AI Is Shrinking Entry-Level Salaries While Grads Still Expect 2022 Money

The Number That Should Shock Everyone

$23,847.

That’s the average gap between what the Class of 2026 expects to earn one year after graduation and what they’ll actually take home.

A new survey from Clever and Real Estate Witch, covered by CNBC, puts the numbers plainly. Students expect to earn $80,000 on average. The real average starting salary clocks in at $56,153. Engineers are even further off, expecting $92,452 when the actual figure is closer to 20% below that.

The gap doesn’t shrink with time, either. Mid-career, graduates project they’ll be earning $144,889. The actual mid-career average? $95,521. The further out you go, the wider the illusion gets.

This isn’t a story about optimism. It’s a story about a structural market shift that hasn’t been priced into how young workers are thinking about their futures.

Where the $80,000 Expectation Came From

There’s a reason grads are anchored to higher numbers. The 2021 and 2022 job markets were genuinely unusual.

Remote work opened geography. Tech hiring went parabolic. Companies competed intensely for talent, and entry-level compensation rose sharply to match. That period created strong mental anchors for what a “good” starting salary looked like, and those anchors have stubbornly persisted in graduation speeches, LinkedIn feeds, and family dinner conversations.

The market has since corrected. Aggressively. But the cultural expectation hasn’t moved with it.

This is the core problem: grads are using outdated pricing signals in a market that has fundamentally changed.

When we wrote about the great AI salary bubble, we noted that AI-adjacent roles were commanding extraordinary premiums that distorted how people thought about tech compensation across the board. That distortion has filtered into how every grad, not just future engineers, is pricing their own value.

The AI Story Nobody Is Attaching to This Data

Here’s what the salary survey headlines are missing.

The $24,000 gap isn’t just a perception problem. It’s being actively widened by AI’s specific impact on entry-level hiring. Most coverage treats the expectation gap as a behavioral story — graduates being unrealistic. But the data tells a different story: the floor of the market is being quietly restructured.

According to Ravio’s 2026 Compensation Trends report, entry-level hiring dropped 73.4% in tech roles year-over-year. Administrative role hiring fell 35.5%. AI/ML hiring, meanwhile, grew 88%. The jobs being created are concentrated at the top; the jobs being eliminated are concentrated at the bottom.

When you reduce supply of entry-level roles while demand from new graduates holds steady, compensation compresses. Basic economics.

A Stanford Digital Economy Lab study found a 16% decline in early-career employment across AI-exposed occupations since late 2022. For developers aged 22 to 25, that figure is closer to 20%. Software development job postings on Indeed have fallen 53% from their late-2022 peak.

Graduates are arriving at a market that has fewer rungs on the bottom of the ladder than it did when their salary expectations were formed.

The Bifurcation Nobody Is Preparing Grads For

The job market in 2026 isn’t uniformly bad. It’s split.

At the top, workers with deep AI expertise are seeing compensation accelerate. Workers who can demonstrate AI skills earn a 56% wage premium over same-role colleagues without those skills, according to PwC’s Global AI Jobs Barometer. That premium has more than doubled from 25% just one year earlier.

At the bottom, routine and administrative entry-level roles are shrinking in both volume and compensation. The tasks that juniors used to learn on — the grunt work that formed the first rung of most career ladders — are being handled by AI agents that don’t need onboarding, benefits, or a salary review.

This is the bifurcation story. And it matters because graduates are being asked to price themselves in a market that no longer has a uniform entry point.

Here’s what the data actually shows about where the pressure is falling:

• Tech entry-level hiring fell 25% at the 15 biggest tech firms between 2023 and 2024, according to SignalFire
• 49% of US Gen Z job hunters believe AI has reduced the value of their college degree, per the World Economic Forum
• Workers aged 25 to 29 fell by 98,000 in Q1 2025, the steepest drop in 12 years
• 40% of employers expect to reduce their workforce in areas where AI can automate tasks, per the WEF Future of Jobs Report

Interview Guys Take: The salary expectation gap is real, but framing it as graduates being delusional misses the point. The job market they’re entering genuinely pays less for entry-level work than it did when the cultural benchmarks for “good starting salary” were established. Grads aren’t wrong to want more. They’re wrong to assume the market will meet them there without a differentiated skill set to back it up.

The Mid-Career Number Is the Most Alarming Part

Most headlines focus on the $24,000 starting salary gap. We think the mid-career number is the one worth paying attention to.

Students expect to earn $144,889 at the midpoint of their careers. The real average is $95,521. That’s a gap of over $49,000, more than double the starting gap in absolute terms.

This suggests the expectation problem isn’t just about miscalibrated optimism at graduation. It’s a sustained, compounding disconnect between how people project career earnings and what careers actually deliver.

Part of this is survivorship bias in how salary data circulates. The people posting big salary wins on LinkedIn are not a representative sample. The $200K engineer announcements go viral. The $52K marketing coordinator role does not.

Our own salary inflation reality check digs into how compensation data gets distorted in the media and social feeds that graduates consume, creating a persistent upward skew in how people benchmark themselves.

This Is Only the Latest Pessimism Signal

The salary gap data lands in an already sobering landscape for new graduates.

We reported that for the first time on record, college graduates are more pessimistic about their prospects than they were a year ago. Employer ratings of the job market for new graduates are now at their most pessimistic since 2020, according to NACE’s Job Outlook 2026 survey.

And it’s not just feeling. Only 30% of 2025 graduates found jobs in their field within a year of graduating. That’s a striking number that isn’t widely discussed in coverage of the Class of 2026 narrative.

The salary gap compounds this reality. When graduates do find roles, those roles are paying less than expected. When they don’t find roles immediately, the gap between expectation and reality grows in a different, more personal way.

Interview Guys Take: There’s a feedback loop forming here that the labor market hasn’t fully reckoned with. Graduates take longer to find jobs, accept lower salaries when they do, and enter careers already behind their mental projection of where they’d be. The compounding effect on financial confidence and career trajectory over the following decade is not trivial.

The Engineering Assumption Deserves Special Attention

Engineers expected to earn $92,452 on average, nearly 20% above their likely actual starting pay.

This is worth unpacking because engineering has historically been one of the most reliable fields for strong starting compensation. The assumption that an engineering degree equals a $90K+ floor has been culturally embedded for years.

That assumption is now outdated. Entry-level software and engineering roles are among the most directly impacted by AI automation. Code generation tools have reduced the apprenticeship value of junior developers significantly, making companies less motivated to hire and train junior talent at premium rates.

The irony is that engineers of all graduates should perhaps be most aware of AI’s impact on the market they’re entering. The tools they’ve been trained to use are directly displacing the roles they’re applying to.

When 29% fewer starting positions are now forcing Gen Z into career workarounds, engineers aren’t exempt. They’re often on the front line.

What the Data Is Actually Telling Us

The $24,000 salary gap is real. But the story underneath it is more structural than it appears.

Here’s the picture the data paints when you look at it together:

• Salary expectations were formed in an anomalous market (2021-2022) and haven’t adjusted
• AI has specifically compressed the entry-level tier where most graduates would land
• The roles being created pay more and require more; the roles being eliminated paid less and required less
• Graduates who can demonstrate AI skills are partially insulated from this compression
• Graduates who cannot are entering a market that has quietly repriced their category downward

The gap between expectation and reality isn’t a communication failure. It’s a structural signal that the entry-level job market has been reorganized from the bottom up, and the cultural narratives around starting salaries haven’t updated to reflect it.

Interview Guys Take: The most useful thing this data reveals isn’t that graduates should expect less. It’s that the market is bifurcating in ways that make skill differentiation more important than a diploma or a GPA. The students who will close the gap are the ones who treat their first job as a skills proof-of-concept, not a title. The ones who will struggle are the ones who priced themselves based on a market that no longer exists in the same form.

Checking out the highest paying entry-level jobs in 2026 shows where the market is still paying well at the junior tier. It’s worth knowing which fields are defying the trend before assuming the $24,000 gap applies uniformly across every career path.

The salary expectation gap is a number. What it’s measuring is a transition. And right now, most graduates are being handed a map drawn before the terrain changed.