I Ran the Numbers on a Seattle House with AI — The Results Were Sobering
A $900K house with a "manageable" monthly payment, prices that "always go up," and a built-in "inflation hedge" — then I had AI crunch every number, and things got complicated.
TL;DR: On this $899K house, 87% of the monthly payment is interest. Breaking even in 10 years requires 6.71% annual appreciation (above historical averages). Selling within 5 years loses money in every scenario. Buying isn't wrong — but do the math first.
For most people, the home-buying decision goes something like this:
See a nice house → run the monthly payment — seems doable → prices always go up long-term, can't lose → make an offer.
Sounds reasonable, right?
I recently used an AI-powered real estate analysis system I built to do a full financial breakdown on a real listing in the Greater Seattle area. The system doesn't tell stories or play on emotions — it just runs the numbers. And after seeing the results, it completely changed how I think about buying a home.
Today, let's follow this one house through the entire buying decision. At every step, we'll let the real data do the talking.
The House
Location: Greater Seattle area, an Eastside school district
List price: $899,000
Layout: 3 bed / 2 bath, ~1,800 sqft
Down payment: 20% ($179,800)
Loan: $719,200, 30-year fixed at 6.75%
Buyer closing costs: $13,485 (1.5%)
Total cost to get in: $193,285
A fairly standard Eastside school-district home. Not a mansion, but at today's rates, the monthly payment isn't small either.
So here's the question —
Checkpoint 1: Can You Afford the Monthly Payment?
The bank calculates a monthly payment (principal + interest) of $4,665. Looks okay?
But that's only the payment you can see. I had the system break it down:
| Month 1 | Share | |
|---|---|---|
| Interest (sunk cost) | $4,046 | 86.7% |
| Principal (actual equity) | $619 | 13.3% |
In other words, of the $4,665 you pay the bank each month, nearly 87% is interest — a pure cost of borrowing, no different from rent. You never see that money again. The portion actually building equity? Just $619 in month one.
People love to say "rent is throwing money away," but the $4,046 in interest out of your $4,665 payment is also throwing money away. The only difference is you're throwing it at a bank instead of a landlord.
After the first year: total payments of ~$55,977, of which $48,312 is interest (86.3%) and only $7,665 goes to principal.
But that's not all. Homeownership comes with a "hidden monthly payment" that never shows up on your bank statement:
| Item | Monthly Cost |
|---|---|
| Property Tax | ~$674 |
| Maintenance | ~$562 |
| Homeowner's Insurance | ~$300 |
| Total | ~$1,536 |
Note: Property tax is an annual tax based on your home's assessed value — roughly 0.9–1.0% of home value in Washington State. It's not a one-time fee. You pay it every year, and it rises as your home value increases.
Add interest and hidden costs together, and you're throwing away $5,582 every month — and maintenance and property taxes keep climbing over time.
Total monthly outlay: $6,201. Only $619 of that builds equity. The remaining $5,582 is gone — just like rent.
So "I can afford the payment" and "this is a good use of money" are two very different things.
Checkpoint 2: Prices Always Go Up, Right?
"U.S. home prices grow ~5% per year long-term" — you've probably heard this. It comes from the Case-Shiller National Index, and as a long-run average, it's true.
But behind that average lies enormous volatility:
- Seattle 2022: prices dropped 14%, took ~2 years to recover
- Seattle 2008: prices dropped 28%, took a full 6 years to recover
More critically, for this $899,000 house, the system calculated a 10-year break-even CAGR of 6.71% — meaning prices need to grow 6.71% per year just to break even after 10 years. That's nearly 1.7 percentage points above the historical average.
In other words, you're not betting that prices will rise — you're betting they'll rise faster than they historically have.
The system modeled four scenarios (10-year hold):
| Annual Appreciation | Sale Price at Year 10 | Net Profit |
|---|---|---|
| Pessimistic 2.0% | $1,096K | -$577,000 |
| Conservative 3.5% | $1,268K | -$418,000 |
| Baseline 5.0% | $1,464K | -$236,000 |
| Optimistic 6.5% | $1,688K | +$71,000 |
Only at an optimistic 6.5% growth rate do you barely turn a $71K profit after 10 years. At the baseline 5%, you still lose $236,000 after a decade.
Checkpoint 3: At Least It's an Inflation Hedge?
"When inflation rises, home prices rise too — buying is at least a hedge against inflation."
Half true. The problem: your holding costs rise too.
Property taxes track assessed value (they go up with the house), maintenance tracks CPI, insurance adjusts annually. Your "monthly payment" effectively gets more expensive every year — even though the mortgage itself is fixed-rate, the hidden costs keep climbing.
There's an even more fundamental issue: the friction cost of buying and selling a home is enormous.
In Washington State, round-trip costs run roughly 8–10%:
- Buyer closing costs: ~1.5% (title insurance, attorney fees, loan origination, prepaid taxes, etc.)
- Seller closing costs: ~5.5% (primarily agent commissions)
- WA State REET: 1.1–2.75% (Washington's Real Estate Excise Tax — a progressive transfer tax where higher-priced homes pay higher rates)
For this house, even if prices rise 63% over 10 years (5% annually), round-trip friction plus cumulative holding costs still exceed the appreciation — net loss: $236,000.
Some people plan to wait for rates to drop and refinance, but refinancing itself costs 2–3% in fees, takes 2–3 years to recoup, and nobody can guarantee where rates will land.
Bottom line: Real estate as an inflation hedge only truly works with low rates (<4%), long holding periods (>15 years), and low transaction friction. At 6.75% interest plus Washington's high REET, that advantage is severely diminished.
Checkpoint 4: But the Internet Says It's Worth the Price?
The first three checkpoints assumed the house is actually worth $899,000. But a more fundamental question: is it really?
Many buyers check Redfin Estimate or Zillow's Zestimate. If the estimate is close to the listing price, they feel the price is "fair."
But these estimates come from algorithmic models called AVMs (Automated Valuation Models). They have critical limitations:
- Median error is about ±2–3%, but for non-standard properties (large lots, unique architecture, major renovations), the error can be much larger
- AVMs don't see the inside of the house. A beautifully renovated home and a gut-job next door can get nearly identical valuations
- Coverage is about 86% — meaning 14% of homes don't even have an estimate
There's another easy trap: listings that say "recently updated" or "beautifully renovated." Sounds reassuring. But renovations come in two flavors — owner-occupied quality renovations with good materials done right, and investor flips with cheap materials and cosmetic-only work.
Key flipper signals: purchase-to-list markup over 30% with less than 3 years of ownership, or only surface-level updates (fresh paint, new light fixtures, replaced countertops) while plumbing, electrical, and roofing are untouched.
When you see "renovated," that's exactly when you should ask: who did the work? What was done? What wasn't?
The right approach: AVMs are useful for rough screening. Real offers should be based on recent comparable sales on the same street, price-per-square-foot adjustments, and in-person inspection.
Checkpoint 5: Just Buy Now, Trade Up Later
"Get on the ladder now, live in it for three to five years, sell when it appreciates, and upgrade."
This mindset is incredibly common — and incredibly risky. Because 5 years is the worst possible holding period.
Why is this idea so popular? Classic FOMO (Fear of Missing Out) — seeing others profit from real estate and feeling like you'll lose out if you don't buy. But what you're seeing is survivorship bias: the winners talk, the losers stay quiet.
Back to the numbers:
Reason 1: Transaction friction hits hardest on short holds.
The 8–10% round-trip cost, spread over 30 years, averages under 0.3% per year. But over just 5 years, it's nearly 2% per year in extra drag. Home prices need to rise 8–10% just to break even — before interest and taxes.
Reason 2: The first 5 years of payments are almost entirely interest.
On the $719K loan, 5 years builds only $44K in equity while paying $236K in interest. You think you're "building a down payment" for the next house, but most of it is just paying rent to the bank.
Reason 3: When you trade up, you're buying into the same inflated market.
Your starter home gained 25%, but the bigger house you want also gained 25% — the absolute price gap is actually larger. "Getting on the ladder" doesn't give you an advantage when upgrading.
The system modeled four 5-year scenarios:
| Annual Appreciation | Sale Price at Year 5 | Net Profit |
|---|---|---|
| Pessimistic 2.0% | $992K | -$328,000 |
| Conservative 3.5% | $1,068K | -$258,000 |
| Baseline 5.0% | $1,147K | -$185,000 |
| Optimistic 6.5% | $1,232K | -$106,000 |
Even at an optimistic 6.5% annual growth, selling after 5 years still loses over $100,000.
The 5-year break-even requires 8.41% annual appreciation — a rate that has rarely been sustained historically.
The more realistic scenario: within 5 years, a job change, family shift, or forced relocation pushes you to sell at the worst possible time. That's not "missing out on gains" — it's losing $100,000 to $300,000 in real money.
After Running the Numbers
After all five checkpoints, back to the original question: should you buy this $899,000 house?
The numbers are clear:
- 87% of the payment is interest — add hidden costs, and monthly sunk costs exceed $5,500
- Breaking even in 10 years requires 6.71% annual appreciation — nearly 1.7 points above the historical average
- At the baseline 5% growth, you still lose $236K after 10 years — only the optimistic scenario barely turns positive
- Online estimates aren't reliable — renovated homes may just be cosmetic flips
- Selling within 5 years loses money in every scenario — "just get on the ladder" is actually the riskiest strategy
But this article isn't telling you not to buy.
A home is family, lifestyle, school district, stability — things that can't be measured in numbers. But when making what might be the biggest financial decision of your life, you should at least know what the numbers are telling you.
Don't make a six-figure decision on gut feeling. Run the numbers before you sign.
The data in this article is based on simulated analysis using real Greater Seattle market parameters: 6.75% mortgage rate, 20% down payment, 1.0% property tax rate, among others. Specific numbers will vary by market and conditions, but the analytical framework is universally applicable.
All figures were generated by an AI-powered real estate analysis system.
If this was useful, share it with someone who's house-hunting right now.
This article is part of a series on data-driven decision making. Next: one request, five AIs, two articles — how this piece was actually written (the behind-the-scenes). If you're curious what happened when I ran the same numbers on renting vs. buying — renting won every time. Subscribe to get the next one.