Most people running some kind of online income — freelance money, e-commerce float, agency retainers, crypto proceeds waiting to deploy — treat their cash balance like furniture. It sits wherever it landed. A payment processor balance here, a branch savings account there, a Wise multi-currency wallet with a few currencies nobody bothered to optimize.

Here’s the uncomfortable math: at the time of writing, a typical mega-bank branch savings account pays 0.01% APY (Wells Fargo’s Way2Save, per Investopedia’s November 2026 rate review). Meanwhile Wise’s personal accounts pay 4.00% APY on USD through their opt-in interest feature, Revolut’s US tiers pay 3.5–4.0% APY depending on plan, and Wealthfront’s Cash Account pays 3.55% APY. These are not exotic products. They are free-ish accounts you can open in an afternoon.

The spread between where your money is and where it could be is a silent invoice you’re paying to nobody. On $40,000 of idle USD, that invoice is roughly $1,400–$1,600 a year. On a multi-currency float — the normal situation if you invoice in USD but live in EUR or GBP land — most people leave the non-USD balances earning literally zero, because “it’s just sitting there before I convert it.”

This post gives you a working Cash Spread Engine: a Python script that scores every legal parking spot for your float, blends promotional rates with what the rate becomes after the promo dies, subtracts subscription costs where a tier requires one, and prints the exact dollar leak per holding. It’s the same “check the spread before you move” discipline crypto arbitrage people apply to perp funding rates — applied to the most boring asset class there is: cash.

Why this is a money-maker, not a chore

Three reasons this bucket belongs next to your other income streams rather than in a personal-finance guilt pile:

  1. It scales with your business, for free. Every extra $10k of float that flows through your operation adds ~$350–400/year at current rates, permanently, with zero additional clients or work.
  2. It’s the closest thing to true passive income that isn’t a scam. No audience to build, no product to launch. The “work” is a one-time audit plus an hourly-per-year review.
  3. The multi-currency angle is genuinely neglected. US-centric advice covers USD only. If you hold EUR and GBP working balances — extremely common for anyone using Wise to receive international payments — the default state is 0% interest. Wise pays 2.00% on EUR and 2.10% on GBP via the same opt-in feature (rates verified from Wise’s own interest page, shown as of November 2026).

The honest fine print (read this before moving anything)

This playbook is legal, boring, and reversible — but it has real fine print. Don’t skip this section:

  • Wise Interest is opt-in and jurisdiction-gated. The 4% USD / 2% EUR / 2.1% GBP rates apply to eligible customers who explicitly turn the feature on; your balance gets swept to a program bank, with FDIC passthrough insurance up to $250,000 in the US. Availability and rates differ per region and per account type — Wise’s business accounts show a different (currently lower) USD rate than personal. Check the page for your region before assuming any number.
  • All of these rates are variable. They track central bank policy and can drop the day after you open the account. Nothing here is a locked return.
  • Promotional rates expire. Wealthfront advertises 4.20% APY for the first 3 months for new clients, then 3.55%. A naive comparison overstates year-one yield; the engine below blends this correctly.
  • Higher tiers can require paid plans. Revolut’s best US savings rate (4.00%) sits on the Metal plan, which costs $16.99/month or $149.99/year per Revolut’s official US fee schedule. Sometimes the tier pays for itself, sometimes it doesn’t — the engine computes the net either way instead of guessing.
  • Some rates are gated behind behavior. Chime’s advertised 4% requires qualifying direct deposits; without them you’re at their standard 1%. If you can’t route a qualifying deposit, that option isn’t real for you.
  • Moving money between providers has friction and (in edge cases) risk. Bank payouts can take days, and some fintech platforms have had well-publicized stability scares. Splitting float across two providers is cheap insurance.

None of this is financial advice — it’s an engineering approach to a spread that is measurable today.

The engine

The design is deliberately simple: a rate table (currently mocked with verified figures — swap in a live scraper when you want it self-updating), a blender that converts “promo rate + base rate + months” into honest year-one and steady-state numbers, and a report function that takes your actual holdings and prints the leak.

Save this as spread_engine.py:

#!/usr/bin/env python3
"""
Cash Spread Engine - stop bleeding yield on float.
Finds the best legal parking spot for every dollar you hold, per currency,
blending promotional rates with long-run rates, and shows what the drift costs you.

Tested with mocked rate feeds (all rates in RATES are hard-coded from public
pricing pages checked 2026-11-02). Plug in a live scraper/API and the logic
is unchanged. Not financial advice - verify rates at the source before moving money.
"""

from dataclasses import dataclass, field
from datetime import date
from typing import List, Optional

# ---------------------------------------------------------------
# Rate feed (mocked - replace with live data if you want)
# Source: provider pricing pages checked Nov 2, 2026. Variable rates.
# ---------------------------------------------------------------
RATES = {
    "USD": [
        # (provider, product, base APY, promo APY, promo_months, notes, restrictions)
        ("Wise",        "Interest (USD, personal)", 0.040, None, 0, "opt-in, swept to program bank, FDIC passthrough to $250k", None),
        ("Wise",        "Interest (USD, business)", 0.034, None, 0, "Wise Business tier, opt-in, as of Sep 18 2026", None),
        ("Revolut",     "High-Yield Savings",    0.035, None,  0, "Standard plan tier", None),
        ("Revolut",     "High-Yield Savings",    0.0375, None, 0, "Premium plan tier", None),
        ("Revolut",     "High-Yield Savings",    0.040, None,  0, "Metal plan tier (subscription cost applies)", "plan_cost"),
        ("Wealthfront", "Cash Account",          0.0355, 0.0420, 3, "new-client promo 3 months, then 3.55%", "one_shot"),
        ("Chime",       "High-Yield Savings",    0.010, 0.0400, 12, "standard 1.00%; 4.00% only with qualifying direct deposit", "direct_deposit"),
        ("BigBankUSA",  "Way2Save-style branch savings", 0.0001, None, 0, "typical mega-bank branch rate", None),
    ],
    "EUR": [
        ("Wise",    "Interest (EUR)", 0.020, None, 0, "opt-in feature, EUR balances", None),
        ("BigBank", "Branch savings", 0.0005, None, 0, "typical EU branch rate", None),
    ],
    "GBP": [
        ("Wise",    "Interest (GBP)", 0.021, None, 0, "opt-in feature, GBP balances", None),
        ("BigBank", "Branch savings", 0.0010, None, 0, "typical UK branch rate", None),
    ],
}


@dataclass
class Holding:
    currency: str
    amount: float
    where: str            # where the money currently sits
    current_apy: float    # what it actually earns today

    def yearly_interest(self) -> float:
        return self.amount * self.current_apy


@dataclass
class Option:
    provider: str
    product: str
    blended_apy: float
    base_apy: float
    year1_apy: float
    notes: str
    restriction: Optional[str] = None
    plan_cost: float = 0.0          # $/yr cost of keeping the tier alive
    year1_net: float = 0.0          # net $ in year 1 on a given balance
    steady_net: float = 0.0         # net $/yr at base rate after promo dies


def blended_apy(base: float, promo: Optional[float], promo_months: int) -> tuple:
    """Weighted APY: promo rate for promo months, base for the rest of year 1."""
    if not promo:
        return base, base
    m = max(0, min(12, promo_months))
    year1 = (promo * m + base * (12 - m)) / 12
    return year1, base


def build_options(currency: str, balance: float) -> List[Option]:
    out = []
    for provider, product, base, promo, pm, notes, restr in RATES[currency]:
        y1, steady = blended_apy(base, promo, pm)
        plan_cost = 0.0
        if restr == "plan_cost":
            # Revolut Metal US: $16.99/mo or $149.99/yr (official US fee schedule).
            # Yearly billing = cheapest, no perks netted out. Source checked Nov 2, 2026.
            plan_cost = 149.99
        opt = Option(
            provider=provider, product=product,
            blended_apy=steady, base_apy=base, year1_apy=y1,
            notes=notes, restriction=restr, plan_cost=plan_cost,
            year1_net=balance * y1 - plan_cost,
            steady_net=balance * steady - plan_cost,
        )
        out.append(opt)
    out.sort(key=lambda o: o.steady_net, reverse=True)
    return out


def spread_report(holdings: List[Holding]) -> dict:
    report = {"date": date.today().isoformat(), "rows": [], "total_leak": 0.0}
    for h in holdings:
        options = build_options(h.currency, h.amount)
        best = options[0]
        earn_now = h.yearly_interest()
        # steady-state (post-promo) comparison - what you keep every year going forward
        best_steady = max(o.steady_net for o in options)
        leak = best_steady - earn_now
        promo_pick = options[0]
        report["rows"].append({
            "currency": h.currency, "amount": h.amount, "where": h.where,
            "current_apy": h.current_apy, "earns_now": round(earn_now, 2),
            "best_home": f"{best.provider} - {best.product}",
            "best_steady_apy": best.base_apy, "best_steady_net": round(best_steady, 2),
            "best_year1_net": round(max(o.year1_net for o in options), 2),
            "leak_per_year": round(leak, 2),
            "runner_up": f"{options[1].provider} - {options[1].product}" if len(options) > 1 else "-",
        })
        report["total_leak"] += leak
    report["total_leak"] = round(report["total_leak"], 2)
    return report


def main():
    holdings = [
        Holding("USD", 25_000, "branch savings at a mega-bank", 0.0001),
        Holding("USD", 15_000, "Wise USD balance (interest OFF)", 0.0),
        Holding("EUR", 8_000,  "Wise EUR balance (interest OFF)", 0.0),
        Holding("GBP", 4_000,  "Wise GBP balance (interest OFF)", 0.0),
    ]
    rep = spread_report(holdings)
    print(f"Cash Spread report - {rep['date']}")
    print("=" * 100)
    for r in rep["rows"]:
        print(f"[{r['currency']}] ${r['amount']:>10,.0f} sitting at: {r['where']} "
              f"(earns ${r['earns_now']:.2f}/yr at {r['current_apy']*100:.2f}%)")
        print(f"    -> best steady home : {r['best_home']} "
              f"(~{r['best_steady_apy']*100:.2f}% base, ~${r['best_steady_net']:.0f}/yr net)")
        print(f"    -> best year-1 net  : ${r['best_year1_net']:.0f} (includes any promo)")
        print(f"    -> runner-up        : {r['runner_up']}")
        print(f"    -> LEAK             : ${r['leak_per_year']:.0f}/yr you are not collecting")
        print("-" * 100)
    print(f"TOTAL LEAK across holdings: ${rep['total_leak']:,.0f} per year")
    print("\nTop USD options ranked by steady-state net:")
    for o in build_options("USD", 40_000)[:5]:
        cost = f", tier cost -${o.plan_cost:.0f}/yr" if o.plan_cost else ""
        print(f"  {o.provider:<12} {o.product:<30} base {o.base_apy*100:>5.2f}%  "
              f"yr1 {o.year1_apy*100:>5.2f}%  net ${o.steady_net:>8,.0f}/yr{cost}  [{o.notes}]")
    assert rep["rows"][0]["leak_per_year"] > 900, "sanity: 25k USD leak should exceed ~$900/yr"
    assert 1_000 < rep["total_leak"] < 3_000, "sanity: total leak in expected band"
    print("\nOK: engine ran clean, sanity check passed.")


if __name__ == "__main__":
    main()

The output

Run it (python3 spread_engine.py) and you get your leak report. On the demo portfolio — $25k rotting in a branch account, $15k of USD float and €8k/£4k of EUR/GBP float sitting at 0% — here is the actual output this script produced when executed on Python 3.14:

Cash Spread report - 2026-11-02
====================================================================================================
[USD] $    25,000 sitting at: branch savings at a mega-bank (earns $2.50/yr at 0.01%)
    -> best steady home : Wise - Interest (USD, personal) (~4.00% base, ~$1000/yr net)
    -> best year-1 net  : $1000 (includes any promo)
    -> runner-up        : Revolut - High-Yield Savings
    -> LEAK             : $998/yr you are not collecting
----------------------------------------------------------------------------------------------------
[USD] $    15,000 sitting at: Wise USD balance (interest OFF) (earns $0.00/yr at 0.00%)
    -> best steady home : Wise - Interest (USD, personal) (~4.00% base, ~$600/yr net)
    -> best year-1 net  : $600 (includes any promo)
    -> runner-up        : Revolut - High-Yield Savings
    -> LEAK             : $600/yr you are not collecting
----------------------------------------------------------------------------------------------------
[EUR] $     8,000 sitting at: Wise EUR balance (interest OFF) (earns $0.00/yr at 0.00%)
    -> best steady home : Wise - Interest (EUR) (~2.00% base, ~$160/yr net)
    -> best year-1 net  : $160 (includes any promo)
    -> runner-up        : BigBank - Branch savings
    -> LEAK             : $160/yr you are not collecting
----------------------------------------------------------------------------------------------------
[GBP] $     4,000 sitting at: Wise GBP balance (interest OFF) (earns $0.00/yr at 0.00%)
    -> best steady home : Wise - Interest (GBP) (~2.10% base, ~$84/yr net)
    -> best year-1 net  : $84 (includes any promo)
    -> runner-up        : BigBank - Branch savings
    -> LEAK             : $84/yr you are not collecting
----------------------------------------------------------------------------------------------------
TOTAL LEAK across holdings: $1,842 per year

Top USD options ranked by steady-state net:
  Wise         Interest (USD, personal)       base  4.00%  yr1  4.00%  net $   1,600/yr  [opt-in, swept to program bank, FDIC passthrough to $250k]
  Revolut      High-Yield Savings             base  3.75%  yr1  3.75%  net $   1,500/yr  [Premium plan tier]
  Revolut      High-Yield Savings             base  4.00%  yr1  4.00%  net $   1,450/yr, tier cost -$150/yr  [Metal plan tier (subscription cost applies)]
  Wealthfront  Cash Account                   base  3.55%  yr1  3.71%  net $   1,420/yr  [new-client promo 3 months, then 3.55%]
  Revolut      High-Yield Savings             base  3.50%  yr1  3.50%  net $   1,400/yr  [Standard plan tier]

OK: engine ran clean, sanity check passed.

That’s $1,842/year of dead money found by a 150-line script, on a modest six-figure-adjacent float. Notice the details the engine got right that a manual “just pick the highest number” approach would have missed:

  • Revolut Metal (4.00%) lands below Premium (3.75%) on $40k once the $149.99/year tier fee is subtracted — until your float crosses roughly $75k, the fee eats the extra 0.25%.
  • Wealthfront’s year-1 column shows 3.71%, not 4.20%, because the promo decays to 3.55% after three months. The steady-state ranking correctly ignores the teaser.
  • Chime doesn’t even appear in the top 5. Its 4% headline collapses to 1% steady-state without a qualifying direct deposit, and the engine refuses to flatter it.

Making the rate feed live

The engine ships with mocked rates — that’s disclosed up front, and every figure in the table came from the providers’ own pricing pages as of November 2, 2026. Rates drift constantly (Wise itself moved its headline USD rate from 3.4% to 4.0% between September and November 2026), so if you want the report to stay honest, automate the feed:

import requests
from bs4 import BeautifulSoup

def fetch_wise_us_apy() -> float:
    r = requests.get("https://wise.com/us/interest/",
                     headers={"User-Agent": "Mozilla/5.0"}, timeout=20)
    r.raise_for_status()
    soup = BeautifulSoup(r.text, "html.parser")
    m = None
    for el in soup.stripped_strings:
        if "APY" in el:
            m = el
    if not m:
        raise ValueError("no APY text found - page layout changed")
    import re
    match = re.search(r"(\d+\.\d+)%\s*APY", m)
    if not match:
        raise ValueError("no rate pattern in text: %r" % m)
    return float(match.group(1)) / 100

One honest caveat: fintech marketing pages are JavaScript-heavy and change layout often, and some of them (Revolut’s US pages at the time of writing) sit behind aggressive bot protection. The function above is written defensively — it raises instead of guessing — but treat scraping as best-effort and always cross-check a moved number against the provider’s help/legal pages before you actually move five figures. A stale rate in the table costs you a few dollars; a wrong rate acted on costs more.

The workflow that actually captures the money

The script is the audit. The money is captured by a boring quarterly routine:

  1. Run the engine on your real holdings. Update the holdings list with actual balances and actual current APYs (open each app and read the real number — “I think it earns something” is how leaks survive).
  2. Move only what’s genuinely idle. Operating cash you need in 48 hours stays where it is. The float that hasn’t moved in 60+ days is the target. Most operators discover 60–80% of their balance qualifies.
  3. Turn on the opt-in features first. If your float already sits in Wise, this is literally toggling interest on in the app — the multi-currency balances most people leave at 0% start earning the same day.
  4. Respect the ceiling. FDIC passthrough covers $250k per US setup. Above that, split across providers — which also hedges the “one fintech has a bad week” risk.
  5. Re-run quarterly. Rates moved three times at one provider this year alone. Ten minutes per quarter keeps the spread captured.

Scale this further

Once the audit mindset is installed, adjacent spreads open up:

  • The EUR/GBP angle for non-US operators. Most non-US readers hold working balances in their local currency at 0%. Wise’s EUR (2.00%) and GBP (2.10%) features are the same opt-in toggle.
  • Promo stacking as a deliberate practice. New-client teasers (like Wealthfront’s 4.20% for 3 months) are small but real: on $40k, the 3-month promo is worth ~$65 over the base rate. Churn-able float across providers is a micro-version of the bank bonus churning you may already run.
  • Compare against your other float sinks. If you hold USDC on an exchange earning a promo rate, price that against its depeg/counterparty risk versus a 4% FDIC-passthrough account. The engine’s table format extends to any instrument with a yield and a risk note.

The bottom line

The Cash Spread Engine turns “I should probably optimize my savings sometime” into a number: $1,842/year on the demo portfolio, found in one script run, recurring, with no clients and no content to produce. The rates will drift, the promos will rotate, and the engine doesn’t care — it just needs a quarterly re-run with fresh numbers.

Your float is already working for someone. The only question this script answers is whether that someone is you.

Tested with mocked rate feeds (figures hard-coded from providers’ own pricing pages checked November 2, 2026) — run once with your real balances. All rates are variable and jurisdiction-dependent; nothing here is financial advice.