On Tuesday, September 29, Bill Pulte, the director of the Federal Housing Finance Agency, said Fannie Mae and Freddie Mac will price mortgages off a single grid that puts VantageScore on equal footing with the FICO Score. The same day, TransUnion said it will keep selling VantageScore 4.0 for 99 cents a mortgage score through 2028. Fair Isaac, the company behind the FICO Score, fell 26.5% to $617.87. By Yahoo's price history, that was its worst day since 1989, and worse than any day of the March 2020 crash.
It was the second blow in a month. On September 4, Pulte told Fannie and Freddie to let every lender use VantageScore, and FICO fell 16.7% that day. From August 31 to September 29 the stock lost 46% of its value, and it is down 63% since the start of the year.
I'll start with what is happening to the company, then turn to where I stand and what my model did with the news. That evening, it ranked FICO fifth in its Top 30. I don't own FICO and never have.
A quick primer: who scores you, and who gets paid
A credit score is a three-digit number that estimates how likely you are to repay a loan, built from the history in your credit report. FICO, short for Fair Isaac Corporation, launched the first widely used one in 1989. VantageScore, its main rival, was created in 2006 by the three big credit bureaus, Equifax, Experian, and TransUnion, which own it together.
Mortgages are where this matters most for FICO. A large share of American home loans are sold to Fannie Mae and Freddie Mac, two government-sponsored companies that package them for investors. Because they buy the loans, they write the rules, including which credit scores a lender may use and a pricing grid that sets how much more a borrower pays for a lower score. Both answer to the Federal Housing Finance Agency. Every time a lender pulls a FICO Score, FICO collects a royalty.
Until this spring, lenders selling loans to Fannie and Freddie in practice had to use FICO. That changed quickly:
| Date | What happened |
|---|---|
| May 1, 2026 | A limited group of lenders is allowed to use VantageScore 4.0 on Fannie and Freddie loans |
| September 4 | Pulte, on X: "EFFECTIVE IMMEDIATELY, I'm instructing Fannie and Freddie to approve ALL lenders to use VantageScore." FICO falls 16.7% |
| September 29 | Pulte says Fannie and Freddie will use one pricing grid, putting VantageScore on equal footing with FICO. TransUnion locks VantageScore 4.0 at 99 cents a score through 2028. FICO falls 26.5% |
| January 1, 2027 | The Federal Housing Administration is set to start accepting VantageScore 4.0 |
The grid is the bigger deal. Until now, FICO sat on the main pricing grid, and a lender that chose VantageScore paid a financial penalty for it. With both scores on the same grid, that penalty goes away, and so does a big reason to stick with FICO.
The chart also shows that September wasn't the whole story. FICO started the year at $1,691 and was already down by half before Tuesday, including a 17% drop on July 30 after a quarterly report that missed revenue forecasts.
Why FICO looked so strong
In the three months to June, FICO's revenue grew 26%, to $674.2 million. Almost all of that came from one place:
| Segment | Apr–Jun 2025 | Apr–Jun 2026 | Change |
|---|---|---|---|
| Scores | $324.3M | $458.9M | +41.5% |
| Software | $212.1M | $215.3M | +1.5% |
| Total | $536.4M | $674.2M | +25.7% |
Source: FICO's third-quarter fiscal 2026 results, filed with the SEC on Form 8-K.
Inside the scores business, revenue from mortgage originations rose 97% from a year earlier, while the number of mortgages grew only by low single digits, according to the company's quarterly presentation. Mortgages made up 71% of its business-to-business scores revenue. In other words, FICO didn't sell many more mortgage scores. It charged a lot more for each one.
By my arithmetic from the company's figures, mortgage originations brought in roughly $285 million in the quarter, about 42% of FICO's total revenue. The rest is less exposed to Tuesday's news: scores used for credit cards, auto loans, and other lending, scores sold to consumers, which grew 5%, and the software business, where recurring revenue from FICO's newer platform grew 62% while its older products shrank 17%.
FICO's per-score mortgage royalty was tiered between 60 cents and $2.75 in 2023. It rose to $3.50 in 2024, $4.95 in 2025, and $10 this year, though lenders can instead choose a $4.95 fee plus $33 per borrower, per score, on each loan that closes.
Those increases drew political fire long before this month. In March, Senator Josh Hawley opened an investigation, writing to FICO's chief executive and urging the Federal Trade Commission to investigate as well. His letter pointed to the jump from $4.95 to $10, which he estimated could add about $500 million to the industry's credit score costs, and to what he called an 88% operating margin. VantageScore, for its part, cites a study by Deep Future Analytics estimating that opening Fannie and Freddie to its score saves more than $930 million in the first year.
FICO's defense is fair to note. When it set the $4.95 price, it said its royalty was about two-tenths of one percent of average closing costs, and that its score is used through the whole life of a loan, from underwriting to "insuring, securitizing, rating, selling."
The growth that made FICO look so strong came from the same pricing power the new rules target. The strength and the threat are the same number.
How fast lenders are switching
TransUnion says more than 1,100 mortgage lenders adopted VantageScore 4.0 between January and September, including 9 of its 15 largest mortgage customers. In the first four months of the limited rollout, VantageScore was the only credit score used on more than 9% of the mortgages Fannie and Freddie securitized. TransUnion is also giving VantageScore away free to mortgage customers who buy a FICO score, which makes trying it close to costless.
FICO's lifecycle argument is the strongest case that this won't spread far. A lender changing which score it pulls is easy, but investors, insurers, and rating agencies further down the chain have to follow for the change to stick. The single grid removes one of those links.
Why I'm looking now
I was always aware of FICO, but it wasn't a stock that was on my radar until it dropped about 45% in a month. I'm looking into the company now because fears of disruption may have presented a buying opportunity.
My first instinct was to check how sticky FICO was. It appears a lot of lenders have switched, and pretty easily. That makes me feel the switches could snowball into something more.
I've been burned by news hitting a business before. AT&T's dividend cut cost me quite a bit, a story I told in the mistakes that cost me the most; in 2022 it cut the annual payout from $2.08 a share to $1.11 as it spun off WarnerMedia. The lesson I took from it was to really evaluate the news, and how it can affect the future numbers.
What my model did
My model doesn't read the news. It scores every company on three things: quality, growth, and value. The score blends them into one number. Here is FICO, as the model reads it:
- Quality: near perfect. An 81% return on invested capital, a 42% free-cash-flow margin, and a 34% net margin, all at full marks. The one zero is return on equity, and only because FICO has bought back so much stock that its balance sheet shows negative shareholders' equity.
- Growth: strong. Revenue up 24% over the past year, and full marks on the forward-growth part, which compares analysts' earnings estimates with FICO's past earnings.
- Value: improving. This is the part that moves with the share price. As the stock falls, FICO looks cheaper, and the score goes up.
The Top 30 allows at most three companies from any one industry group, and software already had three: Adobe, Intuit, and GoDaddy. FICO had been fourth, 0.1 point behind GoDaddy. Tuesday's drop lifted its score from 84.5 to 85.8, past both.
It had happened before. On September 4, the day of the first announcement, FICO's score jumped from 81.6 to 84.9 and it entered the Top 30 at No. 9. It drifted in and out all month.
From August 31 to September 29, the price fell 46% and the score rose 4 points. Both times the regulators moved against FICO, the model moved FICO up.
What it would take to knock FICO back out
The model will only react once the news reaches the numbers, so the useful question is how much damage it would take. Rerunning FICO through the same scorer with Tuesday's figures, changing one input at a time and holding everything else where it is, gives a clear answer. To keep its spot, FICO has to stay above GoDaddy's 84.6.
| If this happens | FICO's score | Still in the Top 30? |
|---|---|---|
| Nothing changes | 85.8 | Yes, at No. 5 |
| Analysts cut their earnings estimate 20% | 85.5 | Yes |
| Analysts cut it 30% | 82.0 | No |
| Revenue growth slows from 24% to 15% | 84.7 | Barely |
| Revenue growth slows to 10% | 83.0 | No |
| Revenue stops growing | 68.4 | No |
| Estimates cut 30% and revenue stops growing | 62.1 | No |
Everything not listed is held at its September 29 value, and the other companies' scores are assumed not to move.
Two things stand out. The model is patient with analysts: a 20% cut to the earnings estimate, which would be a big move, barely touches the score, because FICO's expected growth sits so far above the level that earns full marks. It is much less patient with reported revenue. If FICO's growth over a full year slows to 10%, it falls out of the Top 30, and if revenue stops growing altogether, the score drops to 68.4 even with estimates untouched. In every case it is the growth side doing the damage. Quality only moves if FICO's margins and returns on capital start falling too.
Blind, but not only in a bad way
The model isn't programmed to weigh subjective events. It looks at the numbers. The issue that creates is that it can be blind to events until they actually show up in the fundamentals. From the other side, it may also save you from overreacting if the numbers never get hurt.
It's the same line I drew in my piece on rates: the model's job isn't to focus on subjective maybes. It uses the data to highlight strong fundamentals, and it is up to the reader to make the judgment on something this abstract.
For FICO, the news will reach the score in stages, and at different speeds:
| What changes | How often | Where it stands |
|---|---|---|
| Share price | Every trading day | Down 46% since August 31, already in the score |
| Analysts' earnings estimates | Whenever analysts revise | Down 0.3% in 30 days and 2.6% in 90 |
| Reported revenue and margins | Every quarter | Latest covers April to June, before either announcement |
The middle row is the one I'd watch. The site's estimate revisions page tracks it every weekday. So far, analysts have barely touched their numbers, even as their price targets fell; Wells Fargo has cut its target from $1,350 to $950. And the next quarterly report covers July through September, so it will show only the first few weeks of open competition and none of the single grid.
That is the gap this whole piece is about. The market is pricing a future the numbers haven't reached yet, and the model is scoring a past the market no longer trusts. Neither one is wrong about what it is looking at.
What I'm doing
I'm interested, but I'll be waiting until I see a signal of strength or weakness. I would rather be late and catch some of the upside than buy in early and try to catch a falling knife.
A strength signal would be mortgage revenue holding up once lenders have real choice. A weakness signal would be FICO cutting its price, or analysts finally cutting their estimates. Either way, I want the numbers to tell me which story is true before I pay for one of them.
This is my personal view and not investment advice. Scores on this site are research tools; they don't know your goals or your risk tolerance. Do your own research before buying or selling anything.