30
posted ago by Narg ago by Narg +30 / -0

Of the many aspects and possibilities associated with the rise of AI, the ECONOMIC realities of AI at the current time, will be more consequential in the next few years than most believe.

JMO.


https://www.zerohedge.com/news/2026-06-09/tokenomics-equals-panic

Tokenomics represents the panic borne out of the realization that you spent a fortune developing AI yet there is insufficient demand to generate a positive ROI. It’s an attempt to socialize the losses later this year when OpenAI and Anthropic come public combined with the need for the Trump Administration to show GDP expansion into the mid-term elections. It’s where political expediency meets financial necessity. I don’t plan to fight it.

Investments in Information Processing Equipment: Contribution to GDP Growth

(Chart showing "Investments in Information Processing Equipment: Contribution to GDP Growth")

Similar charts are available for software, data centers, and R&D. Machiavelli taught us that the end justifies the means. I suspect that we should be prepared for these market sectors to rally through year-end despite not making fundamental sense. Whether it’s stories of government investment in the frontier-model companies, mysterious studies that show potential positive ROI, or changing the accounting for model-training to capitalize costs as long-lived assets like electric utilities.

What is “Tokenomics”

I love the smell of napalm in the morning – Colonel Kilgore (Robert Duvall) Apocalypse Now

An AI token is a word and words are the basic building blocks of Large Language Models or LLM’s. “Tokenomics” is simply a top-down incentive system to drive software developers to maximize their use of tokens or units of LLM capacity. It’s an artificial attempt to force Jevon’s Paradox which simply states that increasing the efficiency of a new technology or resource often increases the overall consumption of that technology/resource.

The top tech companies are betting the ranch on forcing AI onto the software developers out of fear of being left behind. Companies such as Microsoft, Amazon, Uber, and Meta are creating internal incentive systems and outright directions for software developers to utilize internal AI tools.

It’s a last-ditch effort to make AI relevant and a “Hail Mary” pass to increase the revenues of frontier-model companies such as Anthropic and OpenAI ahead of IPO’s later this year. These companies are burning capital like it’s gasoline yet they need something to support the insane valuations “assigned” to them.

And it’s pretty clear that these “assigned” valuations are at the heart of the recent rally in AI stocks. One example is the 81% increase in Google’s earnings. But if you look closely, roughly half of Google’s earnings were non-operating, consisting of rising equity valuations on non-marketable equity securities such as Anthropic. Amazon reported similar results.

Is this a sign of success? Or is it a desperate attempt to create the impression that the frontier-model companies are proving to be good investments for the top tech companies instead of capital destroyers.

You can’t argue with the successful moves in the stock market but you can question their sustainability. My personal opinion is that we’re witnessing the “Swan Song” of the technology companies even as they hit absurd valuations in their last show.

I don’t believe it is sustainable because there is still no credible evidence that using AI results in a positive return on investment apart from some bottoms-up applications. When the fundamentals ultimately assert themselves, someone is going to be on the receiving end of a napalm bomb. The only question is whether Wall Street can socialize the losses ahead of time by selling OpenAI and Anthropic to the public. Regrettably, I believe OpenAI, Anthropic, and SpaceX will be jammed into index funds before the end of the year.

Damned If You Do…

The hasty stroke goes oft astray – JRR Tolkien

Major companies such as Meta and Amazon have created incentive systems for software developers to “token max” which basically incentivizes developers to waste as many “tokens” as possible while performing their jobs. Uber used up their 2026 budget for AI tokens within the first four months of this year. To date, there is zero evidence that maximizing tokens offers a positive return on investment despite numerous studies.

The problem for the tech industry is that in order to make Anthropic and OpenAI look good before the IPO’s, they have to embrace “Tokenomics”, pay a fortune in AI usage, which will negatively impact their own earnings. It’s not a coincidence that Microsoft, which owns 49% of OpenAI, told developers that they can no longer use Claude (Anthropic), which was formerly cheap, now that they are being billed for each token used. Those funds have to be kept in-house.

The tech industry is facing an open-ended cost problem with the adoption of token-based pricing models.

But if the industry does not embrace “Tokenomics”, there is no way for OpenAI and Anthropic to report anything other than disastrous financials ahead of their IPO’s.

(more)