Are Data Science Jobs Declining? What the Data Actually Shows in 2026

The honest answer is that data science jobs are not declining in aggregate, but the market has genuinely changed since the hiring boom of 2020-2021, and both of those things are true at once. Anyone weighing this career deserves the full picture rather than either the alarmist "dead career" framing or the uncritical "still the sexiest job" framing that both circulate online.

The Short Answer: Are Data Science Jobs Declining?

1. The official U.S. government projection is strongly positive: the Bureau of Labor Statistics projects 34% employment growth for data scientists from 2024 to 2034, far faster than the average for all occupations.
2. The market did correct after 2021-2022 overhiring, particularly at large tech companies, and that correction is the real basis for most "declining" claims, even though those claims are often now citing stale data as if it still described today.
3. Titles have shifted more than headcount has shrunk: some of the work once labeled "data scientist" now gets hired under data analyst, analytics engineer, or ML engineer titles, which distorts a narrow keyword-based reading of job boards.
4. Growth is uneven by role type, with data engineering and applied ML/AI-adjacent roles currently showing stronger demand signals than generalist data scientist postings specifically, detailed below.

What the Official Data Actually Shows

According to the U.S. Bureau of Labor Statistics Occupational Outlook Handbook, data scientist employment is projected to grow 34% from 2024 to 2034, adding roughly 82,500 positions to a base of about 245,900 in 2024, with around 23,400 annual openings projected on average (accounting for growth plus replacement of workers who leave the field or retire). That 34% figure is categorized by BLS as "much faster than average" against a total all-occupation growth rate that's a fraction of that size. The median annual wage was $112,590 as of May 2024.

That data point alone doesn't fully resolve the debate, since BLS projections are long-range, aggregate, and don't capture short-term hiring cycles or the specific experience of searching for a job in any given month. But it's the most authoritative available answer to the literal question "is this field growing or shrinking," and it points clearly toward growth, not decline, over the medium term.

It's also worth being specific about what that 23,400 annual openings figure actually represents, since it's frequently misread. BLS explains annual openings as a combination of newly created positions from growth and positions that open up as existing workers leave the occupation entirely (retirement, career change, and so on), averaged over the projection decade. That means the real number of postings in any single year can vary meaningfully around that average, and a slower year within a strong long-range trend is fully consistent with the projection, not a contradiction of it.

A Short History of How the Market Got Here

Understanding why the current market feels different from a few years ago helps explain why both the "declining" and "still growing" framings can each point to something real:

1. 2019-2021: rapid, broad hiring. Demand for data science talent grew quickly across both established tech companies and traditional industries newly investing in analytics, coinciding with historically low interest rates that supported aggressive corporate hiring and headcount growth generally.
2. 2022-2023: a documented correction. Multiple large, publicly known technology employers announced hiring freezes and layoffs during this period, and data science and adjacent analytics roles were affected alongside broader tech headcount reductions. This is the period most "data science is dying" content is actually describing, even when published or updated more recently.
3. 2024-2026: stabilization and role differentiation. Hiring didn't return to the 2020-2021 pace, but it also didn't continue contracting; instead, the market reorganized around more clearly separated titles (data engineer, ML engineer, analytics engineer, AI engineer) rather than a single broad "data scientist" umbrella covering all of that work, which is the structural shift described throughout this article.

That sequence, boom, correction, reorganization, is a fairly ordinary pattern for a fast-growing field that briefly overheated, and it looks quite different from a field in genuine terminal decline.

Why the Market Still Feels Tighter Than the Growth Numbers Suggest

A long-range growth projection and a genuinely harder day-to-day job search are not actually in conflict. Several specific, well-documented dynamics explain the gap:

1. The 2020-2021 hiring boom created an unusually high baseline. Many large tech companies over-hired data science and adjacent roles during that period, and the correction that followed (documented hiring slowdowns and freezes at multiple major tech employers starting around 2022) is the real basis for most current "market is shrinking" claims, even where those claims no longer specify how old the underlying data is.
2. Title relabeling absorbs some of the apparent decline. Some roles that would have been posted as "data scientist" in 2021 are now posted as "data analyst," "analytics engineer," or "machine learning engineer," which is partly a genuine skill-mix shift and partly a labeling change, and both make raw data-scientist-titled job counts an incomplete measure of the field's actual health.
3. The applicant pool has grown faster than early-career openings specifically. A large wave of bootcamp graduates, career changers, and new degree holders entered the field over the past several years, which increases competition for entry-level roles even where overall demand is healthy, a dynamic covered from the candidate side in the Entry-Level guide.
4. The skill bar per role has risen. Employers increasingly expect broader applied skill (production deployment awareness, applied statistics depth, tool-specific fluency) for the same title than a few years ago, which makes the search feel harder even when the number of open roles hasn't shrunk.

Where the Growth Is Actually Concentrated

Demand isn't distributed evenly across every role that touches data. Based on current hiring patterns and BLS occupational data across adjacent categories, growth is currently strongest in these areas:

This is also why role type matters more than the single word "data science" when judging your own odds: a search focused specifically on data engineering or ML engineering roles is currently working with a different, generally more favorable supply-demand picture than a search focused only on generalist data scientist titles.

What Current Job Postings Data Shows

Beyond the long-range government projection, shorter-term postings analysis adds useful color, with the caveat that it reflects a specific, smaller sample rather than a government-scale dataset. One 2026 analysis of roughly 1,000 data scientist job postings by 365 Data Science found several notable shifts in what employers are actually asking for: Python appeared in 57% of postings, down from 78% the prior year, while natural language processing skill requirements surged from 5% of postings to 19% over the same period, and machine learning was referenced in 69% of postings. Roughly 47% of postings sought a data science-specific degree, with 30% specifying a master's degree and 24% a PhD, while a meaningful 26% of postings didn't specify formal education requirements at all, broadly consistent with the "typical entry-level education" framing BLS uses rather than a strict requirement.

Read carefully, that data tells a consistent story with the BLS projection and the historical pattern above: the skill mix employers want is shifting quickly toward applied AI and NLP work, formal education requirements remain common but far from universal, and none of it points toward the role disappearing, even as what a "typical" data scientist is expected to know keeps moving.

Is AI Replacing Data Science Jobs?

AI tools are changing parts of the day-to-day work, particularly routine data cleaning, boilerplate code, and first-draft analysis, but this is meaningfully different from AI replacing the roles outright. The more accurate short version: AI is raising the bar for what counts as valuable human judgment in the role (framing the right question, validating a model's real-world tradeoffs, communicating uncertainty to a non-technical stakeholder) while automating some of the lower-judgment tasks around it. This deserves its own full treatment rather than a summary here; a dedicated breakdown of what AI is and isn't currently automating in data science work is covered in a companion article on this site.

Should You Still Pursue Data Science?

For most people asking this question, yes, with a realistic view of what "still worth it" means in 2026. The long-range government projection is genuinely strong, the field isn't disappearing, and demand is healthy across the broader family of data-related roles even where it's uneven within any single title. What's changed is that a portfolio, a clear specialization (data engineering, ML/MLOps, or a specific industry), and honest odds-awareness matter more now than they did during the 2020-2021 boom, when almost any data science credential was in high demand. The Career Entry guide covers exactly what that preparation looks like in practice.

What This Means If You're Currently Job Searching

The data above is reassuring at the aggregate level, but a job search happens one application at a time, and a few practical adjustments follow directly from the patterns described here. First, broaden your search beyond the exact phrase "data scientist" to include data engineer, analytics engineer, and ML engineer postings if your skill set genuinely overlaps, since meaningfully similar work is now distributed across more titles than it was a few years ago. Second, expect the entry-level segment specifically to be more competitive than the aggregate growth number implies, and weight your preparation accordingly, with a stronger portfolio and more targeted applications rather than a high-volume, low-specificity approach. Third, don't over-index on any single company's layoff headlines as a signal about the whole field; a specific employer's correction after a specific overhiring period is a company story, not automatically an industry story. Finally, treat AI-adjacent skills (applied NLP, working with AI-assisted tooling, understanding model deployment basics) as a genuine differentiator right now, since postings data shows demand shifting toward these skills faster than most existing prep content has caught up to.

Common Misreadings of the "Data Science Is Dead" Narrative

1. Citing 2021-2022 hiring-freeze data as if it describes today, without noting how many hiring cycles have passed since, which is the single most common flaw in "market is disappearing" content still circulating online.
2. Treating a specific company's layoffs as a field-wide trend, when large, well-publicized tech layoffs often reflect that specific company's over-hiring correction rather than industry-wide contraction.
3. Reading data-scientist-titled job counts alone as the full market, missing the meaningful volume of equivalent work now posted under adjacent titles.
4. Assuming AI automation and job elimination are the same thing, rather than the more accurate picture of AI shifting which tasks within the role carry the most value.
5. Ignoring the entry-level-versus-experienced distinction, since aggregate growth and a genuinely harder entry-level search can both be true simultaneously, a dynamic covered in more depth in the Entry-Level guide .

FAQ

Are data science jobs declining in 2026?

Not in aggregate. The U.S. Bureau of Labor Statistics projects 34% employment growth for data scientists from 2024 to 2034, categorized as "much faster than average." The market has genuinely tightened since the 2020-2021 hiring boom and shifted in how roles are titled, but that's a change in shape, not an overall decline.

Why do so many articles say data science is a dead career?

Much of that content originated during or shortly after the 2021-2022 hiring correction at large tech companies and continues to circulate, sometimes updated with a new year in the title but not with new underlying data. It reflects a real, specific hiring slowdown at a point in time, not a permanent or field-wide trend.

Is data science oversaturated with candidates?

At the entry level specifically, competition has increased meaningfully as more bootcamp graduates, career changers, and new degree holders have entered the field over the past several years. At the experienced level, and in adjacent roles like data engineering and ML engineering, demand signals remain considerably stronger relative to supply.

Which data-related roles are growing the fastest right now?

Data engineering, ML engineering/MLOps, and AI engineer roles currently show the strongest demand signals among data-related titles, largely because production infrastructure and deployment skills are more scarce than generalist analysis skills.

Is it still worth learning data science in 2026?

For most people, yes, provided the goal is realistic: entering with a clear specialization and a genuine portfolio, rather than expecting the 2020-2021-era ease of hiring to still apply. The underlying long-range demand, per BLS, remains strong.

Did data science job postings actually drop after 2021?

Yes, at multiple large, publicly known technology employers during 2022-2023 specifically, which is well documented and is the real basis for most current "declining" claims. That correction followed an unusually fast hiring period in 2020-2021 and has since stabilized rather than continued contracting, per current BLS projections.

Your Next Step

Browse current openings across the full range of data-related roles, filtered specifically by role type (data scientist, ML engineer, data engineer, analytics engineer, and more), on finddatasciencejobs.com, rather than relying on a single job title to judge the market.

Read next:

• How to Get a Job in Data Science: The Complete 2026 Guide
• How to Get an Entry-Level Data Science Job With No Experience