Reconcile Changes in a Company’s Cash Balance With Claude MCP
A cash-flow statement is supposed to connect two balance-sheet dates: opening cash, plus what came in and went out, equals closing cash. When the bridge is rebuilt from a data feed, it often fails to close, or closes on a different definition of cash from the one the company uses. That matters for any assessment of corporate financial health that starts from the cash line.
This walkthrough reconciles Uber Technologies' (UBER) cash movement for fiscal 2024 and fiscal 2025. Claude first builds the bridge from FMP's standardized cash-flow and balance-sheet records, then rebuilds it from the cash-flow statement and accounting-policy note in Uber's 10-K, and compares the two line by line.
Both filing-based bridges close. The standardized records follow a narrower definition: their opening and closing balances for these two years equal cash and cash equivalents plus current restricted cash, while Uber's filed cash-flow total also includes non-current restricted cash.
The output is a cash bridge for each year, from opening to closing cash, and a comparison table that records, for every difference, whether it equals a disclosed amount and whether the line items in the two sources show where that amount sits. An equal number and a confirmed mapping are reported separately. The workflow reconciles before it interprets. It does not score earnings quality, estimate cash runway or assess capital allocation.
Key takeaways
- The cash-flow statement's own caption defines what "cash" means for the bridge. Uber's includes restricted cash, both current and non-current.
- A bridge can close perfectly and still use a different definition. FMP's standardized records reconcile internally on a narrower total.
- A difference that equals a disclosed amount is a numerical reconciliation. It becomes a documented explanation only when the line items in both sources show where that amount sits.
- The net-increase subtotal is a check, never a further movement, and any additional reconciling line is identified from the filing before it enters the bridge.
- In this run, seven differences have line-item support, three only match a disclosed amount, and the two exchange-rate differences match no disclosed amount and stay Unresolved.
Building the cash reconciliation framework
Uber suits this exercise for three reasons. It holds material restricted cash in both current and non-current assets, it reports a sizable exchange-rate effect, and in both years it made a cash payment to redeem non-controlling interests that a standardized feed could place in a different section from the filing. Each of these is a place where a bridge can drift.
Defining cash
The definition comes from the cash-flow statement, not from the balance sheet. Uber's statement reconciles "cash and cash equivalents, and restricted cash and cash equivalents," and its accounting-policy note lists the balance-sheet lines that add up to that total. Restricted cash there includes amounts pledged as security for letters of credit and insurance collateral, and cash unavailable for immediate use because of legal or contractual limits. The balance sheet splits it between current and non-current assets by the term of the restriction. Standardized fields are built to a common template, as FMP's notes on how statements are parsed explain, so their cash total need not follow any one company's caption.
The bridge and its formulas
|
Measure |
Calculation |
Where it is used |
Purpose |
|
Calculated closing cash |
Opening cash + operating + investing + financing + exchange-rate effect + other reconciling items identified from the filing |
Tables 2 and 5 |
Rebuilds closing cash from its components |
|
Residual |
Reported closing cash − calculated closing cash |
Tables 2 and 5 |
Shows whether the bridge closes |
|
Net-change check |
The movement lines, compared with the reported net-increase subtotal |
Tables 2 and 5 |
Confirms the subtotal without adding it to the bridge |
|
Continuity check |
Opening cash of year t, compared with closing cash of year t−1 |
Table 2 |
Confirms consecutive years use the same definition |
|
Balance-sheet tie |
Cash and cash equivalents + restricted cash, current + restricted cash, non-current |
Tables 4 and 5 |
Matches closing cash to the balance-sheet lines of the same definition |
Additional reconciling items
Some cash-flow statements show lines between the section totals and the closing balance, such as cash reclassified to assets held for sale. FMP's standardized record has no field for such a line, so Prompt 1 tests only opening cash and the four movements. If that bridge leaves a residual, the residual is recorded as open, with no cause assigned, until the filing is read.
Prompt 2 identifies any additional item from the filing itself: a line in the reconciliation from opening to closing cash that is not one of the three section totals, the exchange-rate line or the net-increase subtotal. Each item found is added once to the calculated closing balance. For the net-change check, Prompt 2 tests whether the filing's net-increase subtotal includes the item or excludes it and reports which, so the item is never counted twice. Uber's statement shows no such line in either year, and the filing's bridge confirms it.
Signs and subtotals
Every value keeps the sign the source reports. Uber labels outflow sections "net cash used in" and shows them as negative numbers; FMP returns the same signs, so no conversion is needed. Reading "used in" as an instruction to subtract an already negative value would turn an outflow into an inflow, so the bridge adds every line as presented.
Evidence levels and status rules
A difference that equals a disclosed amount shows the two numbers agree. It does not by itself show that the amount explains the difference. The status separates the two cases.
|
Status |
Rule |
|
Reconciled |
The difference is $0 at whole-million precision |
|
Explained, mapping shown in line items |
The difference equals a disclosed amount, and the line items in both sources show where that amount sits: an FMP component field equals the filing line, and the filing's own line for that component does not account for it |
|
Arithmetic match, mapping not established |
The difference equals a disclosed amount, but no line item in either source shows where the amount sits or why |
|
Unresolved |
No disclosed amount equals the difference; the line is marked Review Required and no balancing line is added |
Where FMP reports a total with no component fields, as with its cash balances, the line-item test can compare only that total with a sum of filed lines; the comparison table states this wherever it applies, because it is the weakest form of line-item evidence. Neither status establishes why FMP builds a field the way it does. Those sources do not document FMP's methodology, so the article reports where differences sit, not why. A missing value is reported as Unavailable after one retry, and a statement line that does not exist is recorded as None disclosed, which is different from a reported zero. These labels are screening rules set for this workflow.
FMP data inputs and access through Claude MCP
The standardized values come from the Cash Flow Statement API and the Balance Sheet Statement API, both part of FMP's financial statements dataset. The filing values come from the Financial Reports Form 10-K JSON API, which returns the 10-K's statements and notes as structured data.
|
Claude connector tool / operation |
Parameters |
Fields used |
Feeds |
|
statements / cashflow-statement |
symbol=UBER, period=annual, limit=3 |
Prompt 1: cashAtBeginningOfPeriod, the three section totals, effectOfForexChangesOnCash, netChangeInCash, cashAtEndOfPeriod, filingDate. Prompt 2: the investing and financing component fields |
Tables 1, 2 and 6 |
|
statements / balance-sheet-statement |
symbol=UBER, period=annual, limit=3 |
cashAndCashEquivalents |
Table 3 |
|
statements / financial-reports-form-10-k-json |
symbol=UBER, year=2025, period=FY |
Cash-flow statement, balance sheet, Note 1 cash reconciliation |
Tables 4 to 6 |
|
statements / financial-reports-form-10-k-json |
symbol=UBER, year=2024, period=FY |
Fiscal 2024 cash-flow lines as originally filed |
Revision check |
The filing values are in the consolidated statements of cash flows and consolidated balance sheets in Item 8 of Uber's fiscal 2025 Form 10-K, and in Note 1, Description of Business and Summary of Significant Accounting Policies, under restricted cash and cash equivalents. The original fiscal 2024 figures are in the fiscal 2024 Form 10-K.
All values are in USD millions and cover Uber's consolidated results. Fiscal years end December 31, and filing dates are kept apart from period-end dates; when a filing became public is a separate question, covered in the as-of-date fundamental screen workflow. Uber's 2025 acquisitions appear in the investing totals, but their consideration is not analyzed here.
Accessing the data through Claude
Add the FMP MCP server under Claude's Settings, Connectors, Add custom connector, with this URL and your key in place of YOUR_FMP_API_KEY:
https://financialmodelingprep.com/mcp?apikey=YOUR_FMP_API_KEY
Every source comes through FMP, so both prompts run in one new conversation with web search off.
The run shown below took place on October 23, 2026, in a single Claude conversation completed by 03:17 UTC.
Running the reconciliation in Claude
The workflow uses two prompts. Prompt 1 works from FMP's standardized records alone and reports every difference without explaining it. Prompt 2 brings in the 10-K, the only source that defines Uber's cash, and tests each difference against it. Values are in USD millions with signs as reported. A later run applies the same rules, but FMP's records can change, so its tables can differ.
Prompt 1: Build the bridge from FMP's standardized records
|
Using Financial Modeling Prep through the connected FMP MCP server, build a cash bridge for Uber Technologies (UBER) for fiscal 2024 and fiscal 2025. Objective Check whether FMP's standardized cash-flow records connect opening cash to closing cash, and whether closing cash matches the balance-sheet cash field. Do not explain differences yet. Data rules - Use only the FMP statements tool (endpoints cashflow-statement and balance-sheet-statement). Do not use web search or filings in this step. - Keep every value with the sign FMP returns. Do not flip signs. - Retry a failed request once; report Unavailable if a value is still missing. Never treat a missing value as zero. Steps 1. Call cashflow-statement and balance-sheet-statement for UBER, period annual, limit 3, so fiscal 2023 is available as the prior period. If fiscal 2023 is not returned, raise the limit until it is. 2. From each cash-flow record keep cashAtBeginningOfPeriod, netCashProvidedByOperatingActivities, netCashProvidedByInvestingActivities, netCashProvidedByFinancingActivities, effectOfForexChangesOnCash, netChangeInCash, cashAtEndOfPeriod and filingDate. 3. For fiscal 2024 and 2025: Calculated closing = opening + operating + investing + financing + exchange-rate effect. These four movements are the only ones FMP's record provides; any other reconciling line can be identified only from the filing in a later step. Do not add netChangeInCash as a separate line; use it only to check the sum of the four movements. 4. Residual = reported closing minus calculated closing. If a residual is not zero, report it as Open and do not assign a cause. 5. Continuity: compare each year's opening cash with the prior year's closing cash. 6. Compare each year's cashAtEndOfPeriod with the balance sheet's cashAndCashEquivalents for the same date. 7. Show values in USD millions. Output Start with a short run summary: run date, company, periods, operations called and any data limitation. Table 1: Fiscal year | Opening cash | Operating | Investing | Financing | Exchange-rate effect | Net change | Closing cash | Filing date Table 2: Fiscal year | Calculated closing | Reported closing | Residual | Net-change check | Continuity with prior closing Table 3: Fiscal year | Cash-flow closing cash | Balance-sheet cashAndCashEquivalents | Difference After the tables, list each non-zero difference in one line. Do not explain causes. |
What Prompt 1 returned
Claude's run summary recorded the run date (October 23, 2026), the two operations called with limit 3, both on the first attempt with fiscal 2023 included, and no data limitations: every requested field was present. It converted FMP's whole-dollar values to USD millions and kept FMP's signs.
Table 1. FMP standardized cash-flow records for UBER (USD millions)
|
Fiscal year |
Opening cash |
Operating |
Investing |
Financing |
Exchange-rate effect |
Net change |
Closing cash |
Filing date |
|
FY2023 (prior period) |
6,677 |
3,585 |
-3,226 |
-95 |
63 |
327 |
7,004 |
2024-02-15 |
|
FY2024 |
5,485 |
7,137 |
-4,028 |
-1,236 |
-920 |
953 |
6,438 |
2025-02-14 |
|
FY2025 |
6,438 |
10,099 |
-3,673 |
-5,604 |
476 |
1,298 |
7,736 |
2026-02-13 |
Table 2. Bridge checks on FMP records (USD millions)
|
Fiscal year |
Calculated closing |
Reported closing |
Residual |
Net-change check |
Continuity with prior closing |
|
FY2024 |
6,438 |
6,438 |
0 |
Sum of four movements 953 = netChangeInCash 953 |
Opening 5,485 vs FY2023 closing 7,004: difference -1,519, Open |
|
FY2025 |
7,736 |
7,736 |
0 |
Sum of four movements 1,298 = netChangeInCash 1,298 |
Opening 6,438 vs FY2024 closing 6,438: difference 0 |
Each FMP record adds up on its own terms, so no residual is open. The series still breaks between fiscal 2023 and fiscal 2024, which an arithmetic check inside one record cannot reveal.
Table 3. FMP cash-flow closing cash and balance-sheet cash (USD millions)
|
Fiscal year |
Cash-flow closing cash |
Balance-sheet cashAndCashEquivalents |
Difference |
|
FY2024 |
6,438 |
6,438 |
0 |
|
FY2025 |
7,736 |
7,105 |
631, Open |
Claude listed two non-zero differences, the fiscal 2024 continuity break of -1,519 and the fiscal 2025 gap of 631 against the balance sheet, both Open with no cause assigned. It also recorded a related observation outside the two target years: FMP's fiscal 2023 cash-flow closing of 7,004 is 1,519 above its fiscal 2023 balance-sheet cash of 5,485, the same amount as the continuity break.
Prompt 2: Rebuild the bridge on the filing's definition and test each difference
Prompt 2 reads Uber's 10-Ks through FMP, identifies any additional reconciling line, rebuilds the bridge on the filing's definition, and tests every difference at two levels: whether a disclosed amount equals it, and whether the line items show where that amount sits.
|
Continue the UBER cash reconciliation using the fiscal 2025 Form 10-K. Objective Establish the cash definition the cash-flow statement uses, rebuild the bridge on that definition, and test each difference from Tables 1 to 3 against disclosed evidence. Data rules - Use only the FMP statements tool: endpoint financial-reports-form-10-k-json, symbol UBER, year 2025, period FY, and for the revision check year 2024; and endpoint cashflow-statement, symbol UBER, period annual, limit 3, for FMP's investing and financing component fields. Do not use web search. - Keep signs as presented. Do not count the net-increase subtotal in addition to its components. - If a difference equals no disclosed amount, mark it Unresolved. Do not add a balancing line. Steps 1. From the cash-flow statement, record the cash caption and the opening, operating, investing, financing, exchange-rate, net-increase and closing lines for fiscal 2024 and 2025, and the individual investing and financing lines. 2. Additional reconciling items: list any line in the reconciliation from opening to closing cash that is not one of the three section totals, the exchange-rate line or the net-increase subtotal. Record None disclosed if there is none. 3. From the accounting-policy note, record the reconciliation of balance-sheet cash lines to the cash-flow total for each year-end. 4. Rebuild the bridge for each year: opening + three section totals + exchange-rate line + each additional item from step 2, added once. Residual = reported closing minus calculated closing. Net-change check: test whether the net-increase subtotal equals the movements with the additional items or without them, and report which. 5. Compare the cash-flow closing balance with the balance-sheet lines that make up the same definition. 6. Compare each FMP total with the filing value: opening cash, operating, investing, financing, exchange-rate effect and closing cash for fiscal 2024 and 2025, and FMP's balance-sheet cashAndCashEquivalents with the filed cash and cash equivalents line at December 31, 2023, 2024 and 2025. For each difference: a. Give the disclosed amount that equals it, or mark it Unresolved. b. Line-item test. Mapping shown if (i) an FMP component field equals the disclosed amount and the filing's line for that same component does not include it, or (ii) every other FMP component in the section equals a filing line or the sum of filing lines of the same kind, and the disclosed amount is the only filing line with no FMP counterpart. Otherwise the mapping is not established. An equal amount on its own never establishes the mapping. c. Status: Reconciled if the difference is 0; Explained, mapping shown in line items; Arithmetic match, mapping not established; or Unresolved. State the line items used for b. Do not state why FMP builds a field the way it does. 7. Revision check: compare the fiscal 2024 lines in the 2025 10-K with the original 2024 10-K. Output Table 4: Year-end | Cash and cash equivalents | Restricted cash, current | Restricted cash, non-current | Total per cash-flow statement Table 5: Line | Fiscal 2024 | Fiscal 2025 | Source Table 6: Fiscal year | Line | FMP value | Filing value | Difference | Disclosed amount that equals it | Line-item evidence | Status Then give the additional-items result and the revision-check result, one line each. |
What Prompt 2 returned
Claude called the 10-K operation for fiscal 2025 and fiscal 2024, both on the first attempt, and reused the cashflow-statement response from Prompt 1 for FMP's component fields. It used the consolidated statements of cash flows, the consolidated balance sheets and the restricted cash table in Note 1, and noted that the 10-K operation returns no filing URL or page numbers. The cash-flow statement's caption is "Cash and cash equivalents, and restricted cash and cash equivalents," and Note 1 reconciles the balance sheet to it.
Table 4. Uber's cash definition, from Note 1 (USD millions)
|
Year-end |
Cash and cash equivalents |
Restricted cash, current |
Restricted cash, non-current |
Total per cash-flow statement |
|
Dec 31, 2023 |
4,680 |
805 |
1,519 |
7,004 |
|
Dec 31, 2024 |
5,893 |
545 |
2,172 |
8,610 |
|
Dec 31, 2025 |
7,105 |
631 |
1,911 |
9,647 |
Each total equals the three balance-sheet lines for that year-end and the cash-flow closing balance, with no difference in any year.
Table 5. Uber cash bridge from the filing, fiscal 2024 and fiscal 2025 (USD millions)
|
Line |
Fiscal 2024 |
Fiscal 2025 |
Source |
|
Beginning of period (cash, cash equivalents and restricted cash) |
7,004 |
8,610 |
10-K cash-flow statement |
|
Net cash provided by operating activities |
7,137 |
10,099 |
10-K cash-flow statement |
|
Purchases of property and equipment |
-242 |
-336 |
10-K cash-flow statement |
|
Purchases of non-marketable equity securities |
-289 |
-676 |
10-K cash-flow statement |
|
Purchases of marketable securities |
-12,765 |
-21,447 |
10-K cash-flow statement |
|
Proceeds from maturities and sales of marketable securities |
10,204 |
20,046 |
10-K cash-flow statement |
|
Proceeds from sale of equity method investments |
17 |
0 |
10-K cash-flow statement |
|
Acquisition of businesses, net of cash acquired |
0 |
-815 |
10-K cash-flow statement |
|
Other investing activities |
-102 |
-336 |
10-K cash-flow statement |
|
Net cash used in investing activities |
-3,177 |
-3,564 |
10-K cash-flow statement |
|
Issuance of term loan and notes, net of issuance costs |
3,972 |
3,359 |
10-K cash-flow statement |
|
Principal repayment on term loan and notes |
-3,986 |
-2,350 |
10-K cash-flow statement |
|
Principal payments on finance leases |
-172 |
-157 |
10-K cash-flow statement |
|
Proceeds from issuance of common stock under ESPP |
156 |
183 |
10-K cash-flow statement |
|
Repurchases of common stock |
-1,252 |
-6,523 |
10-K cash-flow statement |
|
Redemption of non-controlling interests |
-851 |
-109 |
10-K cash-flow statement |
|
Other financing activities |
46 |
-116 |
10-K cash-flow statement |
|
Net cash used in financing activities |
-2,087 |
-5,713 |
10-K cash-flow statement |
|
Effect of exchange rate changes |
-267 |
215 |
10-K cash-flow statement |
|
Net increase (subtotal, not added) |
1,606 |
1,037 |
10-K cash-flow statement |
|
End of period |
8,610 |
9,647 |
10-K cash-flow statement |
|
Calculated closing (opening + three sections + exchange-rate effect) |
8,610 |
9,647 |
Calculated |
|
Residual (reported minus calculated) |
0 |
0 |
Calculated |
|
Net-change check |
1,606 = movements without additional items |
1,037 = movements without additional items |
Calculated |
For fiscal 2025: 8,610 + 10,099 − 3,564 − 5,713 + 215 = 9,647, the reported closing balance, and 7,105 + 631 + 1,911 = 9,647 on the balance sheet. Claude found no additional reconciling item in either year: the filed reconciliation runs from opening cash through the three section totals and the exchange-rate line to the net-increase subtotal and closing cash.
Table 6. FMP values compared with the filing (USD millions)
|
Fiscal year |
Line |
FMP value |
Filing value |
Difference |
Disclosed amount that equals it |
Line-item evidence |
Status |
|
2024 |
Opening cash |
5,485 |
7,004 |
-1,519 |
Restricted cash, non-current, Dec 31, 2023: 1,519 |
Test (ii): FMP 5,485 = cash 4,680 + restricted current 805; the 1,519 line is the only policy-table line with no FMP counterpart |
Explained, mapping shown in line items |
|
2024 |
Operating |
7,137 |
7,137 |
0 |
Not applicable |
Not applicable |
Reconciled |
|
2024 |
Investing |
-4,028 |
-3,177 |
-851 |
Redemption of non-controlling interests (financing): -851 |
Test (i): FMP acquisitionsNet -851, while the filing's acquisition line is 0. Test (ii) also holds: capex, purchases (-12,765 + -289), sales (10,204 + 17) and other investing all match |
Explained, mapping shown in line items |
|
2024 |
Financing |
-1,236 |
-2,087 |
+851 |
Same -851 line |
Test (ii): netDebtIssuance -186 = 3,972 − 3,986 − 172; issuance 156; repurchases -1,252; other 46. The redemption is the only filing line with no FMP counterpart |
Explained, mapping shown in line items |
|
2024 |
Exchange-rate effect |
-920 |
-267 |
-653 |
None; 653 appears in neither 10-K |
Calculated only: equals the change in non-current restricted cash (2,172 − 1,519), which is not a disclosed amount |
Unresolved |
|
2024 |
Closing cash |
6,438 |
8,610 |
-2,172 |
Restricted cash, non-current, Dec 31, 2024: 2,172 |
Test (ii): FMP 6,438 = 5,893 + 545; the 2,172 line is the only line with no FMP counterpart |
Explained, mapping shown in line items |
|
2025 |
Opening cash |
6,438 |
8,610 |
-2,172 |
Restricted cash, non-current, Dec 31, 2024: 2,172 |
Same evidence as the 2024 closing row |
Explained, mapping shown in line items |
|
2025 |
Operating |
10,099 |
10,099 |
0 |
Not applicable |
Not applicable |
Reconciled |
|
2025 |
Investing |
-3,673 |
-3,564 |
-109 |
Redemption of non-controlling interests (financing): -109 |
Test (i) fails: no FMP investing field equals -109. Test (ii) fails: acquisitionsNet -120 and otherInvestingActivities -1,140 match no filing line; only their total, -1,260, equals -815 + -336 + -109 |
Arithmetic match, mapping not established |
|
2025 |
Financing |
-5,604 |
-5,713 |
+109 |
Same -109 line |
Test (ii): netDebtIssuance 852 = 3,359 − 2,350 − 157; issuance 183; repurchases -6,523; other -116. The redemption is the only filing line with no FMP counterpart |
Explained, mapping shown in line items |
|
2025 |
Exchange-rate effect |
476 |
215 |
+261 |
None; 261 appears in neither 10-K |
Calculated only: equals the negative of the change in non-current restricted cash (1,911 − 2,172), which is not a disclosed amount |
Unresolved |
|
2025 |
Closing cash |
7,736 |
9,647 |
-1,911 |
Restricted cash, non-current, Dec 31, 2025: 1,911 |
Test (ii): FMP 7,736 = 7,105 + 631; the 1,911 line is the only line with no FMP counterpart |
Explained, mapping shown in line items |
|
Dec 31, 2023 |
Balance-sheet cash and cash equivalents |
5,485 |
4,680 |
+805 |
Restricted cash, current, Dec 31, 2023: 805 |
FMP has no restricted-cash field, so test (i) fails. Test (ii) fails: FMP's other current-asset fields match the filing only as a group |
Arithmetic match, mapping not established |
|
Dec 31, 2024 |
Balance-sheet cash and cash equivalents |
6,438 |
5,893 |
+545 |
Restricted cash, current, Dec 31, 2024: 545 |
Same structure as Dec 31, 2023 |
Arithmetic match, mapping not established |
|
Dec 31, 2025 |
Balance-sheet cash and cash equivalents |
7,105 |
7,105 |
0 |
Not applicable |
Not applicable |
Reconciled |
On signs, the 2024 exchange-rate row states its relationship without the sign: standardized minus filing is −653, the negative of the +653 change in non-current restricted cash, exactly as the 2025 row says for its +261. Claude applied the same tests to the open items from Prompt 1. The fiscal 2024 continuity break of -1,519 and the fiscal 2023 observation of 1,519 rest on the same evidence as the fiscal 2024 opening row, and the fiscal 2025 gap of 631 equals current restricted cash at December 31, 2025. FMP's zero residuals in Table 2 remain valid within FMP's records; its exchange-rate and net-change fields differ from the filing by the two Unresolved amounts.
Revision check: the fiscal 2024 amounts are unchanged between the fiscal 2024 and fiscal 2025 10-Ks. The only change is presentational: the fiscal 2025 10-K's operating line "Other" of 187 combines three lines from the fiscal 2024 10-K (61, 38 and 88).
Reading the reconciliation
On the filing's definition, Uber's cash bridge closes exactly in both years and ties to the balance sheet: closing cash rose from 7,004 to 8,610 in fiscal 2024 and to 9,647 in fiscal 2025. FMP's standardized records also close, on a narrower total, and the comparison table shows how far each difference between the two can be traced.
The redemptions of non-controlling interests are the best-supported differences. In fiscal 2024, FMP's acquisitions field holds exactly the 851 that the filing reports as a financing redemption, while the filing's own acquisitions line is zero, and every other financing component matches. In fiscal 2025 the 109 is clearly missing from FMP's financing total, but its place in FMP's investing total can be seen only inside the combined acquisitions and other-investing fields, so the investing side stays an arithmetic match.
The cash totals rest on a weaker kind of line-item evidence. FMP's opening and closing balances equal Uber's cash and cash equivalents plus current restricted cash, and non-current restricted cash is the only filed line left over. That is a consistent pattern across both years, but FMP reports no component fields for its cash total, so the test compares a total with a sum of filed lines rather than field with field. The balance-sheet cash field shows the same pattern for December 2023 and 2024 and then matches the filed line alone for December 2025, which is a change in scope any screen comparing that field across years should record.
The exchange-rate differences are the only Unresolved lines. In both years the difference, standardized minus filing, equals the negative of the year's change in non-current restricted cash: −653 against a rise of 653 in 2024, and +261 against a fall of 261 in 2025. That is a numerical reconciliation, calculated from two disclosed balances, and it keeps FMP's narrower bridge closed. It is still marked Unresolved because the amount is a calculated change and not a disclosed line. The evidence shows where the difference sits; it does not show that FMP treats restricted-cash movements as exchange effects, and the article does not say why FMP's fields are built as they are. The revision check found no restated fiscal 2024 amount, a point worth repeating whenever later filings are added, as covered in how financial data revisions and restatements flow through.
The next actions follow the open lines: read FMP's field documentation, or ask FMP, for how restricted-cash movements and redemptions of non-controlling interests are mapped, and until then use the filing's lines from Table 4 for any cash figure that has to match Uber's own definition. A liquidity view needs its own choice of measure, because restricted cash is not necessarily available for general use.
When analyst review is required
Mark a line Unresolved, and record it for review, when any of these occur:
- a difference between the standardized value and the filing equals no disclosed amount;
- the cash-flow caption and the accounting-policy note describe different definitions;
- opening cash does not equal the prior year's closing cash under the same definition;
- a later filing restates a prior-year cash-flow line, and the note does not explain the change.
An arithmetic match is not a review trigger; it is a result that says how far the evidence goes. The next action for either case is a read of the filing note or line items behind the affected figure, and the line keeps its original value until that read settles the question.
Using the reconciled bridge in further research
To run the check on another company, change the symbol and fiscal years in both prompts. Companies that reclassify cash to assets held for sale, or hold restricted cash inside consolidated entities, will show additional reconciling items in Prompt 2, and the net-change check will report whether the subtotal includes them.
The same habit of matching definitions before comparing numbers applies well beyond cash. This SBC and share-count review shows a standardized field matching a broader combined measure, and a bulk fundamentals refresh is the natural place to record a field's definition alongside its values, so a change in scope between years is caught before it reaches a model.

Financial APIs, Claude MCP, and AI-driven research workflows
Pranjal Saxena writes technical content focused on financial data APIs, Claude MCP workflows, AI-driven research systems, and Python-based market analysis. For FMP, his work centers on turning structured financial data into practical, workflow-driven content for developers, analysts, and fintech teams. He combines experience in data science, NLP, generative AI, and financial API workflows to show how APIs, automation, and AI-assisted systems can support modern financial research and analysis.
Financial data for every need
Real-time quotes and 30+ years of historical data, including prices, fundamentals, and insider transactions — all accessible via API.
Create Free Account