One Year of Podcasting Paid $377. Here's What the Math Actually Says.

Here is what one year of podcasting actually paid: $377.56.

That is Bandrew Scott's take-home after producing 40 episodes, paying $915.84 in hosting and production costs, and accounting for taxes. His hourly rate worked out to $1.89—less than a third of minimum wage in most U.S. states. Bandrew published these numbers publicly, which is rare in an industry where income claims skew heavily toward the people selling courses about income claims.

The more uncomfortable fact is that Bandrew is not a failure by industry standards. He is roughly average. According to Buzzsprout's platform data from over 116,000 active shows, the median podcast gets 28 downloads in its first week after release. At a standard CPM of $25 with two ad slots, that is $1.40 per episode—$67 a year before costs. The AI-powered future of audio is real. The passive income version of it, for most people, is not.

So here is the question worth answering: is there a version of this opportunity that works for someone with your background, your expertise, and your available hours—or is "start a podcast with AI" just the new "open an Etsy shop"?

The Ad Revenue Math Is Broken for New Entrants

Before blaming the creator, blame the structure. The CPM ad model does not fail independent podcasters because they lack skill or work ethic. It fails them because nearly half of all U.S. podcast ad spend—49%, according to Magellan AI's Q4 2025 data—flows to just 500 shows. Out of roughly 4.7 million indexed podcasts, that is the top 0.01% capturing half the money.

New entrants are not competing for a smaller slice of the ad pie. They are competing for crumbs the top 500 did not want.

The math makes this concrete. Apply the same 28-download median to a weekly show over a full year, and gross ad revenue lands at $67.20—before hosting fees, recording software, or a single hour of your time. The U.S. podcast advertising market is projected to approach $2.6 billion by 2026, according to the IAB and PwC. That number is real. It simply does not flow to the median creator.

This is where the opportunity splits into two fundamentally different paths. Path A is building a media business where the podcast is the product—a model that requires top 1% download numbers to generate meaningful ad revenue. Path B is using a podcast as a trust-building engine that sells something else: consulting, a service, a course, a client relationship. Path A is statistically improbable for new entrants. Path B can generate real ROI at 200 listeners if those 200 listeners are exactly the right people.

For a marketing consultant, an HR advisor, a financial planner, or any professional who charges for expertise—Path B is the accessible entry point. A niche B2B podcast reaching 500 qualified buyers can command a $1,500 flat-fee sponsorship from a vendor targeting that exact audience, bypassing CPM math entirely. That same $1,500 on 500 downloads equals a $3,000 effective CPM—120 times the programmatic rate.

The medium is not the problem. The business model is.

Two Insiders, Identical Tools, Opposite Decisions

The AI-generated podcast factory model has generated most of the recent headlines, and one company has actually made it work at scale. Jeanine Wright, who previously served as COO at podcast network Wondery, built Inception Point AI with eight employees producing 3,000 episodes per week across 5,000 shows. Production cost: $1 per episode. Profitability threshold: 20 listeners. Revenue in 2025: $770,000. Her logic is direct—"The price is now so inexpensive that you can take a lot of risks."

The factory model is real. The math works in aggregate, using programmatic ad fill across thousands of niche shows covering everything from local weather to celebrity biographies.

I think that people who are still referring to all AI-generated content as AI slop are probably lazy luddites.
by Jeanine Wright, CEO of Inception Point AI

Then there is Jeff Umbro, CEO of The Podglomerate, a premium podcast services firm with deep industry relationships and a full production team. Umbro built the identical pipeline—LLM-generated scripts, synthetic voices, automated RSS publishing. He had every technical advantage Wright had. Then he shelved it entirely. "We actually built this whole model and thought about doing this and we never released it," he said. His reason was not technical failure. It was brand safety risk, thin content quality, and the judgment that the output was devoid of anything human.

Two credentialed insiders. Same tools. Opposite strategic conclusions.

What separates them matters for anyone evaluating this path. Wright operates at a scale where 5,000 shows absorb platform risk across a portfolio—if 500 get filtered as spam, 4,500 remain. For someone betting a career pivot on a single show, that risk calculus is entirely different. Meanwhile, the platform window for undisclosed AI content is closing fast. Spotify removed 75 million AI-generated spammy tracks in 2025 and introduced new spam filtering systems. Apple Podcasts now mandates prominent AI disclosure in both audio and metadata for any show where AI generates a material portion of the content.

Umbro's shelved project is the more instructive case for most readers. He had the tools, the team, and the industry connections—and he still said no. The median outcome of the factory model is not $770,000. It is a show with 15 listeners that earns $0.30 in programmatic fill before being filtered as noise.

What Actually Works for an Individual Operator

With the factory model ruled out and the CPM model structurally inaccessible, what does a viable path actually look like in concrete operational terms?

Start with survival, because it is the primary variable and more controllable than most guides admit. Ninety percent of podcasts do not survive past episode three. Of the ten percent that do, ninety percent never reach episode twenty. According to the Independent Podcaster Report 2025, which surveyed 558 creators, only 15% of independent podcasters make any money at all—and 46% of those who do monetize only break even. The top challenge cited by creators who quit is not production quality. It is audience growth, named by 72% of respondents.

Reaching episode twenty-one puts a show in the top one percent of all podcasts ever created by persistence alone.

We actually built this whole model and thought about doing this and we never released it.
by Jeff Umbro, CEO of The Podglomerate

The survival problem is solvable through structure. Pre-record five to six episodes before launch. Commit to twenty-five episodes before evaluating performance. Measure success using leading indicators—email subscribers, qualified inbound conversations, referrals—rather than CPM revenue. The shows that survive long enough to compound are almost universally the ones that defined their backend product before recording episode one.

The cost structure is more manageable than the guru pitch suggests, and more significant than the "just grab your phone" crowd admits. A mid-tier solo setup runs about $19 a month for hosting, $24 for a remote recording platform, and $23 for editing software—roughly $66 in monthly cash outlay. The real cost is time: five to ten hours per episode for a quality weekly show. At a conservative freelance billing rate, that is $1,200 or more per month in labor foregone. Anyone evaluating the ROI of a podcast needs to run that number honestly against the expected revenue timeline.

The video dimension is now non-negotiable. According to Edison Research's Infinite Dial 2026, 57% of Americans have both listened to and watched a podcast, and YouTube has become the primary discovery engine for the medium. Audio RSS feeds are excellent for retention—deep listening, parasocial loyalty, long-term subscriber relationships. But audio-only distribution is not a growth strategy. Recording in video for YouTube clips handles discovery. Distributing audio via RSS handles retention. Running both in parallel is the baseline for any new show launched in 2026.

Bandrew's $1.89 Is a Map, Not a Verdict

Return to Bandrew Scott's $1.89 per hour—not as a warning about podcasting, but as a diagnostic tool with a specific use case.

That number is the exact cost of podcasting without a defined backend product, without an audience identified before episode one, and without a clear answer to what happens after a listener finishes the first episode and wants more. It is the cost of treating a podcast as the business rather than as the most efficient trust-building machine available to a professional with genuine niche expertise.

The medium itself is not the problem. One hundred thirty million Americans listen to podcasts every week as of 2026, according to Edison Research. The trust mechanics of the format are genuinely powerful—listeners who follow a host through twenty episodes know that person's thinking, values, and judgment better than they know most colleagues. None of that changes the math for a show without a clear answer to: "Why would my first hundred listeners ever pay me anything?"

Before buying a microphone, write one paragraph answering this: Who are the first hundred people you want listening, what problem do they share, and what will you offer them that they cannot get from a Google search? If you can write that paragraph clearly, podcasting is a delivery mechanism worth evaluating seriously. If you cannot, the medium is not your obstacle. The business model is. Start there.

The passive income version of podcasting is not coming. The trust-engine version already exists—but only for operators who know what they are building trust toward before they record a word.

Explore Further