XP took in R$28bn in a quarter when client assets rose only R$6bn

XP Inc. reported R$28 billion of net inflow in the second quarter of 2026. Yet total client assets rose by only R$6 billion, from R$1.529 trillion at the end of March to R$1.535 trillion at the end of June. The difference implies that roughly R$22 billion of market movements, foreign-exchange effects, classification changes or other valuation factors offset new money during the quarter.[1]

The residual is not evidence that client money disappeared. Net inflow records transfers into the platform less transfers out; closing client assets also move with the prices of securities and funds already held, as well as with the measurement boundary. XP reports all three numbers rounded to the nearest R$1 billion. Allowing for that rounding, the unexplained offset lies approximately between R$20.5 billion and R$23.5 billion. The central estimate is therefore robust even though it is not an exact ledger balance.[1][2]

The quarterly headline combines a flow and a stock

A flow measures activity across a period. A stock measures the value present at a particular date. Starting with R$1.529 trillion, adding the reported R$28 billion inflow would have produced R$1.557 trillion if nothing else changed. The actual closing balance was R$1.535 trillion. Subtracting the expected balance from the reported balance produces the approximate negative R$22 billion residual.[1]

That calculation is a reconciliation, not a claim about the cause. XP does not publish a complete quarter-on-quarter bridge in the release. The company defines client assets as the aggregate value of assets it services across investment advice, self-directed brokerage, financial planning and other models. Its annual filing says the total varies with transfers and with fluctuations in security prices and fund net asset values. Any instrument denominated in another currency can also change in reais even when its local price and the number of units held are unchanged.[1][2]

The quarter’s asset result therefore tells a different story from the inflow result. Commercially, XP gathered substantially more net new money than in the comparable period: R$28 billion against R$10 billion in the second quarter of 2025. Economically, the value of the total asset base was almost unchanged from March. Both can be true at once. A platform can win deposits during an adverse market period, just as it can report rising assets during net withdrawals if markets rise enough.[1]

The annual comparison supplies the bridge the quarter lacks

XP did provide a year-on-year decomposition. Total client assets increased from R$1.372 trillion in June 2025 to R$1.535 trillion in June 2026, a gain of R$163 billion. The company attributed that increase to R$103 billion of net inflow and R$60 billion of market appreciation. Those two components add exactly to the reported change before considering rounding.[1]

This longer window is useful because it shows that the second-quarter offset did not erase the broader trend. About 63.2% of the R$163 billion annual increase came from net inflow and 36.8% from market appreciation. But it also demonstrates why the same percentage growth in client assets can arise from very different sources. Money gathered from clients can be more persistent and commercially informative than a market revaluation, while market gains can still increase fee-bearing balances without any new account transfer.[1]

The release does not identify whether the R$103 billion year-on-year inflow was concentrated in retail, corporate or institutional relationships. It says retail net inflow in the second quarter was R$20 billion, leaving R$8 billion between retail and the reported total for that quarter. That arithmetic should not be treated as a clean institutional-flow measure without confirmation, because the total may include categories, eliminations or definitions that are not fully reproduced in the headline table.[1]

A change in the asset boundary complicates comparisons

XP disclosed in 2025 that it had begun including institutional client assets in total client assets and restated the series back to the first quarter of 2024. The current total is therefore broader than the older version of the metric. This does not invalidate the 2026 quarterly comparison, because both March and June figures use the current basis, but it matters when readers compare newer releases with historical figures copied from older presentations or databases that may not have been restated.[2]

The annual report’s definition is deliberately broad: client assets are assets serviced by XP across different service models. That is not the same as assets under management, where XP itself makes portfolio decisions, nor is it the same as client cash on XP’s balance sheet. A brokerage platform can service securities held in custody, distribute third-party funds, advise on portfolios and manage some products directly. Combining them is useful for measuring commercial reach, but it should not be read as a single accounting liability or a pool wholly controlled by XP.[2]

The denominator also prevents a tempting but misleading calculation. Dividing R$1.535 trillion by the 4.772 million active-client count would suggest about R$322,000 per active client. That ratio is not like for like. Total client assets include institutional assets, while active clients are defined as retail accounts across XP Investimentos, Rico, Clear, XP Investments and XP Private Europe. A person with qualifying accounts in more than one brand can also be counted once for each account.[1][2]

Client growth and trading activity moved differently

XP’s active-client count fell from 4.790 million in March to 4.772 million in June, while remaining 1% above the prior year. Retail daily average trades fell 13% sequentially, from 2.7 million to 2.4 million, but were 3% higher than a year earlier. Meanwhile, total advisers increased slightly to 18,400. These movements show why a single asset headline cannot describe engagement, acquisition and monetisation at the same time.[1]

A small decline in active accounts does not establish that 18,000 individual people left, because the metric counts qualifying accounts and can count one client across multiple brands. Nor does lower trading activity necessarily imply weaker relationships: clients may hold fixed income, funds, pensions or cash without placing frequent equity or futures trades. XP’s annualised retail take rate rose two basis points sequentially to 1.20% but remained five basis points below the prior year, adding another separate measure of how balances converted into revenue.[1]

The missing quarterly bridge is the material disclosure

The key unanswered question is what produced the approximate R$22 billion offset. A complete bridge would separate market-price changes, currency translation, acquisitions or disposals, reclassifications, institutional-custody changes and rounding. It would also distinguish retail from institutional assets on the same basis as the net-flow figures. Without that breakdown, investors can calculate the combined residual but cannot identify its components.[1][2]

The stock-flow bridge is a residual, not a record of withdrawals

The arithmetic is attractive because it is simple: beginning customer assets of R$1.529 trillion plus R$28 billion of net inflows would, absent other changes, imply R$1.557 trillion. Reported ending assets were approximately R$1.535 trillion. The unexplained difference is therefore negative R$22 billion. But an asset balance measured at period-end is not a bank deposit account into which every new real of net inflow must remain. Security-price movements, distributions, foreign exchange and consolidation rules can change the mark-to-market value even when no customer withdraws funds.[1][2]

That residual is about 1.4% of beginning assets and about 79% of the reported net inflow; neither ratio proves a specific market loss or operational failure. The company would have to disclose the detailed market, FX and perimeter movement schedule to attribute it precisely. The practical follow-up is whether the firm can convert inflow into stable fee-bearing assets across more than one market cycle, not whether the closing total rose by the full amount of a single quarter’s deposits.[1][2]

There are benign explanations. Brazilian and international asset prices can move sharply within a quarter; exchange rates can alter the real value of foreign holdings; and a broad servicing metric can change when accounts or products are reclassified. A less favourable explanation would be that inflows went into lower-value or lower-revenue assets while existing balances declined. The current disclosure does not allow those possibilities to be ranked.[1][2]

The next earnings release is scheduled for after the end of this reporting window, so the second-quarter figures remain XP’s latest published operating results as of 8 November. The most useful test will be whether third-quarter asset growth again diverges materially from net inflow and whether XP supplies a bridge. For now, its own numbers establish the central finding: a quarter of strong gathering produced almost no increase in the closing asset stock because other effects offset about four-fifths of the reported inflow.[1]

Report an error or submit an update