One of the most talked about phenomena across markets over the past three months is the extent to which OpenAI and Nvidia are in the process of creating a symbiotic AI ecosystem wherein the success of one company assumes, and in many cases even depends upon, the success of the others.

On some level, it makes sense: We’ll bind our fates such that you have to do what’s in my interests because your interests are synonymous with mine, and that being the case, I have no choice but to reciprocate.

The problem with that’s pretty straightforward. It takes for granted that the underlying venture and, just as importantly, prevailing assumptions about the nature of that venture, will be successful and are correct, respectively. If the venture fails to live up to expectations or if someone comes along and disrupts the market such that the assumptions upon which your ecosystem’s built are rendered invalid, everyone fails together.

With that in mind, OpenAI and Nvidia each announced new tie-ups on Monday. Both of the deals fit the mold, which is to say they’re amenable to the same skepticism which has AI bubble watchers on high alert.

Starting with OpenAI, a press release described a new “ownership stake” in Thrive Holdings, an interesting choice given that it’s an offshoot of Thrive Capital, one of OpenAI’s biggest investors.

If your question’s whether that means OpenAI’s effectively paying for a stake in Thrive’s AI incubator with Thrive’s own money (remember: money’s fungible), the answer’s “no.” Or “not exactly.”

Instead, Thrive’s giving Sam Altman a stake in exchange for access to OpenAI’s research, product and engineering teams, who’ll be “embedded” in Thrive Holdings’ portfolio companies with the goal of “establish[ing] a repeatable model that can expand across other industries.”

The jokes write themselves. Here’s one: There is indeed something “repeatable” about OpenAI’s deals. Tempting as it is, I won’t overindulge. Suffice to say we’re pretty far down the self-referential rabbit hole in the OpenAI complex, which by now is the most complicated Venn diagram you’ve ever seen.

Justified or not, this deal will be lampooned by critics, who recently began to question whether Altman’s strategy was (and still is) to make OpenAI so ubiquitous that in the event someone outcompetes it (someone like, say, Google), it’s too embedded to abandon and, perhaps, too big to fail.

Meanwhile, Jensen Huang bought $2 billion in Synopsys stock, which means Nvidia now owns a near 3% stake in a company that makes design software for chip companies like — wait for it — Nvidia.

Bloomberg parroted Huang in claiming the deal “is not circular in nature,” but… well, I guess that depends on your definition of “circular.”

Synopsys isn’t obligated to buy any of Nvidia’s chips as part of the deal, and Huang was keen to suggest Synopsys’s other customers will benefit from the arrangement.

Still, if the point is to more deeply embed Huang’s technology in the chip-design process, it’s hard to understand how that’s not “circular.” Indeed, that seems like the very definition of circular.

The language employed to describe these tie-ups is now just as recursive as the deals themselves. Commenting on the Nvidia partnership Monday, Synopsys chief Sassine Ghazi said, “Together we will re-engineer engineering.”

And rest assured, whenever that re-engineered engineering itself needs re-engineering, Huang will be right there to engineer a solution.