The AI Is Doing the Work. Why Am I Paying More?

September 30, 2026 · Mike Gingras

What an Equity Trust crypto offer taught me about fees, custody, marketing, and the questions investors should ask before handing over retirement money

I have money at Equity Trust.

That matters because I am not writing this as someone who received a random advertisement and decided to complain about a company I have never used.

I am a customer.

Recently, Equity Trust emailed me about its Automated Crypto IRA, an investment option powered by Animus Technologies. Equity Trust describes the product publicly as an algorithmically managed approach to Bitcoin and Ethereum. The company also promotes zero setup and application fees for the Automated Crypto IRA. Equity Trust

Then I read the investment material.

And I started asking questions.

Start with the fees

According to the Animus investor material I received, the investment manager charges an annual management fee of 2 percent, plus a performance fee equal to 25 percent of qualifying new profits above the account's previous high water mark.

Trading costs are also passed through.

Those charges exist in addition to whatever custodial and account fees apply through Equity Trust.

That immediately raised a question for me.

Why does automated investment management cost so much?

The entire premise of automation is scalability.

A human investment manager has limits. One person can only research so many investments, watch so many markets, manage so many portfolios, and make so many decisions.

Software does not have the same limitation.

An AI system can theoretically monitor thousands of accounts at the same time. It can analyze markets while everyone is asleep. It can process information continuously. It can execute the same decision logic across hundreds or thousands of customers.

That is part of the sales pitch.

But scalability normally creates efficiency.

And efficiency normally creates an opportunity to lower cost.

So if much of the monitoring, analysis, and execution is being automated, I think investors are entitled to ask a very simple question:

Where did the savings from automation go?

Maybe there is a good answer.

Do not simply tell me that AI manages my money and expect the letters A and I to justify a premium price.

Zero to open is not the same thing as inexpensive

This is one of the more interesting lessons in financial marketing.

Equity Trust can truthfully advertise zero setup or application fees for this product while the underlying investment strategy can still carry substantial ongoing costs.

Those statements do not contradict each other.

They are simply talking about different parts of the transaction.

That is why investors need to stop asking only:

What does it cost to get in?

The better question is:

What does the entire economic relationship cost me after everyone gets paid?

Custodian fees.

Management fees.

Performance fees.

Trading costs.

Administrative fees.

Asset specific charges.

Whatever else applies to the particular account.

Government investor guidance specifically warns that fees associated with self directed IRAs can be significantly higher than fees associated with other investment accounts and can include annual fees, administrative fees, transaction fees, and asset specific charges. FINRA also cautions that fees on alternative and emerging investments can materially erode returns. Syndication

That is the number I care about.

Not the number on the front door.

Then there is the AI

The Animus material describes a sophisticated system.

It says the system processes information from more than 150,000 sources across 70 languages and evaluates several families of data before producing signals used to allocate money among Bitcoin, Ethereum, and cash.

That sounds impressive.

But the number of things an AI reads tells me almost nothing about the quality of its judgment.

An AI can consume enormous amounts of information and still reach a bad conclusion.

Humans can too.

The important questions come afterward.

Who decided which information matters?

Who decided how much each signal matters?

Who determines what constitutes excessive risk?

Who changes the model?

Who approves those changes?

What happens when the system makes a bad decision?

Does somebody perform an analysis afterward?

Can investors see why the model reduced Bitcoin, increased Ethereum, or moved into cash?

Is there an audit trail?

Is the production model actually learning continuously, or is it periodically retrained and replaced with a newly tested version?

Who has authority to override it?

Those are governance questions.

They are not sexy marketing questions.

But if software is making decisions with retirement money, I think they are more important than how many languages the software can read.

AI does not eliminate humans

This may be the most misunderstood part of artificial intelligence.

Putting AI between the customer and the decision does not eliminate human judgment.

Someone designed the system.

Someone chose the training data.

Someone decided what success looks like.

Someone established the rules.

Someone decided how much Bitcoin, Ethereum, or cash the system may own.

Someone decides when the software gets changed.

Someone decides when a new model is safe enough to manage real money.

The human did not disappear.

The human moved farther away from the customer.

That can be an improvement.

It can also make accountability harder to see.

A long track record deserves another question

The Animus material I received also describes an older Bitcoin strategy with a live record dating to December 2020.

That sounds reassuring until you keep reading.

The current generation strategy described in the same material went live in May 2026.

The company's disclosure says the older strategy and the current strategy are related but distinct and that the older results are not necessarily representative of the current strategy.

That distinction matters enormously.

An investor should know whether a performance number represents the actual system receiving their money today, an earlier version of the system, or a historical simulation.

Those are three very different things.

A five year track record for a predecessor strategy is not the same thing as a five year live track record for the strategy running your money.

The custodian is not the investment manager

Here is another distinction that deserves far more attention.

Equity Trust says plainly that it is a directed custodian and does not provide investment, tax, or legal advice. Equity Trust

That means the presence of an investment inside an Equity Trust account should not be interpreted as Equity Trust saying the investment is good.

The SEC has warned investors about exactly this misunderstanding. In a self directed IRA, the custodian generally holds and administers the assets. It generally does not evaluate the quality or legitimacy of the investment or its promoter. SEC

That is not an accusation against Equity Trust.

It is how this part of the financial system works.

And that is precisely why customers need to understand the difference between:

The company holding the retirement account.

The company marketing access to an investment.

The company managing the investment.

The technology making decisions.

And the investor ultimately accepting the risk.

Those roles can easily blur together in a customer's mind even when the legal documents keep them very separate.

This is bigger than crypto

This is actually why I think this belongs on BuiltPredictable.com.

I work on business systems.

And the same rule applies whether somebody is selling investment management, scheduling software, marketing automation, payroll software, an AI receptionist, or an AI system that trades cryptocurrency.

When a vendor tells you automation makes their product better, ask another question:

What did the automation make cheaper?

If the answer is nothing, ask why.

If AI replaced hours of human monitoring, where did that efficiency go?

If one system can now manage thousands of customers, who captured the economic benefit of that scale?

The customer?

Or the provider?

There may be an excellent explanation.

But sophisticated technology should withstand simple questions.

Technology should earn trust, not borrow it

I am not saying the Animus strategy is bad.

I do not know that.

I am not saying Equity Trust has done anything improper by offering access to it.

I have seen nothing establishing that.

I am saying that phrases like artificial intelligence, algorithmically managed, institutional technology, automated risk management, and round the clock monitoring should start a conversation.

They should not end one.

If someone wants a percentage of my assets every year, a quarter of certain profits when things go well, trading expenses when the system acts, and additional fees elsewhere in the account structure, I want to understand exactly what every party is doing.

Especially when the argument for the product is that computers are doing much of the work.

Technology can create extraordinary value.

But automation should not make us less curious about the business model behind it.

It should make us more curious.

If the machine makes the service dramatically more scalable, the customer has every right to ask who received the benefit of that scale.