Money Stuff: Malaysian Fryer Oil Arbitrage

You can make jet fuel from palm oil. It doesn’t even need to be fresh. You can take palm oil, put it in a deep fryer, use it to make french fries, reuse it again and again until it becomes gross and the fries taste bad, and then take the used oil and sell it to a refiner to make jet fuel. Also works with olive oil, soybean oil, lots of cooking oils. This is a very nice fact about the world! You can get jet fuel without (1) drilling for oil or (2) cutting down forests to create dedicated agricultural land just for the jet fuel. You use the existing agricultural land to make cooking oil, you cook with it until it’s no longer good, and then when you are done with it you get an extra bonus use as jet fuel.

In 2025, this extra bonus use is especially valuable, because energy companies and airlines, at least in Europe, want to demonstrate that they are environmentally sustainable. So finding an environmentally sustainable way to make jet fuel is worth a lot of money to them. Therefore they will pay a lot of money for used cooking oil (or “UCO”). 

You could imagine decomposing the value of used cooking oil into (1) its energy value plus (2) its environmental value. That is, a European refiner will pay $X for a gallon of vegetable oil that can be turned into jet fuel, and will pay $Y for the ability to say “we have reduced our carbon impact by reusing cooking oil,” so the total price of a gallon of used cooking oil, on the international market, is $(X + Y).

Do you see the problem here? A refiner would pay $X for any cooking oil, but $(X + Y) specifically for used cooking oil. The energy content of the oil is a real measurable intrinsic fact about the oil, and there is a market for it. The fact that the oil was used is a much fuzzier fact, but there is a market for that too.

It turns out that restaurants, street food stalls and home cooks in Malaysia — which is “among the world’s leading suppliers of both UCO and virgin palm oil” — will pay less for fresh cooking oil than the international market will pay for used cooking oil. Fresh cooking oil is more useful to cooks than used cooking oil (it tastes better), but it is less useful to refiners and airlines than used cooking oil (it doesn’t reduce their carbon impact). Also fresh cooking oil is subsidized by the government in Malaysia: “Subsidised cooking oil sells for RM2.50 per kg versus the UCO trading price of up to RM4.50 per kg.” So if you run a restaurant, you can buy fresh cooking oil for about $0.60 (USD), use it to fry food a few times, and then sell it to a refiner for $1, which is a nice little subsidy for the difficult, risky, low-margin business of running a restaurant.

But if you don’t run a restaurant, you can buy fresh cooking oil for $0.60, not use it to fry food any times, and then say “oh yeah we totally used this oil” and sell it to a refiner for $1. The refiner probably isn’t going to taste it. This is in a sense less efficient than cooking with the oil first, but it is also easier to scale, and it seems like a simpler, safer, higher-margin business than actually doing the cooking.

Obviously it is worse for the environment: Instead of reusing a product that was previously used for cooking, you are creating incremental demand for palm oil (and incremental deforestation, etc.). You shouldn’t try to pass off fresh cooking oil as used cooking oil. But if the price of used cooking oil is higher than the price of fresh cooking oil, someone will.

Here’s an investigation from the Straits Times and Climate Home News into the sustainable aviation fuel (SAF) market:

As SAF producers scramble for limited raw materials to meet new blending quotas in Europe and growing demand elsewhere, barely used and virgin palm oil is being passed off as UCO to traders that supply fuel companies, experts and industry operators told us. Palm oil that is not considered waste is not permitted under European Union rules for SAF because of its links to deforestation. …

A source at a leading Malaysian UCO supplier to companies including Repsol told The Straits Times that some UCO collectors and restaurants are committing fraud by providing oil that does not qualify as used, although it is difficult to prove.

In Malaysia, which is among the world’s leading suppliers of both UCO and virgin palm oil, government-subsidised cooking oil is cheaper than UCO – providing a clear incentive for fraud. …

The restaurants, street stalls, households and factories from which the UCO is pooled self-declare the origin and authenticity of their contributions. Aside from ad-hoc spot checks and sampling, there is no way of knowing that all of these providers are telling the truth.

“The opportunity, or incidents, of fraud is very high,” said Vasu R Vasuthewan, the former Malaysia head for the ISCC.

Malaysian authorities recently uncovered criminal syndicates that had pocketed thousands of dollars a day by getting hold of large amounts of subsidised cooking oil, mixing it in with UCO, and then selling it on to industrial UCO traders.

Industry sources told Climate Home and The Straits Times that many households and restaurants are motivated to replace cooking oil after a single use – contrary to standard practice – and then sell it on as UCO. Cooking oil is considered waste when it is no longer fit for frying – generally after being used between three and five times.

I guess the upside is that all the fried food in Malaysia tastes really fresh, but that’s not really the point of the sustainable aviation fuel regime.

You could have a crude model of acqui-hiring that goes like this:

1. You are very good at some important job that some big tech company really needs done. Obviously in 2025 that job is “AI stuff.”
2. The big tech company, however, has some fairly arbitrary maximum amount of money it can pay employees, and your value is higher than that maximum. If you go to the chief executive officer of the company and say “you know and I know that if I come to work for you, I will increase your profits by $1 billion a year, so I would like you to pay me a $500 million signing bonus and a salary of $50 million a month,” she will say “you are right, and you’d be worth it at that price, but I just can’t do it. My board would never let me. Other employees would be demoralized by seeing someone get paid that much. I think that price is fair, but it’s impossible.”
3. You start a startup. The startup’s only employee is you, its business model is “I will do some AI stuff, whatever,” and you raise a little bit of venture capital from (1) your friends and/or (2) you’re doing AI in 2025, if you say “SoftBank” three times in a mirror they’ll deposit $100 million in your bank account.
4. You call the big tech CEO back and say “I know that there is a cap on your salary budget, but surely that cap does not apply to your mergers-and-acquisitions budget. I’ll come work for you for $400,000 a year, but you have to buy out my startup for $2 billion.” And she says “yes that’s a good deal, I will immediately shutter your startup, but getting you for $2 billion will increase my profits by $1 billion a year and is worth it.”
5. You get most of the $2 billion, though your friends and/or SoftBank, who invested in your startup (to make it look credible), get a cut (to make the acquisition look credible).

I don’t think that this is the main form of acqui-hire; most deals are probably people who really wanted to start startups, rather than this sort of M&A-for-salary arbitrage. But in the AI hiring arms race of 2025, where people are apparently getting nine-figure offers from Mark Zuckerberg, the way to get eleven figures is by having an AI company to sell to Zuckerberg.

And then you can go up one level of abstraction and think “what if I had, like, twenty AI companies to sell to Mark Zuckerberg when he wanted to hire me?” The Information reported last week:

Meta Platforms is in advanced talks to hire the prominent artificial intelligence investors Nat Friedman and Daniel Gross to help lead its AI efforts, according to a person familiar with the discussions. As part of those talks, Meta is in discussions about partially buying out Friedman and Gross’ venture capital fund, NFDG, which holds stakes in top AI startups and is worth billions of dollars on paper, the person said.

If the talks are successful, Gross would leave Safe Superintelligence, which he co-founded with former OpenAI chief scientist Ilya Sutskever last year. At Meta, Gross is expected to work mostly on AI products while Friedman’s remit is expected to be broader. Both Gross and Friedman are expected to work closely with Meta CEO Mark Zuckerberg and Scale AI CEO Alexandr Wang, whose hiring by Meta was finalized last week in a $14.3 billion deal.

As part of the talks, Meta is discussing buying out a substantial portion of NFDG’s holdings and cashing out the fund’s limited partners in the process, the person familiar with the discussions said. The social media giant will have minority stakes in the startups that NFDG has invested in, which could include SSI, but it will not get information about and control over these startups, the person said.

Although financial terms of the discussions couldn’t be learned, a partial buyout of the fund could cost more than $1 billion.

The point is that if you are a prominent AI researcher, you want to be ready for the day when Mark Zuckerberg approaches you. You want to be ready to say “okay but it’ll cost you,” and to have an itemized bill to hand him.

Elsewhere, here is a funny Wall Street Journal story about Mark Zuckerberg’s personal efforts to hire every AI researcher:

Mark Zuckerberg is spending his days firing off emails and WhatsApp messages to the sharpest minds in artificial intelligence in a frenzied effort to play catch-up. He has personally reached out to hundreds of researchers, scientists, infrastructure engineers, product stars and entrepreneurs to try to get them to join a new Superintelligence lab he’s putting together.

Some of the people who have received the messages were so surprised they didn’t believe it was really Zuckerberg. One person assumed it was a hoax and didn’t respond for several days.

Also: We talked last week about how Meta is going around offering OpenAI researchers signing bonuses of “your grandchildren will never need to have jobs,” and how Sam Altman is tut-tutting that those offers risk creating a culture at Meta that is all about the money rather than the work. But, to be fair, if Meta is going around offering OpenAI researchers that kind of money, it’s probably creating a culture at OpenAI that is all about the money, because, you know, that money is out there. The researchers can’t unsee the money. And so:

Andrew Bosworth, Meta's chief technology officer ... said in an interview with CNBC's “Closing Bell: Overtime” on Friday that Altman “neglected to mention that he's countering those offers.”

Right, I mean, imagine! You work at OpenAI, where you are paid $400,000 plus the satisfaction of a job well done. Mark Zuckerberg calls you and is like “I’ll pay you $100 million to come here.” And, what, you say no? You go back to your normal job? Come on! You walk into Sam Altman’s office and you get in line behind all the other researchers, and when you get to the front of the line you say “Meta, 100,” and he says “SoftBank” three times and a bag labeled “$100 million” appears on his desk, and he hands it to you and you go back to work, and the next person on line does the same thing. Imagine if the world of AI labs was divided into (1) the ones that were all about the work and (2) the ones that paid tons of money! Who would want to work at the ones that were all about the work? 

Classically, when a company did an initial public offering, it paid its investment bankers 7% of the amount it raised. That is no longer always true, and big companies doing big IPOs generally pay less than that, but 7% is the traditional number and still what a lot of smaller companies pay. [1] Fees for mergers and acquisitions are less standardized, but have a similar structure: If you are selling your company, you will probably pay your bankers a percentage of the deal value, and that percentage could be several percentage points if your company is small or under 1% if it is large. [2]

But as a rough rule of thumb, it seems approximately correct to say that, if you run a small company and you want to engage an investment bank to do some sort of important, life-changing transaction, you will probably pay the bank something on the order of a 5% commission. And if you run a giant company and you want that, (1) you will find the transaction less life-changing and (2) you will pay less. (Also, a giant company will probably hire several banks, and the fees will be split among them.)

There is some market segmentation here: The investment banks that do deals for small companies are often not the investment banks that do deals for large companies. But there is an important source of overlap: Many big fancy investment banks that do giant deals for giant companies will also intensively cover certain small companies, and will compete hard to do deals for them. Specifically they will cover the small companies that they think will become giant companies. At least some venture-backed startups will get early investment banker attention out of proportion to their size, because the bankers are hoping to lead their giant IPOs one day.

The bankers covering those startups are looking for a long-term payday, not a short-term one. If you help a teeny company raise $10 million, you can charge them $500,000 in fees, which is not nothing but doesn’t really move the needle for a big bank. The point is to be well positioned for when the company goes public and raises $1 billion in an IPO, or gets acquired for $10 billion, and you can charge tens of millions of dollars of fees. You’ll do the small deal, and probably (not necessarily!) even bill for it, but you’re not there for the $500,000. You’re making a longer-term bet.

Here is one way to formalize the longer-term bet. You do the small deal for the small company, but you sign an engagement letter that focuses on the long term. “Instead of  paying me 5% now,” you say, “let’s just sign a contract saying that you’ll pay me 5% of the deal value when you eventually do a merger farther down the line.” “Hmm,” thinks the company, “that’s a good deal. Paying 5% later costs me nothing now!”

You might notice, though, that paying 5% later will cost the company more, later, than paying 0.5% later. The going rate for the deal now, as a small company, is 5%; the going rate for a bigger deal later, in the state of the world where the company is very big, might be 0.5%. If the company agrees to pay 5% later, it is implicitly betting that it will not become a very big company. And if you, as the the banker, agree to take 5% later, you are betting that one day the company will be very big. And then it will do a $20 billion merger and you will collect a $1 billion fee. And people will say “that is not the market rate for a $20 billion merger?!?” And you will say “but it was the market rate when I signed the deal.”

Also you’ll get sued, but still, possibly worth it! At FT Alphaville, Sujeet Indap has a story on “the M&A banker who might be sitting on a more than $5bn deal fee.” He’s Steve McLaughlin, the founder of Financial Technology Partners, who signs engagement letters like this: “In exchange for providing to emerging fintech companies what it says are valuable connections and advice, and in some instances capital, the firm extracts high fee rates and ironclad contracts that can extend decades.” Various emerging fintechs (1) sign the deals, (2) emerge and then (3) try to get out of them. Indap quotes from an FT Partners court filing in a legal dispute with Circle Internet Group Inc., the stablecoin company:

FT Partners has been a pioneer in offering its clients more varied and flexible fee structures than investment banks typically are willing to.

FT Partners’ clients frequently agree to a long-term engagement agreement and/or a fee structure that recognises the substantial risk that FT Partners is taking by providing that, in the event that the company eventually completes a Company Sale at a valuation far higher than its valuation at the time of entering into the long-term engagement agreement, FT Partners’ fee for the Company Sale will typically approach or even exceed 5-20% of the aggregate consideration or incremental aggregate consideration in the transaction.

Circle understood that to secure FT Partners’ highly sought-after services, it would need to agree to terms that provided FT Partners with significant upside.

Circle’s position is “heads, we win; tails, you lose.” Circle secured FT Partners’ assistance with an ironclad guarantee that FT Partners would be fairly compensated, including that, in the unlikely event that Circle entered into a multi-billion dollar Company Sale, FT Partners would receive a fee of roughly 9%-10% of that amount. But once Circle achieved a massive valuation, it greedily sought to nullify the parties’ arrangement so it would not have to pay fees to FT Partners going forward.

Indap points out that “A tenth of Circle’s current $57.8bn market cap comes to nearly $5.8bn,” a pretty good M&A fee. 

Obviously:

Apple was sued on Friday by shareholders in a proposed securities fraud class action that accused it of downplaying how long it needed to integrate advanced artificial intelligence into its Siri voice assistant, hurting iPhone sales and its stock price.

The complaint covers shareholders who suffered potentially hundreds of billions of dollars of losses in the year ending June 9, when Apple introduced several features and aesthetic improvements for its products but kept AI changes modest. …

Shareholders led by Eric Tucker said that at its June 2024 Worldwide Developers Conference, Apple led them to believe AI would be a key driver of iPhone 16 devices, when it launched Apple Intelligence to make Siri more powerful and user-friendly. …

Shareholders said the truth began to emerge on March 7 when Apple delayed some Siri upgrades to 2026, and continued through this year's Worldwide Developers Conference on June 9 when Apple's assessment of its AI progress disappointed analysts.

We have talked about the general case of “is AI securities fraud?” The point is that there is a lot of uncertainty about the development of AI and its disruptive effects, and uncertainty creates the conditions for securities fraud lawsuits: There are a lot of opportunities for disappointing surprises. If a public company says things like “we expect our business to grow 2% next year,” and then AI rapidly undermines demand for its products: securities fraud lawsuit. If a company says “we expect to deploy AI to make our products better,” and then it’s harder than they thought: securities fraud lawsuit.

Young Investor Demand for Alternative Assets Is Reshaping Wall Street's Playbook. Compass Sues Zillow as Fight Over Private Home Listings Heats Up. Regional bank stablecoin. Germany and Italy pressed to bring $245bn of gold home from US. FICO to Offer Credit Scores Including Buy-Now, Pay-Later Data. ‘Kirklandisation’ of Big Law pushes firms to launch salaried partnerships. Chinese Stocks and American Exchanges Head for a Breakup. Bank of New York Mellon Approached Northern Trust to Discuss Potential Merger. Donohoe Says He Plans to Keep Irish Bank Bonus Cap in Place. Revolut chief in line for Musk-style payday at $150bn valuation. Why the serial CEO has fallen out of fashion. Applebee’s and IHOP Plan to Introduce AI in Restaurants. “It would get us to the point where we were struggling to make payroll for our 50-person company because he wanted that Hamptons house.” Bear Whose Head Was Stuck for Two Years Is Freed.

If you'd like to get Money Stuff in handy email form, right in your inbox, please subscribe at this link. Or you can subscribe to Money Stuff and other great Bloomberg newsletters here. Thanks!