The Art of Slowing Down
156: How Lack of Due Diligence Cost QIA $250M - Is Anthropic Next & Why Sovereign Wealth Funds Are Easy Targets for Overhyped Tech Startups
If you’ve been following the news, you would have seen that the Qatar Investment Authority (QIA) has invested in Anthropic’s Series F round that raised $13 billion bringing the company’s total valuation to $183 billion which is up from $61.5 billion in March 2025. Anthropic is a US-based AI company behind the Claude model. This move is part of QIA’s strategy to back cutting-edge technology firms, particularly in artificial intelligence. What interesting about this is that Anthropic CEO Dario Amodei had previously stated that taking money from Middle Eastern sovereign wealth funds and investors was “dangerous to democracy”.
This investment supposedly strengthens the QIA’s growing AI portfolio, which includes stakes in Databricks, Applied Intuition, Cresta, Turkish AI Insider, and Elon Musk’s xAI. Unfortunately, I believe this is another case of Middle Eastern Sovereign Wealth Funds being taken advantage of due to their ability and willingness to write large checks to “check boxes”. Let’s get into why.
In May 2023 it was announced that the QIA led the Series D investment round into a British no code startup called Builder.ai. Builder.ai, once considered one of the UK’s most promising tech start-ups, ultimatley collapsed after uncovering evidence of potentially fraudulent sales. An internal probe revealed that the company’s previously reported revenues were vastly overstated. Estimates for 2024 were slashed from $220 million to around $55 million, while 2023 figures were cut from $180 million to roughly $45 million.
The investigation centered on questionable sales practices involving “resellers,” especially some based in the Middle East, raising doubts about whether these entities and their reported sales were legitimate. These concerns emerged after unpaid invoices remained outstanding for extended periods. Following the findings, lenders declared a default and seized company cash, forcing Builder.ai into insolvency. Ultimately sales were exaggerated by 300%.
Builder.ai had raised over $500 million from investors, including Microsoft and the QIA, on the promise of using AI to simplify app and website creation. The QIA is reported to have lost almost $250 million on this investment. “But Pierre these guys are professionals and surely when deploying that much capital they would have done due diligence!” Unfortunately, in this case, along with many others companies prey upon Middle Eastern SWFs fear of missing out on opportunities that could provide major returns. What seems to be lost is that the point of a SWF (similarly to a family office albeit on a much larger scale) is to preserve capital and not “shoot the lights out”.
While I understand where Mohammed Al-Hardan and his team were most likely coming from here there are ways they could have done DD internally or have brought people in to assist with it given the size of the investment.
The simplest ways the QIA diligence team (or the Bain, McKinsey or BCG PIPE consultants they may have outsourced it to) could have looked for problems would be to start with GlassDoor reviews. I’ve pulled a few that I found and have included them below. Some of my favorite quotes are:
“Deceptive business practices, misleading tech to consumers and investors.”
“The product Builder sells is a complete scam… Despite their marketing claims there is absolutely no AI involved in their building process. They boast about leveraging artificial intelligence, but in reality, they outsource development to contract developers from all over the world.”
“No AI in any of the products”
“No AI in product”
“Product is also awful, delivery are far far far behind on 90% of projects and the AI element to the product is exaggerated” – Review May 2022 assuming QIA DD team should have caught this
“Limited addressable market”
“Product is not unique, there are many competitors”
“The product we sell is not driven by AI at all; the work is outsourced and there is an awful amount of code. We don’t build apps in a day; rather months or years; definitely nice PR packaging branding but nothing else.”
“Company product and follow up service is a big problem for customers.”
“Ok Pierre, however wouldn’t you agree that many GlassDoor reviews may be written by former employees with an axe to grind and may not be reflective of the actual business?”
Sure, so the next thing I would do had I been on the QIA DD team would be to do customer verification calls. In many cases the QIA would have outsourced this to Bain, McKinsey or BCG who would then use expert networks like GLG, AlphaSights or Tegus. These calls generally cost €1.200 each, however, can be valuable when you’re looks to do customer verification especially if you’re deploying millions of dollars into an investment. MBB consultancies will then bill these back to SWFs like the QIA as part of their $5.000.000+ engagement fee.
It also turns out that I am very good at this myself. Given that Builder.ai’s “Case Studies” page no longer works I used the “Wayback Machine” which captures websites at various intervals and logs them.
What’s interesting here is that it was incredibly easy to see that the majority of the customer testimonials were fake without doing checks. The first give away is that a simple Linkedin search for the individuals resulted in no bios. Ok fine, not everybody is on Linkedin. However searching via Google and reverse image searches also pulled up nothing except in some cases random Indian guys on freelance platforms like Fiverr who use the same generic pictures for headshots. I have included the screen captures from 2022, 2023 and 2024 which is what investors like the QIA would have seen during their DD process and also when they should have been conducting maintenance on the name.
Secondly if you look at the customer’s that do exist, they are very small operations, for example one is a small pub in the UK. If you’re a company doing a few million in ARR why on earth would you be highlighting a customer that spends at most a couple of thousand pounds a year on the service?
The findings above should have raised a lot of questions. My guess is that similar to the FTX due diligence debacle with Bain and Tiger Global, the QIA relied on somebody else to do their DD and the end result was them getting fucked and seeing a few hundred million dollars vaporized.
Very candidly I think something similar is going to happen with the QIA investment into Anthropic. While I do not think that Anthropic is a fraud, it seems highly unlikely that the company will ever be profitable or get to a point where SWFs like the QIA can recoup their investments. In fact, if you look at the accounting treatment that Microsoft and Google are giving Anthropic on their P&L and balance sheet, it looks like they’re writing down the value for both (for those interested especially if you’re from the QIA, ADIA, PIF etc. I’m fairly sure I have the “kill shot” on this).
The other issue with companies like Anthropic is that they are going to require an insane amount of capital to continue running and I will write a separate piece that explains this in more detail. However, the TLDR is that I am very confident that we are in an “AI Bubble” and at some point, Anthropic is going to blow up. It also seems like Anthopic is desperate to raise cash given the CEO’s flip-flop on taking money from Middle Eastern SWFs.
Another example is that in March 2025, The Abu Dhabi Investment Authority (ADIA) participated in a $200 million seed funding round for the United States-based life sciences company, Lila Sciences. This investment was intended to advance the development of Lila Sciences’ AI platform and autonomous labs. So what does this company do? According to Lila’s website:
“At Lila, we believe that science, spanning from physical science to life science and beyond, is governed by scaling principles that mirror the pathways of progress in AI. Just as larger AI models unlock emergent abilities, scaling scientific experimentation enables discoveries that remain hidden at smaller scales.
When AI turns the wheel of science, new paths are illuminated, new results brought into the world faster than could be imagined. The challenges we face, in health, sustainability, and beyond demand urgency and an acceleration of science. We cannot wait for incremental progress.
For transformative science we must transform the scientific process. To turn the incremental to the exponential, we are rethinking the scientific method from first principles.
Join us in reimagining what science can become.”
What the fuck does this even mean? I’m an unpaid intern so maybe it’s just going over my head, however if anybody can explain I’m open to listening.
As is the case with many of these AI related investments there seems to be a big red flag to me and that is employee churn. For example Lila has seen the following people leave in the past year with the average tenure appearing to be very short:
Dr. Sam Cross – Senior Principal Engineer (10 months at Lila)
Dr. Stephen GoldFless – Vice President of Life Scient (6 months at Lila)
Dr. Jaclyn Lunger – First Employee | Machine Learning Scientist
Dr. Alex Ramos – Senior Director Applied AI (7 months at Lila)
Dr. David Bellamy – Founding AI Researcher
It’s also important to look at CEO Geoffrey von Maltzahn’s track record. He seems to be involved in quite a few companies that have blown up. Notably:
Co-Founder Seres Therapeutics NASDAQ:MCRB: -98.2% all-time
Co-Founder & VP of Discovery at Axcella Health: Liquidated and dissolved 2023
Co-Founder Sienna Biopharmaceuticals: Filed for bankruptcy
You get the idea. If anybody knows Jean-Francois Couture (Head of Healthcare Private Equites Americas at the ADIA) or if he is reading this, feel free to ping me as I have some info on this one that may be helpful as you do maintenance on the name.
I understand why larger investors like Sovereign Wealth funds do not want to miss out on various opportunities however I’ve seen time and time again how they are taken advantage of. There are actually some very interesting ways for SWFs like the QIA, ADIA, PIF etc. to manage the risk on these AI investments and also hedge (with little cash outlay) many of their currently public equities investments, especially given the massive exposure they have to US names. Ultimately if there is a downturn and they hedge properly, they could do fairly well instead of getting crushed without paying massively for “downside insurance”.
I will be in Abu Dhabi in the next few weeks and if there is interest in connecting to discuss (I can easily get to Doha, Riyadh etc.) happy to provide the insights I have as an unpaid intern.
As Alfred Tong mentioned in his last write up, I will also have
