Potential Upstart Cloud

By acquisition and explosive growth Upstart now dominates my portfolio. Due to the weight it has in my portfolio I read everything I come across that relates to Upstart, either directly or tangentially.

My broker, Fidelity, always posts articles on my account summary page. The articles are divided into two categories, those that have a direct relationship to my positions. This group often includes new Edgar postings, so strictly speaking, the listings are not all authored news “articles”.

The second category is for more general financial news. The following link was for an article in this second category. You won’t see Upstart named in this article, but it may as well have been. Thanks to Jon Wayne’s incredible research on Upstart I am aware of the fact that they quite some time ago got the equivalent of an endorsement from the CFPB which allays some of the potential worries that might arise from this article. Nevertheless, regulations can be blunt instruments. I would like to hear from other folks in order to learn what they may think of the issues raised in this article as related to Upstart. I hope that you needn’t be a Fidelity customer in order to access it.



I’m also on Fidelity and, frankly, a lot of what they post as “relevant” really is not (I learned the hard way by sharing some of those articles here…).

From the article you posted:

"Since early last year, Google has also blocked new AI features analyzing emotions, fearing cultural insensitivity, while Microsoft ( MSFT )
restricted software mimicking voices and IBM ( IBM ) rejected a client request for an advanced facial-recognition system."

UPST is data driven, PRECISELY to avoid cultural insensitivity and bias that might arise under more traditional/conventional ways of evaluating loan applications.

UPST is not analyzing emotions, or using facial recognition.

I do understand what it’s like to have a large position and feel the need to read about the company…and, following this board trains us to look at the numbers.

I am long UPST (~11% of portfolio)


I think the concern Google had is that historical data can reflect past discrimination, and thus perpetuate it. But then I wonder, perhaps the quality candidates Upstart finds in the sub-prime pool are people who have been traditionally discriminated against.

It’s definitely something I want to be aware of.

"Google said it was presented with its money-lending quandary last September when a financial services company figured AI could assess people’s creditworthiness better than other methods.

However, its ethics committee of about 20 managers, social scientists and engineers who review potential deals unanimously voted against the project at an October meeting, Pizzo Frey said.

The AI system would need to learn from past data and patterns, the committee concluded, and thus risked repeating discriminatory practices from around the world against people of color and other marginalized groups."



Thanks for bringing this article to my attention.

From article:

Indeed some AI advances may simply be on hold until companies can counter ethical risks without dedicating enormous engineering resources.
After Google Cloud turned down the request for custom financial AI last October, the Lemonaid committee told the sales team that the unit aims to start developing credit-related applications someday.
First, research into combating unfair biases must catch up with Google Cloud’s ambitions to increase financial inclusion through the “highly sensitive” technology, it said in the policy circulated to staff.
“Until that time, we are not in a position to deploy solutions.”

I think this reinforces the idea of a “regulatory” or “fear of regulatory” moat.

The idea that “any big bank can copy Upstart” is just not true - AI/ML models and its associated data are not a commodity, it can indeed be highly differentiated as UPST has proven - it’s really really hard to do AI well in the world of credit.

As we know Goldman Sachs has trouble in its foray into credit underwriting (https://discussion.fool.com/unfair-ai-underwriting-for-net39s-co…), so too Google has realized the same difficulties lie ahead and smartly shelved their attempt for the time being.

Upstart has a long term first mover regulatory advantage in AI underwriting, with its CFPB NAL and proactively working with SBPC and NAACP on fairness in its models.


Since various kinds of discrimination are systemic in the culture, that means they work their way into data as well as into more subjective criteria.

Upstart’s use of AI for lending, therefore, put them under the watchful eye of both the NAACP and the Student Borrower Protection Center last summer, which resulted in a demand letter to adjust their algorithms. That letter is here: https://www.naacpldf.org/wp-content/uploads/Upstart-Letter-R….

By December of 2020, both organizations announced they had entered into an agreement with Upstart to have a two-year monitoring period to make sure any issues were fixed. That agreement is here: https://protectborrowers.org/naacpldf-sbpc-upstart-agreement….

That, in addition to the No Action Letter from the CFPB, tells me that Upstart is responsive to all such concerns and is willing to do whatever it takes to become an ethical lender.


Hi, first time poster as I have been reading this board for the last 90 days.

It’s like drinking from a fire hose the volume of investing wisdom and fact-based reporting of companies.

Hopefully this will meet the high standards of this board. The best posts I have read don’t tell you what to do but give you the information so you can come to your own logical conclusion.

About me: I am a retired (2 years ago) IT tech that held many positions at small companies to a large investment bank and most recently at one of the 3 leading credit rating agencies. In my last 10 years I concentrated on Information Security with certifications in (CCSP) Certified Cloud Security Professional (CCSP) and Certified Information security professional CISSP.

What I wanted to add to this Upstart discussion is the fact that Upstart is a quasi-Credit rating agency. I think in the short term Upstart is facing mostly sunny skies when it comes to using their AI for loan approval. The reason is Upstart was granted a Consumer Financial Protection Bureau Issues No Action Letter to Facilitate the Use of Artificial Intelligence for Pricing and Underwriting Loans.


This is a 3-year allowance to use their Upstart uses artificial intelligence (AI) techniques and alternative data in its model that started November 30th, 2020.

So I think in the short term ~2 years Upstart will be fine but longer term it is something to keep an eye on. Although I do believe there are tailwinds for ways to properly assess credit risk of populations that aren’t currently in the credit rating agency’s system.

As a side note, at the credit rating agency I did a stint as third-party risk assessment of vendors for Information Security and Compliance. There was a constant search for vendors that could assess risks using non-traditional metrics. If a company can solve that problem (which I believe Upstart has), it a win-win for everyone.

One last note is another thing to keep an eye out longer term is Upstarts ability comply with the Equal Credit Opportunity Act. I am not a subject matter expert but am familiar with some of the pain points when it comes to proving compliance. They have at least 2 years to figure this out which I’m confident they can do. It’s not as simple as proving non-discrimination of loan approvals. Auditors also want proof that the mailers they send out aren’t being held back on the basis of where someone lives etc.

Upstart is my largest position.


This topic keeps popping up here.
Yes, Upstart is FinTech … but FinTech is not Upstart. Same for Upstart and AI …

Lumping Upstart with all Fintech companies and all AI applications and the related racial biases is not accurately looking at the whole picture.

Upstart AI and ML models are scrubbing current data and actions looking for correlations to the ability to pay back a loan or creditworthiness that we humans can’t see. By looking for NEW predictive data points to include in the model vs basing decisions on past (potentially racially biased data) the models become less racially biased.

Do AI models have the potential to be racially biased? They most certainly do, especially when using historical data points that were biased to begin with. But when this is monitored and addressed, they can open up the arena to those who have traditionally been under-served and discriminated against.

Dave Girouard, Upstart CEO stated at a congressional hearing in May “The concern that the use of AI in credit decisioning could replicate or even amplify human bias is well-founded,” and “It’s very reasonable to be concerned about AI because it’s a very sophisticated system and it does have the potential to introduce bias or unfairness. Our answer to that concern is very rigorous testing.”

Upstart also submits the results of the internal testing for bias every quarter to the Consumer Financial Protection Bureau (CFPB).

The data has proven this to be successful to a certain level. Upstart’s AI lending model (per UPST in 2020):

  • approved 30% more Black borrowers than “a traditional model” with interest rates that were on average 11% lower
  • approved 27.2% more Hispanic borrowers with an average of 10.5% lower rates

This in no way proves that racial biases don’t exist in the UPST models, in fact UPST believes biases can and do occur, thus the effort to monitor and address. It does show that UPST models are less biased than traditional models and are heading in the right direction.

I agree with Jabbo’s take on the demand letter from the NAACP and SBPC in July 2020 and have read the report released in Feb 2020 which led to the letter. At the time of the report, Upstart had already demonstrated that alternative credit data they use had helped them approve 27% more loans than traditional models since late 2017, and that the APR on those loans is on average 16% lower.

Upstart agreed that while their models were already more inclusive than more traditional models (including those used by other Fintech companies), they would increase the scrutiny of the models even further because one of the elements of their objectives to remove racial biases from lending.

Upstart stopped using SAT scores and never zip codes in their models – historical data points proven to be racially biased and used in FICO scores (which are narrow in scope and inherently backward looking). Upstart doesn’t disclose what correlations the data scrubbing has unearthed nor the data points in their constantly improving models, but other companies in China using AI/ML for personal loans for years have shared that use less obvious predictive data points such as how fast the user enters data on the phone, level of phone battery charge, as well as timing of the repayments relative to the due date – data points that I would find hard to think are racially biased.

I believe Girouard has installed a corporate culture to address concerns head on vs deflecting and defense which is so often prevalent in the lending institutions. And I believe this culture will serve them (and the stock price) well going forward.


Who is:

  • Long UPST (though late to the party)
  • Pleased with the Giants performance against the Dodgers this past weekend
  • Starting to think Fidelity stock news is becoming as relevant as Yahoo news

I have a small position in UPST. I am not an IT guy, nor a finance/lending guy. I based it solely on the numbers. So I found your post very informative. I don’t understand the AI that much, nor lending laws, nor how the two have to work together. Or at least I didn’t. Your outline of the situation (and the subsequent response from another poster) was useful.

I was unaware of the 2 yr time horizon, and so will be watching that. Many thanks.

1poorguy (long UPST, less than 5% allocation)

“Nevertheless, regulations can be blunt instruments.” Indeed they can. I am quite concerned about the potential downside for my Upstart investment. The regulatory agencies and NAACP don’t much care about genuine, legitimate risk assessment. They care about outcomes. If the pool of approved loan applicants does not match their expectations, regardless of the reasons and despite the best possible AI algorithms, my investment is potentially in trouble. The good news is it appears I have a couple of years before that trouble would likely begin.


Thanks all for your well informed replies.

This board is an incredible gift. We are all indebted to Saul for having drawn together this community. I have benefitted greatly from it as I am sure many other people have.

I failed to mention that the root of my concern about regulations was born from my 30 years in IT at The Boeing Company. I have been reluctant to post the name of my former employer due to the numerous NDAs I signed during my tenure there, but I retired more than 10 years ago and at this time virtually everything I was sworn to protect by now has become public domain or irrelevant.

Aerospace is a highly regulated industry. I have witnessed seemingly minor infractions result in massive penalties. I am all too aware that nothing anyone posts will eliminate the potential for Upstart to be found in violation of some regulation irrespective of the effort and resources they may have expended in order to avert such a finding.

Nevertheless, having read these posts does very much help to put my mind at rest. I hope it serves others as well.

Thanks again

very long Upstart.