Upstart: An enigma or a phoenix

With the markets closed today, hoping that everyone is having some respite to unwind and enjoy the weekend and reflect.

I think my post which tried to explain a couple of points about Upstart got lost amid the noise in the thread that has finally ended.

Here’s how I felt in 2021 after everyone was talking about Upstart ( including the “What does Upstart do” guy on CNBC)…

Cons: As the Upstart stock rose exponentially in 2021, I felt having missed out and also got intrigued and curious. Though I was tempted a number of times but I was always in doubt about investing in it as it seemed to reside outside my comfort zone. The primary reason was that it was non SaaS. Additionally, I don’t know anything about lending AND the proclamation to disrupt sub-prime lending with ML and AI. So, watching the rise of Upstart last year, I started thinking really hard if their AI was really so exceptional that Credit Unions and Banks will rapidly embrace it to start lending while minimizing the risk of default. Since, I’ve worked very closely with some highly regarded data scientists in developing ML models and which are being used by services catering to millions of users everyday, I had my inhibitions. So out of curiosity, I started having conversations with my friends who are experts in the field and leading AI/ML efforts in some of the biggest organizations ( some of them are from the same company that the Upstart founders worked). The conversations didn’t give me enough confidence and also left me a little cautious. I do admire Paul Gu and believe that Upstart’s AI development is in great hands but at the same time could not vouch that their AI could indeed achieve all it intended to do without hiccups at it’s current state. There is no doubt that in the future most of the manual data driven insights and human decision making will transition to AI to prevent fraud, anomaly detection and identifying risks. All of those aspects make the value proposition for an AI driven lending service to be marketable and see mass adoption and integration BUT I couldn’t see that happening right then and fast enough. So, with whatever I understood, I decided to give Upstart a pass and probably missed some great profits.

Pros: Now to the other side of the story… Remember that Upstart is NOT a company that was born yesterday; it was founded in 2012 and has some amazing folks leading their engineering and business with a vision set to disrupt lending in personal loans, auto loans and maybe housing someday. Specially, since the inception of Covid, every credit union and bank are trying or are being forced to make the digital experience better for customers. They are also trying hard to have lifetime returning customers. So, it is inevitable that in the lending space those underwriting decisions will be taken over by AI in due course. That would lead to a huge increase in efficiency and ROI. What I’m not fully sure today is, if “Upstart” would be that number one provider. But when you think about the head start that Upstart has, they do they have a very big advantage ; each day they are increasing the volume of data on their platform (akin to the the billions of driving miles data advantage that Tesla has over competitors). And AI and ML are all about having the right and high quality data to get better accuracy. So, that’s where I think Upstart’s real advantage lies. The models will improve incrementally and there will be a time when the accuracy of the models predicting credit worthiness with minimum risk will be fine tuned and optimum. And at that inflection point most manual processes across all credit unions and banks will be forced to adopt such a solution to stay competitive and in business. And it does appear that Upstart has the best chance to be that top AI disruptor in lending.

So with all that known and unknown information, here’s what I did yesterday! I established a small position in Upstart ( about 2%) and had bought a few long term call options (that expire in 2023) some days back. It’s my smallest position behind ZoomInfo which I finally got into a couple of days back. Time will prove if these decisions with Upstart have been right or wrong but for now I can sleep well and that 2% wager won’t kill my portfolio even if it were to go to zero :).

So, that’s enough of Upstart! Like Saul my top conviction names like Datadog, Zscaler and Snowflake are what I’m focussed on right now due to their track record and predictability of SaaS ( although Snowflake is usage based, I think their usage is going to grow exponentially!). The three names in my Security bucket ( Zscaler (ZS), SentinelOne (S) and Crowdstrike ( CRWD) was already overweight making up around 50% of my portfolio, so I trimmed some SentinelOne to add to ZoomInfo and Upstart. I think Security will have a lot of tailwind in 2022 and beyond; since hackers out there already have Ransomware as a service (RaaS) :slight_smile: and someone needs to keep organizations protected 24x7.

Cheers!

ronjonb (@ronjonbSaaS on twitter)

And cheerleader of Datadog (DDOG) since it’s IPO :slight_smile:

P.S. Please don’t clutter this board which is a treasure trove for many of us with offensive, emotional and off-topic posts. They do no good to anyone and make it hard to find the ones worth reading and above all creates a lot of work for the board managers who are serving in a selfless way.

69 Likes

There’s a small error in my earlier post…

The three names in my Security bucket ( Zscaler (ZS), SentinelOne (S) and Crowdstrike ( CRWD) was already overweight making up around 50% of my portfolio, so I trimmed some SentinelOne to add to ZoomInfo and Upstart. I think Security will have a lot of tailwind in 2022 and beyond; since hackers out there already have Ransomware as a service (RaaS) :slight_smile: and someone needs to keep organizations protected 24x7.

That should be 40% and NOT 50% as the bucket before I trimmed SentinelOne comprised of Zscaler: 20%, SentinelOne: 15% and CrowdStrike: 5%.

Now SentinelOne and CrowdStrike are at 2:1 instead of 3:1 and the total allocation in the Security bucket is 35% of my portfolio. Zscaler is at 20% because they have redefined network security and I don’t see any viable competitor.

I have 2x more in SentinelOne compared to CrowdStrike not really based on technical advantages but rather the probability of SentinelOne being a slightly better investment opportunity at this point based on their revenue growth potential. Tech-wise, I still like CrowdStrike more ( as a complete platform ) and with the CISA endorsement CrowdStrike has a big opportunity ahead. With Security becoming integral and spending getting bigger, I want to own both names as I feel both of them will have a fair share of that market.

A long time ago when CrowdStrike was a huge position for me > 20% ( which I rarely do) I had written about CrowdStrike’s extremely lightweight agent with this example of how easy it was to deploy and scale. I still believe that’s one of CrowdStrike’s unique selling point.

"…CRWD’s agent/sensor that is installed on end-points is extremely lightweight and unobtrusive…there’s no UI, no pop-ups, no reboots, and all updates are performed silently and automatically. Requires less than 20 MB of disk space, less than 10 MB memory and less than 1% CPU when active. And installs in a few minutes. There are no controllers to be installed, configured, updated or maintained and no on-premises equipment.

I believe at least in one customer’s case, CRWDs sensors were deployed to over 70K nodes via silent installs in about 2 hours and no helpdesk calls!.."

That was way back in 2020 and I think the story hasn’t changed much. Here’s a reference to that post: https://discussion.fool.com/thanks-for-the-write-up-ethan-here39…

Cheers!

ronjonb

8 Likes

From my post…

"…Pros: Now to the other side of the story… Remember that Upstart is NOT a company that was born yesterday; it was founded in 2012 and has some amazing folks leading their engineering and business with a vision set to disrupt lending in personal loans, auto loans and maybe housing someday. Specially, since the inception of Covid, every credit union and bank are trying or are being forced to make the digital experience better for customers. They are also trying hard to have lifetime returning customers. So, it is inevitable that in the lending space those underwriting decisions will be taken over by AI in due course. That would lead to a huge increase in efficiency and ROI. What I’m not fully sure today is, if “Upstart” would be that number one provider. But when you think about the head start that Upstart has, they do they have a very big advantage ; each day they are increasing the volume of data on their platform (akin to the the billions of driving miles data advantage that Tesla has over competitors). And AI and ML are all about having the right and high quality data to get better accuracy. So, that’s where I think Upstart’s real advantage lies. The models will improve incrementally and there will be a time when the accuracy of the models predicting credit worthiness with minimum risk will be fine tuned and optimum. And at that inflection point most manual processes across all credit unions and banks will be forced to adopt such a solution to stay competitive and in business. And it does appear that Upstart has the best chance to be that top AI disruptor in lending…

And from Dave Girouard’s statement it looks like the above surmise has a great chance of materializing!

“With triple-digit growth and record profits, Q4 was an exceptional finish to a breakout year for Upstart. 2021 will be remembered as the year AI lending came to the forefront, kicking off the most impactful transformation of credit in decades.”

All this isn’t too shabby :grinning:

UPST FY2022 Guidance

  • Revenue of approximately $1.4 billion
  • Contribution margin of approximately 45%
  • Adjusted EBITDA margin or approximately 17%
  • Auto transaction volume of approximately $1.5 billion

I’m inclined to add to my Upstart position as an exception ( along with Affirm) even if I’m not too comfortable investing in non SaaS. And it won’t be my top level high conviction positions like Datadog, Snowflake, Zscaler or MongoDB.

Cheers!

ronjonb ( https://twitter.com/ronjonbSaaS)

15 Likes