BACK TO INSIGHTS
FIELD REPORT · NO. 04/UNITED STATES · SEARCH SIGNAL/2026.04

The founder wave has a sound. We heard it.

For two decades, the searches that map American economic distress moved in tight, familiar patterns. In the last six months, four of them broke out of those patterns by 20× to 74×, and the breakout signature is not “recession.” It is something we have never seen before. This is the data behind why DFX exists in the form it does, why we built it in six months, and why it had to be built now.

DFX INTELLIGENCE RESEARCH

Authored from internal signal probes across 40 economy, employment, and finance-coded queries.

SOURCE · GOOGLE TRENDS · UNITED STATES · JAN 2004 — APR 2026

01
74×
"ai replace my job"
02
56×
"automate my job" / "ai layoffs"
03
20×
"wind down business"
04
> 2008
"help with mortgage" peak
01 / THE SIGNAL

Four queries broke history.

We probed forty candidate search terms across economy, employment, and finance themes, all measured against a 2015-2023 baseline, all U.S., all from January 2004. Four cleared a 4× cutoff with their all-time peak inside the last six months. Together they describe a single phenomenon. None of them are classic recession queries.

OUTLIER 01
"ai replace my job"
RATIO
74×
0255075100200420082012201620202024PEAK · 100
PEAK
Mar 2026
BASELINE
0.0
RECENT 6MO
74 / 100
OUTLIER 02
"automate my job"
RATIO
56×
0255075100200420082012201620202024PEAK · 100
PEAK
Nov 2025
BASELINE
0.0
RECENT 6MO
56 / 100
OUTLIER 03
"ai layoffs"
RATIO
56×
0255075100200420082012201620202024PEAK · 100
PEAK
Mar 2026
BASELINE
0.6
RECENT 6MO
56 / 100
OUTLIER 04
"wind down business"
RATIO
20×
0255075100200420082012201620202024PEAK · 100
PEAK
Mar 2026
BASELINE
3.6
RECENT 6MO
72 / 100

People are not searching “unemployment benefits.” They are searching whether the machine is coming for them.

02 / THE PARALLEL WAVE

The other half of the signal: founders.

The same probe surfaced a second wave running in parallel. As people ask whether their job is about to disappear, they are also asking how to start something. “Build startup,” “finance startup,” and “change business model” all broke out of two-decade baselines. And the most economically meaningful search of all, “help with mortgage,” just exceeded its 2008 financial-crisis peak.

02550751002004201020162022PEAK · 100
"build startup"6× baseline
02550751002004201020162022PEAK · 100
"finance startup"Mar 2026 peak
02550751002004201020162022PEAK · 100
"change business model"Feb 2026 peak
02550751002004201020162022PEAK · 100
"replace employees"Feb 2026 peak
02550751002004201020162022PEAK · 100
"close business"Mar 2026 peak
02550751002004201020162022PEAK · 100
"help with mortgage"> 2008 peak

This is not a recession. It is a workforce reset, and the founder wave that always follows.

03 / READ THE PATTERN

What did not spike is the tell.

We tested every classic distress query in the same probe. Not one of them broke out. The terms that defined 2008 and 2020 are flat. The terms defining 2025-2026 are about agency: who controls the replacement, who builds the next thing, who exits cleanly.

DID NOT BREAK OUT

The 2008 playbook is flat.

“Unemployment benefits.” “Apply for food stamps.” “Loan modification.” “Stop foreclosure.” “Chapter 7 bankruptcy.” Every one of them peaked in the financial crisis or the COVID shock and is currently at or below its 2015-2023 baseline.

Searchers in 2026 are not asking the government for help. They are asking themselves what to build next, and whether they should wind down what they have to do it.

PROBE RESULTS
40 TERMS · 2004-2026
40
Terms tested
4
Cleared 4× cutoff
7
Near-miss (2.5-3.5×)
0
Classic recession terms broke out
04 / CORROBORATING SIGNALS

The search data is not alone.

Four public datasets, none of them Google Trends, all pointing the same direction. These are the canonical series labor economists and policy researchers track, and every one of them is available monthly from a named source.

01 / BLS JOLTSPUBLIC · MONTHLY

Openings-to-unemployed compression.

The ratio of job openings to unemployed workers peaked in 2022 and has compressed steadily since, even as the headline unemployment rate stayed low. The quits rate has fallen in step. Workers are not being fired in a 2008 shape, but the bargaining slack is gone. The JOLTS release is public and monthly — pull it and judge for yourself.

02 / Challenger, Gray & ChristmasPUBLIC · MONTHLY

The first AI-attributed layoffs in the monthly report.

Challenger's monthly job cut report began citing AI as a stated employer reason for layoffs in 2023, a category that did not exist in 2022. By 2024 it had a permanent row. The report is public, monthly, and widely quoted by BLS economists. The category only moves one direction.

03 / Census Business Formation StatisticsPUBLIC · MONTHLY

A sustained record in new business applications.

New business applications broke their all-time record in 2020, then stayed above the pre-pandemic trend line every quarter since. This is the cleanest public indicator of the founder wave the search data describes, and it is published monthly by the Census Bureau.

04 / Kauffman Indicators of EntrepreneurshipPUBLIC · MONTHLY

Opportunity-share of new entrepreneurs.

Kauffman tracks the share of new entrepreneurs who started their business because they saw an opportunity, versus those who started because they needed income. The share historically moves against economic distress. The current cycle is reading as a mix: both opportunity and necessity rising at the same time. That is the fingerprint of a workforce reset, not a recession.

05 / WHY DFX, WHY NOW

We saw the wave. We built for it in six months.

DFX Intelligence is not a one-tool product. It is the bundled toolkit that the next million displaced operators, people pivoting out of laid-off jobs, founders raising on a closing window, fund managers moving capital before the consensus catches up, will need to do the work that used to require a team of fifteen and a six-figure stack.

PILLAR I01

One platform, not fifteen tools.

The post-layoff founder will not pay for fourteen SaaS subscriptions to do what one operator could do with one system. We bundled the relationship graph, the signal engine, and the decision layer into a single sealed surface. Investor mapping, donor intelligence, deal diligence, capital pathway, oppo research, network resolution. One login. One bill. One source of judgment.

Bundled · Sealed · Operator-Owned
PILLAR II02

Six months to powerful. Not eighteen.

The wave is not waiting. We compressed what would have been a two-year roadmap into a six-month build by treating AI as engineering leverage, not feature decoration. Every module, from Property X-Ray to Donor Signal to the Stealth runtime, ships against a real customer in weeks, not quarters. Speed is the moat the incumbents cannot copy.

6-month build · 7 modules · 0 idle quarters
PILLAR III03

Built for the people the wave displaces.

The four outlier searches name our customer. Someone whose role is being automated. Someone winding down a business that no longer works. Someone deciding whether to build the next thing instead. DFX gives that operator the same intelligence layer the largest funds and the most sophisticated campaigns are using, without a team of fifteen, without a six-week onboarding, without a $40k invoice.

Operator-first · Sub-hour setup

For the people who can’t afford to find out last.

DFX Intelligence is the operating layer between public-data chaos and the next decision a serious operator has to make. Built for the people moving capital. Built for the people building something new. Built, deliberately, in the six months before the wave the data above describes hits its peak.

The signal is in. The system is live. The window is open.