Who doesn’t love the idea of Arnold and Danny as twins, it was a great movie and concept, but in the real world we need twins to be more similar than Arnold and Danny.
Funny side note, I married a twin, I mean an identical twin. There are pictures of my wife and her sister as kids that their mom can not actually say who is who. Two birthdays ago, they bought each other the exact same birthday card, that is the power of TWINS.
But for Venture VINE we are here to learn or share about Venture | Innovation | Networking | Ecosystem, so that obviously includes technology and one of the most interesting technologies today that gets very little press is Digital Twins.
AWS has this definition on their website
“A digital twin is a virtual model of a physical object. It spans the object’s lifecycle and uses real-time data sent from sensors on the object to simulate the behavior and monitor operations.”
I like it but would say there are Digital Twins of non-physical objects as well which I will cover down below. This diagram from Toobler shows a good example of Physical Digital Twins, in a car design operation.
You don’t need a manufacturing background to see how helpful a Digital Twin can be in cars or airplanes. You can see how the object will behave in the real world without the real world consequences, you can easily see this benefit in a car or plane.
Benefits
Better R&D
I love R&D which is one reason I am excited to be moving some of my efforts to Deep Tech from Fintech. R&D to me is the lifeblood of any good technology business. One of my favorite quotes about R&D is from Steve Ballmer, a well known sales and cheerleader but not a techie by any means:
“The lifeblood of our business is that R&D spend. “ Steve Ballmer of Microsoft
Even a lifelong sales leader knows that you need R&D to invent and perfect new solutions.
I have often (maybe too often) made the point that Nvidia invested $5-$7B in R&D each year from 2019 to 2024 and the outcome for Nvidia was an outsized increase in enterprise valuation:
Nvidia Valuation in 2020 was $14 Billion
Nvidia Valuation in 2024 is $3 Trillion
Nvidia invested about $20 Billion in R&D and got a multi trillion dollar return on that investment.
Ok, we can all agree R&D is a value driver but not everyone can do that type of R&D, some folks need to outsource R&D, which is how I look at CVC or Corporate Venture Capital for Banks and other non-technical organizations.
(please do NOT tell me your bank is a Tech Company with a Bank License)
Better Efficiency
Every company today, even startups are trying to become more efficient as the “grow at all costs” days of 2019 are well behind us and investors of all types require Profit above Growth.
In the hardware world it is hard to over emphasize how efficient a digital twin is compared to a physical Proof of Concept or Prototype. Yes there are advances in prototyping with 3D printing and I don’t want to ignore that but for now I will focus on Digital Twins.
What If Scenario
A Digital Twin can operate “What If” scenarios that can predict outcomes based on various inputs.
What if the car we are designing heads into a head wind of 90 MPH in the real world, we need to know what would happen BEFORE it happens. You could reasonably argue that a wind tunnel would be great for that test. Fair enough after we spend millions to build a prototype of the car and hope we got it mostly right. And I am not sure but I think Wind Tunnels run in the billions of dollars these days.
With a Digital Twin, we can see how the car would react BEFORE we so much as put clay to frame and that will save time and money. If I am honest it does look fun to build a full scale car in clay, but I also think it is incredibly time consuming and expensive.
What If Analysis and Scenario Planning sounds similar but the key differences are in Scope, Time Horizon and Complexity.
- Scope: Scenario planning looks at a variety of interconnected external and internal factors. What-if analysis focuses on the impact of isolated variables or small combinations of them.
- Time Horizon: Scenario planning tends to be long-term, whereas what-if is usually short-to-medium term.
- Complexity: Scenario planning is more holistic and multifactorial, whereas what-if is more direct and tactical.
The above are what ChatGPT said when I asked for the differences between What If Analysis and Scenario Planning. Thanks ChatGPT, well said.
Note: I will always point out when I use ChatGPT directly, but mostly I use it for formulating ideas not directly having ChatGPT write for me.
Banks
Since I have been in and around Fintech and Banking for a long time, I wanna talk about how a Digital Twin could help a bank. I have no doubt a new design for an ATM could benefit from a Digital Twin strategy to test how it would work in the real world.
Full disclosure I am working with a very cool startup called Delfi that does something that is best viewed as a Digital Twin of a bank’s balance sheet.
Having recently been going deeper into Deep Tech, (pun intended) when the Delfi founder mentioned the Digital Twin concept I immediately got it. Not many folks in banking think about Digital Twins. I should note my good friend ChatGPT said this when I asked about Digital Twins for non-physical goods:
“That’s an interesting angle! Digital twins have traditionally been viewed in the context of physical objects — like manufacturing processes or infrastructure — but applying the concept to finance models is a logical leap, especially for banks where predictive accuracy is critical.” ChatGPT October 18, 2024
Again thanks ChatGPT that is exactly my point.
Get Dion F. Lisle’s stories in your inbox
Join Medium for free to get updates from this writer.
I have long worked in and around banks but have never been a banker. That sounds oxymoronic but it means I worked for Citibank but never opened an account or underwrote a loan for anyone. Same at Huntington Bank where I was in Corporate Ventures and not responsible for ensuring our loan ratios met the Fed’s requirements. I have clearly never been near a bank’s balance sheet or the associated ALCO (Asset Liability Committee) often overseen by the bank’s CFO.
But what I do with banks is help them innovate.
Silicon Valley Bank (SVB)
Now if you know the sad story of Silicon Valley Bank (SVB) as of March 2023, they were the 16th largest bank in the USA. Also, SVB was a well respected brand in the Startup and Innovation Economy.
It has been estimated that nearly 50% of VC wire transfers would go to SVB for their startup clients.
How could such a large and respected bank fail so badly?
From the post-mortem report issued by the Fed on SVB:
“SVB’s highly concentrated business model catering overwhelmingly to the venture capital community, and its reliance on uninsured deposits which left the bank “acutely exposed to rising interest rates” amid a slowdown in the tech sector.”
That sounds like the most innocuous way a bank could spiral, but it happened. Honestly you could use the popular “you had one job” meme for this, SVB is a bank and interest rates are a bank thing.
“But you had ONE JOB”
Yes, there were other factors at play but the net was that SVB didn’t manage interest rate risk. It did not help that they did not have a Chief Risk Officer (CRO) for over 9 months.
This is where a digital twin could have helped predict and likely prevented the worst case scenario that SVB found themselves in.
Here is how Delfi would have addressed the sad SVB outcome before it became inevitable.
“Delfi first develops a joint simulation of both the entire interest rate yield curve and other key driver rates as well as a dynamic evolution of the bank’s balance sheet. Delfi can assess the sensitivity of a bank’s balance sheet to potential interest rate volatility.”
This is the ultimate use of a Digital Twin. Sadly for SVB shareholders Delfi was not operational at the time of SVB’s meltdown, but in hindsight it is clear that Delfi’s sophisticated scenario planning would have at the very least mitigated the damage of SVB’s poor policies and decision making.
The key point from the above diagram which makes it look much simpler than it is as the yellow AI\ML Optimization box is the “black box” where the magic happens allowing for the green box output of an actual recommended Hedging Action. The key for every banker I have met for Delfi is that there is an action plan at the end not simply more data to analyze. Bankers are looking for plans not inputs. More data without answers is not helpful.
We can look with the benefit of hindsight and see the following dramatic potential. Delfi’s Digital Twin of SVB had the following outcomes:
Delfi simulations suggest that SVB’s performance would have been a gain of +$3.3bn instead of losses of -$2.0bn, and the volatility of its performance would have been reduced by 95%!
The hypothetical effective cost of this hedging program was calculated to be about $150 million, relatively small given the high liquidity of the futures contracts Delfi uses to construct its risk management strategies.
Imagine any CFO at a bank being able to see this swing of $5.3 Billion. It is a bit like my obsession with Nvidia and their investment in R&D resulting in trillions of enterprise value.
Why wouldn’t a bank CFO want a Digital Twin Balance Sheet to share with the board and the ALCO (Asset Liability Committee) that are tasked with ensuring the bank gets it right. In software development there is the concept of a Sandbox, a place you can play with code and not break the real world IT systems that run your company.
I love the idea of having a Digital Twin for a bank where I can test all the crazy inputs that might impact my bank.
- Interest rates rise by a basis point, what should I do?
- Inflation figures come out lower than expected, what should I do?
- The USD weakens against the Euro, now what?
I would much rather do all of this in the Digital Twin than in the real world.
Phew
A somewhat happy ending for SVB is that they are doing fantastic after they were picked up by First Citizens Bank and are back to hosting events in the valley and beyond. Plus my son is working at SVB and loves it.
But since meeting Delfi I realize the outcome for SVB could have been very different if they had run a Digital Twin Balance Sheet through the various scenarios that were likely for the bank in 2022 and 2023.
Wine
Since this blog is Venture VINE, I will now go over to the Wine portion of the blog. Yes there is a Digital Twin for wines… of course there is.
You can store your wine’s digital twin on the Blockchain run by Winechain and then trade or take possession of those wines. This may sound like a stretch but wine like all expensive collectibles runs into forgeries that cost collectors millions of dollars per year.
https://winechain.co/how-it-works
Thanks and If you want to find out more about getting your bank a Digital Twin Balance Sheet let me know and I can introduce you to the Delfi founders, both PhDs from Harvard and really nice guys, oh also they are brothers, imagine founding a cool fintech startup with your brother, that would be so cool.
Cheers from Sonoma County, come enjoy the Fall Colors.
