UPST's time has not yet come

My case against holding UPST is best explained by an Upstart Update by David Girouard. Update on Upstart’s Credit Performance and Funding Model - Upstart Updates.

The update is in four parts. The first makes the case that the Upstart AI credit model does a superior job of assessing risk. I do not dispute this claim.

The second part states that the credit cycle affects the default rates and the aim of Upstart is to be the fastest to react. I do not dispute this, either. But I believe that the credit cycle dampens enthusiasm of banks to make loans and for institutions to buy them and for underwriters to securitize them. It seems obvious to me that growth in loan volume will disappoint in this phase of the credit cycle. Thus, earnings will disappoint.

The third point made confirms the point I just made. “As a result, some of [our parnters] have paused or reduced their originations in unsecured lending, which has limited our ability to grow.” There have not been any securitizations issued since at least June. This means that the volume of loans, regardless of consumer demand for them, is limited by the willingness of banks to make and hold the loans and Upstarts capacity and willingness to add additional loans to its balance sheet. Therefore they are seeking committed capital arrangements similar to what Affirm has arranged. But this will take time and it seems to me that in the early stages of this, Upstart is not in a strong negotiating position to get the best terms. In any case, I would not expect to see tailwinds from available capital for several quarters. Upstart is confident to add loans to their balance sheet but I don’t believe the institutional holders of UPST stock will share that confidence so this will drag on the UPST stock price.
Finally, the Update makes the point that they have $800 million of unrestricted cash on the balance sheet. This is fine, but in the good old days the securitizations were in the range of $350 million. Now the cash flow for last quarter was… not easy to find. The 10-Q does not show it–is far as I can see. They give the 6-month cash flow only. But if I go back to IR and download 1Q, I see positive cash flow of $43 million. The 6-month cash flow is negative $320 million. This due to $1.125 Billion of loans “held for sale”. (Held for sale because they are “not a bank”). Obviously they need securitizations and those are not being generated. So how long will $800 million of unrestricted cash last? As auto loans ramp up, those will have to go on the balance sheet, too.

Now, I do think that UPST has a durable business. The American consumer lives on credit and UPST does a better job of assessing credit. But I think there are better opportunities for investing now. I will nibble at $17 but be tempted to take profit at…$23? I will own some UPST to write calls against. At some point UPST will be a stock to hold for 2 to 3 years but for me, not now.



I see 38 “reads” of this post in 4 hours. This must be the result of the longer subject lines and the tagging. We used to be able to sort of hide here in WAFtT, but now we are more open to comment, correction, and garden variety humiliation. A good thing, really. Have to put our big boy britches on. More work. Like having to look at Q1 FCF to get the Q2 only FCF. And being open to “No, dummy, look at page xy, Q2 FCF is right there”. I kind of liked a quiet corner where everyone knew that “Its only KC’s slightly demented ramblings…”.

As you were,


If this is true, why do we not see more lenders coming onboard? Upstarts data shows an A+ URG (upstart risk grade) has a 0.7% default rate, while an E- URG has a 11.5% default rate. That looks like an amazing number. For a $2B market cap, why are we not seeing multiple lenders trying to acquire?

I feel like I am missing something from the big picture.

1 Like

Hi 5761796E65,
Is your username also a Swiss bank account, containing significant wealth?
Asking for a friend…


Maybe, you can try it. It’s actually my first name spelled out in hexadecimal.


Not to highjack the UPST thread, but I note you have a few NVDA posts in your history. Still a fan?

I have two targets out there: $106 and $88. Kinda/sorta made up, but they look good on a 5-year chart. Those levels start to unwind the covid/gaming boost, but they would have to get to about $65 to complete get back to pre-covid.

Maybe we take the discussion onto a new thread, but I am very interested in NVDA if it can get down to a level where I can easily see it doubling/tripling in the following 4-5 years.

More and more, I think I need to wait for $88 and then average down if opportunity arises, as $106 may still be too rich.


I will be back into NVDA eventually. I’ll start looking at it once it hits $100. Next Q earnings/revenue are going to be bad.

1 Like
  • why do we not see more lenders coming onboard?*
    why are we not seeing multiple lenders trying to acquire?

As to more lenders, or why not more, I think two reasons. One is that the onboarding process, the “sales cycle” for adding these small banks and credit unions is just lengthy. Seems as though it takes the better part of a year to get all the reviews and approvals from the institutions management. And second, I don’t think there is competitive pressure among the institutions in their market area to be first to use the Upstart platform. I could easily be wrong about this, but the platform is not necessary in order to make the personal, unsecured loans. The performance of the loans is better but are the loans a major part of the bank’s business? The usage has not hit a critical mass.

As to UPST being acquired, I don’t think any local banks or credit unions would have means or interest. It would have to be a major bank, would it not? And they could chose to use the platform and not have to make the acquisition. They could use it only for their own loans, and until the platform is proven for auto and mortgages and SMB loans, would it be worth their while? If Wells Fargo bought UPST, I don’t think B of A or Citi or Chase would use the platform.

Just my thoughts.


I would think any mega-corp that likes “information about consumers/people” would be interested in acquiring them.

What would AMZN do with that info. And combined with AFRM, suddenly they would be all over the digital finance sector. I dunno…something like that wouldn’t surprise me.

I probably lack imagination here, but how would Amazon monetize the platform. UPSt doesn’t have information, do they? They have an AI platform that analyzes the info obtained from public sources and input from potential borrowers. Amazon may very well have sufficient data already to input into the AI software, but what do they do with that. Buying Affirm makes more sense…, I think.

Regarding the other question about bank partners, I think the number of partners is increasing 50% per year. At what percentage of the banks/credit unions being partners does the rush to join up become irresistible? At the current rate of growth it would take 24 to 25 years for half of the credit unions to be partners. (71, 57, 42, 31 the last 4 quarters; 5,000 credit unions in U.S.)

I don’t know what to make of this other than long, long runway where you can compound at 125% annually for two decades to get to 50% market penetration.

As always, check the old guys math–my napkin is soggy.


1 Like


“I think the number of partners is increasing 50% per year.”

I guess that was what I thought when I started writing the post. Or maybe I thought 50% per quarter. Or… Such a pain in the rear to actually go to the original data source. Sooo much easier to just wing it.


WAFTT = Winging And Faking Thoughts Technique

I am with ya, buddy…

Dreamer ← Just a guy on the internet, and not sure the internet is real.

Family Guy Peter Griffin GIF - Family Guy Peter Griffin Of Course - Discover & Share GIFs

Is this a tiny step towards UPST’s time? KBRA report on auto loans.

“September remittance reports showed mixed credit performance across securitized prime and non-prime auto loan pools. Delinquency rates edged higher in KBRA’s prime auto loan index but improved in our non-prime auto loan index on a month-over-month (MoM) basis. Annualized net losses in KBRA’s prime auto loan index climbed 6 basis points (bps) MoM and 15 bps year-over-year (YoY) to 0.3%, while prime delinquency rates (60+ day) rose 3 bps MoM and 12 bps YoY to 0.4%. Meanwhile, annualized net losses in KBRA’s non-prime index increased 46 bps MoM and 314 bps YoY to 6.72%, while the percentage of non-prime borrowers 60+ days past due fell to 5.27%, down 3 bps MoM and up 161 bps YoY.”

Main point here (which is hardly news, but new data to ignorant me) is that the annualized net loses for prime auto loans is rising and now 0.3%, compared to non-prime rising more and up to 6.72%. Huge difference between prime and non-prime. If UPST platform can knock a few percentage points off annualized net loses should be a large TAM.

I sold all my UPST at $25.30 and will look to buy back in closer to $21, $22. At $22.55 pre-market. Market isn’t going anywhere. Maybe make money trading.



Earnings on the 8th, AMC. How are things setting up?

Stock-wise, UPST made a new 52-week low Thursday at $19.52, up to $19.93 at close of after hours trading ($19.77 regular session close). Short interest is reported at 37.6%, or 5.21 days of recent trading volume. That is 30.6 million shares short. There is another 1.9 million of Off Exchange Short Volume, whatever that means.

It seems to be an indication of retail participation in the short selling, but the increase/decrease is as important as the actual volume… ???

I would say that there is a bit of a compressed spring here so a few bucks of upward bias. Not terribly exciting.

Business-wise, what could go wrong? Upstart laid off 140 hourly employees, about 1% of headcount. These people processed loan applications. Bullish interpretation is that a greater percentage of loans are automatically processed. Bearish side says loan applications are down and UPST is controlling costs. I note that there have been no securitizations reported again for October. Meaning that the “held for sale” loans are increasing, and furthermore if the auto lending business is gaining traction, the “held for sale” is increasing even more. A case of good news is bad news, depending on the mood of Mr. Market.

The wildcard, of course, is the committed funding from the White Night. What will they say in the press release and earnings call? A combination of auto loan traction and good news on funding will launch this puppy, don’t ya think? And how much downside is there? (Same question as in May at $75 when there was a quick 67% down).

My WAG is that flat loan growth with personal loans down and auto loans up, combined with no new on capital commitment might drive UPST down to…$16…? Good auto growth and some positive news on funding could pop UPST to $28. Anyone have a Windex wipe? I don’t see a good way to play the 11/8 options. But maybe place some limit buy orders laddered down at $17, $16, $15. The short interest could dampen the downside for a day or two.

Anyone interested in UPST?


1 Like

I’ve been trying to listen to the conference call. My audio acuity is waning and there is too much domestic activity at this morning hour. I will try to post something substantive later today. I did not place laddered buy orders prior to earnings and I closed out the UPST puts last week, so for once I was correctly positioned WRT UPST.

Today I just bought the “dip” without conviction. I’ll see if I can find cause to establish an investment position.


I caught part (most?) of Q&A…the general sense I got was that the UPST mgmt knew their business and had reasons for everything, and when pressed about losing money and needing to raise money, they basically said we have $800m cash and that will last a long time…no need for raising.

Basically the analyst was trying to paint a worst-possible scenario and whichever exec handled the question did ok, I thought.

pretty clear that macro is hitting the company.
question I did not learn about (may have missed it) is whether the UPST model is still benefitting their clients…is the UPST AI engine producing better results than FICO. Crappy macro market aside…does UPST do what they say they can do, or not?

If answer is yes, and based on cash, and based on idea (hope?) that any recession isn’t too long-lasting, I see them surviving and being a growth story again down the line, when macro is healthy.

I tried for $13.99 but no joy. Bad timing. May see what tomorrow brings.
I did math and about $13.25 was about 30% which I thought would be the extreme end.


1 Like

I listened to Cathy Woods today. She is still pounding the table that deflation is around the corner, a rough recession is at hand, rates are too high and they will come down sooner than expected.

Some of what she said made sense. Some seems iffy, and I have rarely seen a macro economic prognostication make people money. It happens but the ones that are right and make money are so rare that they a literally 1 in a billion.

Having said all that, if her call is correct rates will come down. So loans held on the books at these rates will go up in value, but in the current environment, the book value of all loans held are dropping as interest rates rise and they are falling as default rates rise.

So, it is a logical expectation that UPST stock would come under pressure. However, when rates start falling and default rates start falling, then if Upstart is still holding a lot of loans on the books their asset value will rise.


1 Like

Management makes an attempt to answer that, or to pretend to answer that, or to obfuscate the issue. Their slide #9 claims they are 5 times better a predicting default than FICO. But, what is the data set? Ah, the fine print. Data from the last 10 quarters which included all the normal times, the COVID shutdown, the massive infusion of capital and the last 12 months of tightening. Of course, defaults do take some months to occur, so that data on the tightening period is not as available.

But, what can we guess? A little Pareto tells me that the default rate for q3 originated loans is going to be low. With false precision, I would guess 2.5 to 3%. Why? Because from what I heard on the call, they are approving a lot lower % of loan applications. The number of loans is down 40% and if they are chopping off the Upstart grade D and E applicants (historically 6.5 and 12.7% default rates) then you are left with 0.8, 2.4 and 4.4% historical default rates. Management commented on this saying, I think, that the loans they actually made are (will be) more profitable. Among the confusing numbers, is this Upstart Macro Index at 1.5 with a projected value of 1.7 and management are saying that they are pricing loans at 2.0. Say what? They are approving loans with expectation or risk mitigation of the macro getting worse.

Too much for me. Can’t pretend to make quantitative analysis with this mush. But can we put numbers down and see what market cap would be with a p/e of… 20? In four years there might be 150 million shares. Share price 60. Market cap 9 billion. Earnings need to by $3 per share. $450 million after tax profit. Dang, my napkin is maxed out.


1 Like

AIs are only as good as the data they are trained on. “Overfitting” can be a real issue. We will see