Cyber Job Ads Hit a Four-Year Low While AI Security Engineer Postings Jump 268%: What the CyberSN Report Means for Hiring (AI Trends, 20 October 2026)

admincybersecurityTagged , , , , , , , , , ,

What’s trending in AI on 20 October 2026: The cybersecurity job market just sent two signals at once. A new CyberSN report counts 235,786 US cybersecurity job postings from January to September 2026, down 26.4% on the same months of 2025 and the lowest total for that stretch in four years. In the same data, postings for AI security engineers rose 268.3% and postings for cyber data scientists rose 203.5%. Classic roles such as security engineer, security analyst and penetration tester all shrank. The report does not prove AI caused the slowdown, and job ads are not hires. But read alongside how fast vendors are selling AI security tools, it shows where employers are putting their money. Below: what the numbers say, which roles grew and which fell, why the “AI replaced the analysts” story is too simple, and an if-then playbook for security leaders, job seekers and small businesses.

Key takeaways

  • Postings hit a four-year low. CyberSN counted 235,786 US cybersecurity postings from January to September 2026, down 26.4% from 320,210 a year earlier and below 2023 and 2024 too.
  • AI security is the exception. AI security engineer postings rose 268.3% to 2,409, and cyber data scientist postings rose 203.5% to 1,493.
  • The core shrank. Security engineer (down 32.3%), security analyst (down 30.5%), DevSecOps (down 27.4%) and penetration tester (down 45.9%) all fell sharply.
  • Governance is gaining ground. GRC’s share of postings rose from 9.4% in 2023 to 14.8%, and cybersecurity and privacy attorney postings grew 9.1%.
  • Ads are not headcount. The data counts deduplicated job ads, not hires, and does not show that AI caused the decline. Staff for the work AI tools create before you cut.
Cyber job ads hit a four-year low as AI security roles surgeTitle card. Headline: Fewer cyber jobs. More AI security jobs. Subhead: CyberSN counted 235,786 US cybersecurity postings from January to September 2026. Two large figures: minus 26.4 percent total postings versus 2025, in teal on a dark card; plus 268.3 percent AI security engineer postings, in orange. A small arrow graphic shows one line falling and one line rising. AI TRENDS · 20 OCTOBER 2026 Fewer cyber jobs. More AI security jobs. CyberSN counted 235,786 US cybersecurity job postings from January to September 2026, a four-year low. −26.4% all cyber postings vs Jan–Sep 2025 +268% AI security engineer postings, same period delana.co · Source: CyberSN Cybersecurity Jobs Report, 2026 through September (data pulled 7 Oct 2026)
The overall market cooled while one small, AI-focused corner of it boomed.

1. What the CyberSN report found

CyberSN, a cybersecurity recruiting and workforce firm, publishes a running count of US cybersecurity job postings. Its latest edition covers complete months through September 2026, using deduplicated postings from the data provider JobSpikr, sorted into 45 security roles. The data was pulled on 7 October, and eSecurity Planet covered it on 19 October.

The headline figure is blunt. From January to September 2026 there were 235,786 postings, against 320,210 in the same months of 2025. That is a 26.4% drop. It would be easy to dismiss as a hangover from a hot 2025, but the 2026 total is also below 2024 (244,911) and 2023 (273,498). Eight of the nine months this year ran below the same month last year. September was the exception, with 27,951 postings against 27,582.

CyberSN projects a full-year 2026 range of roughly 289,000 to 326,000 postings, based on normal seasonal patterns. It stresses that this is a pace estimate, not a forecast. For context, full-year 2025 closed at 392,356.

US cybersecurity job postings, January to September, 2023 to 2026Vertical bar chart. January to September postings: 2023, 273,498; 2024, 244,911; 2025, 320,210; 2026, 235,786, highlighted in orange as the lowest of the four years. Side note: 2026 is down 26.4 percent on 2025 and below 2024 as well. January–September postings, four years Deduplicated US cybersecurity job ads, not hires 273,4982023 244,9112024 320,2102025 235,7862026 Not just a tough comparison −26.4% vs 2025 Also below 2024 and 2023. 8 of 9 months trailed 2025. delana.co · Source: CyberSN Cybersecurity Jobs Report, 2026 through September; eSecurity Planet (19 Oct 2026)
2026 is the weakest January–September for cybersecurity job ads in CyberSN’s four-year series.

2. The roles that grew, and the ones that shrank

The total hides a sharp split. CyberSN tracks dozens of individual roles, and only a handful grew. The biggest jump by far was the AI security engineer, a role that barely registered a year ago: 2,409 postings, up 268.3%. Cyber data scientist postings rose 203.5% to 1,493. Cryptography and PKI specialists grew 22.8%, cybersecurity and privacy attorneys 9.1%, and data loss prevention engineers 7.2%.

Almost everything else fell. The security engineer is still the most advertised role at 44,083 postings, but that is down 32.3%. Security analysts fell 30.5%, cloud security engineers 34.5% and incident responders 33.1%. Some of the steepest drops hit hands-on offensive and investigative work: penetration testers fell 45.9%, threat hunters 40.0%, cyber risk analysts 51.5% and insider-threat analysts 55.6%.

RolePostings, Jan–Sep 2026Change vs 2025Q3 vs Q2 2026
AI security engineer2,409+268.3%n/a
Cyber data scientist1,493+203.5%n/a
Cryptography / PKI professional2,497+22.8%+72.3% (one-month spike)
Cybersecurity / privacy attorney24,193+9.1%+3.1%
Security engineer44,083−32.3%+7.9%
Security analyst24,095−30.5%−11.3%
DevSecOps24,317−27.4%−2.8%
CISO2,832−38.9%+19.5%
Penetration tester3,063−45.9%−26.6%
Threat hunter1,013−40.0%+1.5%
Cyber insider threat analyst1,088−55.6%−6.2%
Source: CyberSN Cybersecurity Jobs Report, 2026 through September. Selected roles; “n/a” where CyberSN gives no quarterly figure.

The last column adds nuance. Comparing the third quarter with the second, CISO postings rose 19.5%, security architects 14.0% and cyber risk analysts 20.2%. Security engineer postings also ticked up 7.9%. That is not a recovery, but it suggests the fall is slowing in some senior and design roles while hands-on testing and analysis keep sliding.

Winners and losers: change in postings by role, 2026 vs 2025Diverging horizontal bar chart of change in January to September postings versus 2025. Growing roles in orange: AI security engineer plus 268.3 percent; cyber data scientist plus 203.5 percent; cryptography and PKI plus 22.8 percent; cybersecurity and privacy attorney plus 9.1 percent. Shrinking roles in teal: security engineer minus 32.3 percent; security analyst minus 30.5 percent; threat hunter minus 40.0 percent; penetration tester minus 45.9 percent; cyber risk analyst minus 51.5 percent; insider threat analyst minus 55.6 percent. Who’s hiring, who isn’t Change in US postings by role, January–September 2026 vs 2025 AI security engineer+268.3% Cyber data scientist+203.5% Cryptography / PKI+22.8% Cyber / privacy attorney+9.1% Security analyst−30.5% Security engineer−32.3% Threat hunter−40.0% Penetration tester−45.9% Cyber risk analyst−51.5% Insider threat analyst−55.6% delana.co · Source: CyberSN Cybersecurity Jobs Report, 2026 through September. Bars to common scale.
Growth is concentrated in a few small AI and data roles; the big traditional roles are all down by double digits.

3. Governance is taking a bigger slice

CyberSN also groups roles into ten functional areas, and the mix has shifted since 2023. Defense, which covers security engineers, analysts and SOC work, is still the largest area, but its share of postings fell from 47.1% to 41.5%. Governance, risk and compliance (GRC) went the other way, from 9.4% to 14.8%. Research, which includes the cyber data scientist and cryptography roles, rose from 1.3% to 2.3%.

That pattern fits what companies are dealing with. AI brings new rules, contracts, audits and liability questions, from the EU AI Act to state privacy laws, and those land on GRC teams and lawyers. We have tracked that pressure here, from the EU’s first AI Act enforcement moves to AI safety claims becoming evidence in court. GRC is also the most open to newcomers in CyberSN’s data: 24.1% of GRC postings are entry level, against 9.5% across all roles and just 7.3% in Defense.

Share of cybersecurity postings by function, 2023 vs 2026Dumbbell chart comparing January to September share of postings in 2023 (grey dot) and 2026 (colored dot). Defense fell from 47.1 percent to 41.5 percent. GRC rose from 9.4 percent to 14.8 percent. Product security fell from 14.3 percent to 12.8 percent. Research rose from 1.3 percent to 2.3 percent. Sales rose from 1.0 percent to 2.3 percent. Note: GRC has the highest entry-level share at 24.1 percent. The mix is shifting toward governance Share of US cybersecurity postings by function, Jan–Sep 2023 vs Jan–Sep 2026 0%20%40% Defense47.1% → 41.5% GRC9.4% → 14.8% Product security14.3% → 12.8% Research1.3% → 2.3% Sales1.0% → 2.3% Grey dot = 2023 · colored dot = 2026 · GRC has the highest entry-level share of any function: 24.1% delana.co · Source: CyberSN Cybersecurity Jobs Report, 2026 through September
Defense still dominates, but governance and research are taking a bigger share of the ads.

4. Did AI take the jobs? Not so fast

It is tempting to draw a straight line from AI tools to fewer security jobs. The timing fits. This year Anthropic and OpenAI have both pitched AI models for defensive security work, eSecurity Planet notes, and AI-found vulnerabilities have flooded maintainers, as we covered when AI broke the bug bounty. Penetration testing and threat hunting, two of the steepest declines, are exactly the tasks vendors say AI agents can now help with.

But the evidence does not support that conclusion yet, and the report itself does not claim it. Three cautions matter:

  • Ads are not jobs. As eSecurity Planet puts it, “The figures measure deduplicated advertisements, not actual hires.” Fewer ads can mean fewer reposts, more internal moves or slower budgets as easily as fewer roles.
  • Causation is not shown. eSecurity Planet stresses the report does not establish whether AI adoption caused the decline. Budget caution, a strong 2025 base and hiring through contractors all play a part. Contract work rose to 9.2% of postings, up 1.4 points.
  • Fewer ads do not mean full teams. ISC2’s 2024 study estimated a global cybersecurity workforce gap of about 4.8 million, a measure of need rather than open vacancies. Funded jobs and real staffing needs are different numbers.

Even the share of postings that mention AI is mixed. CyberSN found AI or machine learning terms in 14.6% of postings from January to September, slightly below 15.1% a year earlier, though September hit 20.0%, the highest since January 2025. A separate Cisco-led study from the AI Workforce Consortium, reported by Help Net Security in August, found AI skills in 28.5% of cybersecurity postings across G7 countries, roughly double a year earlier. The two studies use different methods and markets, so the exact number varies. The direction does not: AI fluency is becoming a hiring filter.

The broader pattern resembles what we saw in Anthropic’s own data, where junior hiring fell to 9% of technical hires as agents took on more work. Entry-level postings make up just 9.5% of CyberSN’s total. That is a problem for anyone trying to break in, and a long-term problem for employers who will need tomorrow’s senior staff.

5. AI creates security work too

The 268% jump in AI security engineer postings is the other half of the story. Every AI tool a company deploys adds something to protect: models, prompts, agent credentials, data pipelines and the tools agents can call. Someone has to inventory them, set permissions, test them for prompt injection and watch what they do.

Vendors are building products for that job. On 19 October, GeekWire reported that Microsoft is adding risk scores for individual AI agents to Microsoft 365 E7 and its Project Perception security system, now in private preview. Tools like that need people who can read the scores, tune the rules and decide when to pull the plug. And when AI finds problems faster than people can fix them, the bill lands somewhere: we saw that when Bank of America hired 1,000 contractors to fix what AI found.

eSecurity Planet’s guidance for security leaders makes the same point. Staff for the work AI tools create: validating findings, maintaining integrations and handling incidents. Measure missed threats and wrong findings, not just hours saved. And separate tools that only recommend from tools that act, such as disabling accounts, changing permissions or pushing patches, because the second kind needs tight permissions, human approval and logs.

6. Your if-then playbook

Recent guides here have been traffic lights, scorecards and step-by-step plans. This one is a set of if-then rules: find the line that describes your situation and do what it says.

If-then playbook for the AI-era security job marketThree columns of if-then cards. Security leaders: if you are cutting headcount because of AI, then first list the work the AI creates; if an AI tool can act, then require human approval and logs. Job seekers: if you are a SOC analyst or pen tester, then add AI red-teaming and agent security; if you are new to the field, then look at GRC, where 24.1 percent of postings are entry level. Small businesses: if you cannot hire an AI security engineer, then assign one named owner for AI tools; if you buy AI security tools, then ask who reviews their findings. If this is you… then do this Six rules for the AI-era security job market Security leaders IF cutting roles for AITHEN list the work theAI creates first IF a tool can actTHEN require humanapproval and logs Job seekers IF analyst or pen testerTHEN add AI red-teamingand agent security IF new to the fieldTHEN look at GRC:24.1% entry level Small businesses IF you can’t hire oneTHEN name one ownerfor every AI tool IF buying AI securityTHEN ask who reviewsits findings delana.co · Based on CyberSN (Oct 2026) data and eSecurity Planet guidance for security leaders
Pick your column. The full list follows below.

If you lead a security team

  1. If you are planning to cut roles because of AI, then first write down the work the AI will create: reviewing its findings, fixing what it finds, maintaining its integrations and investigating its mistakes. Cut only after a pilot shows the team can catch and recover from those mistakes.
  2. If an AI tool can take action, such as disabling accounts or deploying patches, then limit its permissions, require human sign-off for high-impact steps and keep logs good enough to reconstruct what it changed. Our 14-point AI insider-risk scorecard is a good starting checklist.
  3. If you are measuring an AI tool’s value, then track missed threats and false findings alongside time saved. A faster wrong answer is not a saving.
  4. If you have stopped hiring juniors, then decide where your future senior staff will come from. Pair new hires with AI tools under supervision instead of leaving the pipeline empty.

If you work in security or want to

  1. If you are a SOC analyst, pen tester or threat hunter, then add AI to your skill set in a way you can show: test an AI agent for prompt injection, review an AI-generated vulnerability report, or write detections for agent misuse. These map directly to the AI security engineer role.
  2. If you are just starting out, then look hard at governance, risk and compliance. Nearly a quarter of GRC postings are entry level, and AI regulation is adding work there every month.
  3. If you write code, then learn the skills the Cisco-led study saw most often alongside AI in security ads: Python, prompt and context engineering, agent orchestration and MLOps.

If you run a small or midsize business

  1. If you can’t hire an AI security engineer, then give one named person ownership of every AI tool and agent your staff use, starting with an inventory. Shadow AI is the most common blind spot.
  2. If you are buying an AI-powered security service, then ask the vendor who reviews its findings, how quickly a human responds, and what happens when the AI is wrong.
  3. If you use a managed security provider, then ask how it is staffing for AI-related risks, and whether its own analysts are being replaced by tools you have not evaluated.

7. The bottom line

The cybersecurity job market has not collapsed, but it has cooled and changed shape. Fewer ads are going out for the big generalist roles, and the growth is in a narrow band of AI security, data and governance work. Nobody can yet say how much of that is AI replacing people and how much is budgets, timing and a strong 2025. What is clear is that employers now want security people who can secure AI and work alongside it. For security leaders, the smart move is to staff for the work AI creates before cutting the work it might replace. For everyone in the field, it is to make “I can secure an AI agent” a line on the CV. And for every business, the job of AI agent security needs a named owner, whether or not that title appears on a job ad.

Frequently asked questions

How much have cybersecurity job postings fallen in 2026?

According to CyberSN, US cybersecurity job postings from January to September 2026 totaled 235,786, down 26.4% from 320,210 in the same months of 2025. It is the lowest January to September total in the four years CyberSN reports, below 2023 and 2024 as well.

Which cybersecurity jobs are growing?

AI security engineer postings rose 268.3% to 2,409, and cyber data scientist postings rose 203.5% to 1,493. Cryptography and PKI professionals grew 22.8%, cybersecurity and privacy attorneys 9.1%, and data loss prevention engineers 7.2%. Most other tracked roles declined.

Is AI replacing cybersecurity jobs?

The data does not show that. CyberSN counts job ads, not hires, and the report does not establish that AI adoption caused the decline. Budget caution, a strong 2025 comparison and more contract hiring may all contribute. AI is clearly changing which skills employers ask for, though.

What does an AI security engineer do?

An AI security engineer protects an organization’s AI systems and the data and tools they touch. Typical work includes inventorying AI tools and agents, setting their permissions, testing them for prompt injection and data leakage, monitoring what agents do, and reviewing findings from AI-powered security tools.

What is the best entry point into cybersecurity in 2026?

Governance, risk and compliance stands out in CyberSN’s data: 24.1% of GRC postings are entry level, against 9.5% across all roles. GRC’s share of all postings has also grown, from 9.4% in 2023 to 14.8% in 2026, helped by new AI and privacy rules.

Should my company cut security staff because of AI tools?

Not before testing. AI tools create work of their own, such as validating findings, maintaining integrations and handling mistakes. Pilot the tool on a defined workflow, measure missed threats and false findings as well as time saved, and confirm your team can detect and recover from its errors before changing headcount.

Sources