Original title: Charts of the Week: DExit . . . real or feigned?
Original source: a16z New Media
Original translation: Deep Tide TechFlow
Deep Tide introduction: a16z's chart weekly report this issue covers four topics, each worthy of a separate article: the decline in AI costs triggering the Jevons Paradox, the true scale of capital expenditure by tech giants, Kalshi's prediction market outperforming professional forecasters, and the comprehensive delay of life milestones for the American 30-year-old population. With solid data sources and a calm and restrained perspective, it serves as a high-quality reference for understanding the intersection of current technological and macro trends.
DExit… Is it a real trend or an illusion?
Delaware remains the preferred state for company registration in the United States, but this position is quietly loosening:
According to data from Ramp, Delaware's share of new company registrations has been declining since 2023, with a drop of about 10% by the third quarter of 2025.
History does not simply repeat, but it often rhymes… maybe.
Delaware has not always been the holy land for business registration.
About a century ago, Delaware replaced New Jersey—formerly known as the "Mother of Trusts"—as the preferred state for company registration. New Jersey lost its advantage because then-Governor Woodrow Wilson sought to curb "corporate misconduct", severely deteriorating the business environment in New Jersey. Delaware's corporate law was originally modeled after New Jersey law from before the Wilson era, making it naturally welcoming to businesses fleeing, and subsequently collaborated with Delaware's Court of Chancery, taking nearly 100 years to establish a reputation as a mature and fair venue for resolving disputes between businesses and investors.
However, what took a century to build began to shake within a few years. Right or wrong, Delaware's Court of Chancery has adopted a more lenient stance on shareholder lawsuits in recent years (particularly in several high-profile cases, including but not limited to Tesla), and companies have started genuinely moving their registrations elsewhere. Goodnight, good luck, Delaware.
This is at least the mainstream narrative, but other data indicates a more complex situation.
First of all, even the founding myth of Delaware itself is not entirely accurate.
It wasn't until the 1980s (about 60 years after Governor Wilson's tenure) that Delaware truly surpassed New Jersey to become the number one state for business registrations in the United States:
New Jersey's dominance lasted much longer than the mainstream narrative suggests. The catalyst for Delaware's eventual surpassing was likely its passage of a series of laws regarding board responsibilities that made it particularly attractive to public companies, along with a self-reinforcing network effect that formed its own inertia.
Secondly, regardless of what high-profile public companies (and companies in Ramp's data) are doing, Delaware as a whole still appears to be performing well, if not exceptionally well:
According to data from Harvard Law School's Corporate Governance Forum, Delaware's share of total U.S. businesses has actually increased significantly from the end of 2024 to 2025.
In fact, if you're looking for a clear "DExit" case, it is probably this one, and it has nothing to do with Tesla, but rather involves a specific company structure:
Wyoming's LLCs have seen rapid growth since around 2015.
Why? This is likely related to specific asset protection and privacy provisions in Wyoming's LLC laws; the state promotes this corporate structure as a "Cowboy Cocktail."
In summary, the point is not to say that DExit is not happening (because at least some data suggests that it indeed is— even just a few high-profile companies moving out is significant), but the actual situation is undoubtedly more complex than the mainstream narrative presents.
The reality is that Delaware still enjoys the advantage of being the default option, not to mention all the associated network effects, which are difficult to shake.
We have released an earlier version of this chart before, but as data increases, the effect is becoming increasingly remarkable.
Token costs are down, Token consumption is up:
Since the beginning of this year, the price of paid Tokens has dropped from about $0.90 per million Tokens to $0.50, while the number of processed Tokens has nearly doubled, from about 6,000 to 12,000.
This is a typical Jevons effect. The cheaper AI becomes, the more we use it. It's delightful.
Remember when someone said that when newer, better GPUs are released, the old GPUs would be abandoned?
The facts don't seem to be so:
According to data from Silicon Data, the rental prices of Nvidia's H100 and A100 have both increased this year.
The market shows no signs of a surplus in computation power; rather it appears that we haven't even scratched the surface of existing demand.
This comparison is not a perfect analogy, but if history can provide a reference, we may need some time before we can truly "see" what an AI-driven "economy" looks like:
From Faraday and Henry's initial discussions on current to the actual explosion of industrial productivity in the first half of the 20th century, it took about 100 years.
Since the 1820s, the pace of technological iterations has indeed accelerated, but the variables involved in a platform-level transformation are still extremely diverse.
Roy Amara has a famous quote: "We often overestimate the changes that will occur in the next two years and underestimate the changes that will occur in the next ten years."
Capital expenditure, viewed through the axes
Let's look at this timeless set of data: the scale of AI capital expenditures is substantial.
Consider the following comparisons:
AI capital expenditure in 2026 is expected to approach the total net new loans of all U.S. banks in 2025:
Capital expenditures are about 33% higher than the total revenue from U.S. corporate income tax, roughly three times the total from tariffs:
Capital expenditures are about six times the total military budget of any G7 member country outside the U.S.:
So, yes, the scale of capital expenditures is indeed substantial.
Kalshi enters macro forecasting
Federal Reserve researchers believe that prediction markets are quite decent.
At least on one metric, Kalshi's predictions for the federal funds rate have already outperformed professional forecasting institutions:
For the prediction of the federal funds rate 150 days later (i.e., after three FOMC meetings), the mean absolute error of Kalshi is very close to that of professional forecasting institutions. However, unlike surveys that provide modal path snapshots every six weeks, Kalshi offers continuously updated full probability distributions… We found that the median and mode predictions from Kalshi have maintained a perfect record the day before the FOMC meeting, which is statistically significant improvement compared to predictions from federal funds futures.
In other words, although all forecasters start from a similar point, Kalshi's continuously updated predictions optimize over time and ultimately achieve a perfect prediction record the day before the rate is officially announced. Additionally, Kalshi's performance also surpasses that of futures market predictions.
Kalshi's advantage is not limited to the federal funds rate. As highlighted by Federal Reserve researchers, due to macro indicators like inflation, growth, and unemployment rates lacking other options markets, Kalshi is the only place that can offer "high-frequency, continuously updated, and rich probability distributions" as a "benchmark" to reflect the "public" judgment on the trends of these economic indicators.
Sounds quite significant.
Delay of adulthood
This is a thought-provoking chart, accompanied by (some) commentary:
The percentage of 30-year-olds achieving major life milestones has been on a rather steep decline since at least the 1980s.
Among 30-year-olds, fewer and fewer are:
Living independently;
Ever married;
Living with children;
Owning their own homes.
The only exception is college enrollment rates—the percentage of 30-year-olds with bachelor’s degrees has nearly doubled since 1995.
So, is going to college worth it?
Milestones? More like a millstone around the neck, right?!
Maybe so, maybe not, but the feeling of "buyer regret" seems to be permeating the air.
免责声明：本文章仅代表作者个人观点，不代表本平台的立场和观点。本文章仅供信息分享，不构成对任何人的任何投资建议。用户与作者之间的任何争议，与本平台无关。如网页中刊载的文章或图片涉及侵权，请提供相关的权利证明和身份证明发送邮件到support@aicoin.com，本平台相关工作人员将会进行核查。
