PM Openings Are Up 56%. PM Hires Are Down. The Posting Is for a Different Job.

Lenny Rachitsky posted the PM job openings chart from his late-2026 job market report on LinkedIn today and called it a narrative violation: open PM roles are up 56% since January 2023, ahead of every other core product role. This is a Rewrite entry. The count is right, PM hires went the other way over the same stretch, and the posting itself now describes a different job.

The short version

• PM job openings are up 56% since January 2023 in Lenny Rachitsky's chart, from TrueUp data, against 43% for software engineering, with product design down 22% and UX research down 55%.
• His line is that "Product managers are benefiting most in the AI era." Openings are ads. His own report says actual PM hiring is flat or declining, and Aakash Gupta's October report has outside PM hires down 18% since 2023 while postings rose 77%.
• The ad also changed: LLM is now the most mentioned skill in PM job descriptions, Claude is third, and evals is in the top ten.
• AI PM is 46% of open PM postings and about 7 of every 100 PM hires. In April I audited 200 PM job descriptions and 196 could have been written in 2022.
• My read is that the PM seat is winning because it absorbed drawing, research, and the eval, and that the people who benefit are the ones who can already do that job. Read the posting as a spec and check yourself against it this week.

What the chart says

The chart is titled as a change in open roles since January 2023, sourced to TrueUp, open roles at tech companies through October 2026. Product management is up 56% with 7,303 open roles. Software engineering is up 43% with 65,351. Data and analytics is up 18% with 10,307. Product design is down 22% with 1,963. UX research is down 55% with 116.

The report behind it adds that open PM roles are at their highest in nearly four years, up 75% from the March 2023 low. Lenny's point is aimed at the "AI replaces PMs" story, and on the openings data he is right. Nobody is posting 7,303 ads for a role they plan to delete.

I wrote about the design half of this report yesterday in Lenny Counts 3.8 Open PM Roles per Designer. Kill the Ratio. Count the Surfaces With a Rubric. This is the PM half, and it needs two more numbers before it means what the headline says.

Openings are ads

The same report says that actual PM hiring is flat or declining. PM's share of hiring grew, from 16% in 2021 to 21% today, and it is the only core role whose share grew. A bigger share of a hiring pool that shrank every year since 2022 is a relative win. It is not more PMs getting jobs.

Aakash Gupta put the two series side by side in his October 22 report, using TrueUp for postings and Live Data Technologies for hires. Postings up 77% from the 2023 low. Outside PM hires down 18% since 2023, with 2026 an early read, and down 45% since 2022. He calls it "a totally different trend," and he is careful about the limits: his hire count is outside hires only, so promotions and internal moves do not show, and fewer hires does not have to mean fewer PMs. Hires per exit in his data ran 0.92 in 2025, against 1.16 in 2022.

So the chart measures how many chairs are advertised. The other series measures how many got filled from outside. One went up 77% and one went down 18%, and a PM reading the chart on LinkedIn today should hold both.

The posting changed

Here is the part of Lenny's report I would put on the chart. "LLM" is now the most mentioned skill in PM job descriptions. "Claude" is third. Cursor, OpenAI, and MCP are in the top eight, and "evals" is in the top ten. Two years ago the top skills were SQL, Jira, AWS, and Python.

In April I did my own version of this, smaller and by hand. I pulled 200 PM job descriptions and scored each on one test: could this have been written in 2022? 196 could. Four named the tools, a prototype to ship in the first 90 days, and a metric the PM owns. I wrote it up in 200 PM Job Descriptions Reviewed: 98% Hire for a Dead Job.

The two are not the same measurement. Mine was a sample of 200 read line by line. His is a ranking of skill mentions across TrueUp's listings, and a term can top a ranking without being in most postings. But they point the same way six months apart. The skill list has moved to where my four were.

Gupta has the number that shows how far the hiring is behind the posting. AI PM was 2% of open PM jobs in February 2024 and is 46% today. AI PM hires are about 7 of every 100 PM hires. Nearly half the ads, about a fifteenth of the hires.

46% vs 7 in 100

Where the other roles went

Put the five bars together and my reading is that the PM bar is not rising next to the design and research bars. It is rising because of them.

A PM prototype takes four to six hours in my practice. That is the first draft of the design, made on the PM side of the line. Yesterday's post has the Figma trial that measured it. The question a researcher used to take to users, five people and one question the same week, is now something the PM or the designer runs. And the eval, the page that says what good looks like, is product's job outright, which is the argument of The Eval Is The Spec.

That is three jobs' worth of output landing on one seat. The seat got bigger, companies want more of them, and the posting was rewritten to say so. This is the AI Product Management thesis in one chart: AI collapsed the cost of the parts PMs were trained to be good at and left the parts nobody trained for. The 7,303 ads are for the second group of parts.

I can't prove that from two datasets I did not open. It is the reading that fits all three facts: openings up, hires down, skill list replaced.

Who is benefiting

The seat is. That is a different sentence from the one on LinkedIn.

A PM whose last 90 days were PRDs, stakeholder syncs, and a roadmap deck is not who the 56% is about. Those ads would not have matched that PM's resume in April and they match it less now. A PM who shipped a prototype to customers last month and owns an eval set is who every one of those ads is trying to reach, and the hire count says there are not many of them yet.

Pay says the same thing more quietly. Gupta has median PM total compensation up 14% since 2022, and about 4.5% at L3 once you hold level constant, with the share of L1 reports falling from 17% to 11%. The mix moved up. The door at the bottom got narrower, which I wrote about for CPOs in Anthropic Runs 17 Engineers per PM. The Ratio Is the Wrong Question.

So I would not tell a PM the chart means they are safe. I would tell them the chart is the best news in years for a job they can learn this quarter, and that the posting already tells them what it is.

This week

Take twenty minutes. Write the six terms from the posting down the left side of a page: LLM, Claude, Cursor, OpenAI, MCP, evals. Next to each one, write one thing you shipped with it in the last 90 days. A prototype a customer touched. An eval set that blocked a release. An agent that runs on a schedule and that somebody reads. Reading about it does not count, and neither does a course.

The filled rows are your resume for the 7,303. The blank rows are your plan for the quarter, and the Product Builder job ladder is where I would start on what each level owes. If you hire PMs, run the same page on your own posting first.

Related answer: Why are PM job openings up while PM hiring is down?

Sources: Lenny Rachitsky, LinkedIn post with the chart "PM is outpacing eng, data, and design," November 11, 2026, for the five bars and open role counts, and "State of the tech job market: late 2026," Lenny's Newsletter, November 10, 2026, takeaways 1 to 6, for the 75%, the hiring share, and the skill rankings; takeaway 7 is paid and is not cited. Aakash Gupta, "The Definitive Report on PM Compensation and Hiring in 2026," Product Growth, October 22, 2026, for postings against outside hires, hires per exit, the AI PM shares, and compensation. Every market figure is as they report it from TrueUp, Live Data Technologies, and Levels.fyi, which I did not open. The 200 job descriptions, the 196, and the four to six hours are my own.