rising bar, while watching AI-led pricing deflation on the services side.
4. What to watch
i. CDMO leading indicators: discovery-pipeline partnership wins in FY27-28, and BIOSECURE / China+1 milestones into 2028-29, as the early tells that the $21 bn PaaS pool is converting.
ii. ECLGS 5.0 disbursement pace: progress toward the ₹2.55L Cr envelope is both a credit booster and an MSME asset-quality buffer, mainly for banks.
iii. FCNR (B) and ECB flows: the margin swing factor for deposit-strained private banks; watch Q2 FY27 liquidity and any mortgage-rate cuts that follow.
iv. MFI durability: whether ~13% microfinance-book growth and sub-1% PAR hold as recoveries mature, without a return to the FY24 peak.
v. AI capex and the rotation: whether the $725 bn build-out sustains after the BIS leverage warning, the swing factor for global AI and for capital rotating into an underowned India.
Regards,
Team Sowilo
DISCLAIMER
Any information contained in this material represents Sowilo’s views and research analysis and shall not be deemed to constitute an advice, an offer to sell purchase or as an invitation or solicitation to do for security of any entity and further Sowilo Investment Managers LLP and its employees/directors shall not be liable for any loss, damage, liability whatsoever for any direct or indirect loss arising from the use of this information.
Sowilo Investment Managers LLP – SEBI Registered Portfolio Manager (INPO00008127).
