The Great Disconnect

Total U.S. tech job postings have followed a choppier path over the past two years than headline numbers suggest. Two dips stand out: one in December 2024, another in November and December of 2025. But the recoveries tell different stories. After the more recent dip, growth kept accelerating instead of leveling off. Postings have now risen for seven straight months off the December 2025 low, reaching the highest point in this two-year window in June 2026.

This is a steadier, more sustained climb than the market has seen recently. Critically, it’s concentrated in specific industries and specific skills, not spread evenly across the board, as the next two sections show.

Just like industries and job titles, tech hiring growth doesn’t stop at geography. The states posting the strongest six-month gains aren’t the ones most people would name first, and all ten are outpacing California and Texas, the two largest tech markets by volume. Growth right now is spread across the map, not concentrated in the coasts or the hubs everyone already watches. Sourcing strategies anchored only to the traditional hubs are missing where the momentum actually is.

Every angle in this chapter, industries, job titles, candidate priorities, and geography, points to the same pattern: demand is real, but it’s concentrated, not universal. That distinction is the difference between a message that resonates with candidates and clients and one that falls flat.

• Growth is real, but narrow. It’s concentrated in specific industries (Insurance, Consulting, Manufacturing), specific titles (AI-adjacent and infrastructure roles), and specific states (NJ, MD, MI, MA, IL), rather than signaling a broad-based rebound.
• This is a steadier, more sustained climb. Postings have risen for seven straight months from the December 2025 low, reaching the highest point in this two-year window in June. However, growth remains concentrated in the industries and skills detailed above rather than reflecting a broad-based surge.
• “The market’s growing” isn’t a message on its own. The useful version of this story is specific: which industry, which title, which state. Generic optimism won’t match what candidates and clients are actually experiencing.
• Where you’re spending outreach matters as much as how much you’re spending. If business development is still concentrated on software and tech company logos, that’s the slowest-growing ground in this data. Insurance, consulting, and the emerging state markets above are where the fastest-moving opportunity sits, and moving early tends to win better terms than waiting for the rest of the market to catch up.

When we asked tech professionals how they feel about today’s job market, the picture was stark:

In short:roughly 1 in 7 candidates feel good about today’s market.

Much of this frustration traces back to how AI is used in candidate screening.

Tech professionals overwhelmingly agree AI tools filter out good candidates who don’t optimize for the “right” keywords. The pressure to exaggerate qualifications is highest among candidates under 40 (94%), and women alter their resumes for AI at a notably higher rate than men (75% vs. 60%), often stripping out personality and real accomplishments to “match the keyword model.”

These aren’t candidates being dramatic. They’re accurately describing the process they’re going through.

Ghost jobs, unrealistic qualifications, and AI screening that candidates don’t trust aren’t isolated complaints. They add up to one conclusion: this frustration isn’t manufactured, it’s earned. Roughly 1 in 7 candidates feel good about today’s market, and the data in this chapter explains why. Here’s what that means going forward.

• The frustration is rational, not perception. Ghost jobs, AI screening gaps, and unrealistic postings are measurable. Candidates are responding accurately to a broken process, not overreacting to a good one.
• Trust is the currency being spent. 92% believe AI misses qualified people, and candidates are already adapting, or gaming, their resumes in response. That erosion compounds if it goes unaddressed.
• This is where organizations differentiate. Candidates aren’t just tired of the market. They’re tired of being screened by systems that don’t see them. A human-backed process is a real advantage here, and it’s also a sourcing opportunity: every candidate who stripped their resume down to keywords is a candidate actively looking for someone to treat them like a person.

We asked candidates directly: compared to a year ago, how often are AI-related skills showing up as a requirement for the roles they pursue?

The posting data backs this up. U.S. tech job postings requiring AI skills have climbed from roughly 19,000 in January 2024 to nearly 67,000 in May 2026, more than tripling in just over two years. Growth has accelerated particularly sharply in 2026, reinforcing what candidates are experiencing firsthand: AI skills are increasingly showing up in the roles they pursue, reshaping what employers look for in qualified tech talent.

AI requirements are showing up more often, and the language describing them is changing just as fast. Some skills are surging in job postings while others fade, even when they describe roughly the same underlying capability. Here’s where some of that shift shows up.

The underlying capability hasn’t disappeared. The language describing it has changed. Candidates and hiring processes still using last year’s terms are going to miss each other.

This shift isn’t contained to obviously-AI job titles. Some of the fastest-growing titles in tech right now don’t sound like AI jobs at all. These are growing alongside Agentic AI and AI Agents (+805% and +591%), and connecting these systems is real, measurable demand of its own:

Building the model is one part of the equation; connecting it into everything else a business runs on is real, measurable demand of its own.

This also tracks with a broader shift happening inside organizations. As AI-assisted coding tools become common, more of the actual code is being written by people outside traditional engineering roles, while software engineers increasingly shift toward reviewing and integrating that work rather than writing all of it themselves. That shift is part of what’s fueling demand for roles like Forward Deployed Engineers, people who can take AI-generated work and responsibly connect it into real business systems.

Across postings, skills, and job titles, the same story keeps repeating: it’s not that AI expertise is disappearing, it’s that the vocabulary describing it is changing faster than most hiring processes can keep up. That shift reaches well beyond AI-titled roles, and it’s already deciding who gets seen as qualified and who gets filtered out. Here’s what that means going forward.

• It’s a vocabulary shift, not a shrinking pool. Machine Learning Algorithms is declining while Agentic AI explodes. The skills didn’t disappear, the language around them did.
• This isn’t contained to AI roles. Traditional titles are growing fast too, and they increasingly require integration and infrastructure skills.
• Static screening will keep missing good people. If “qualified” keeps moving and a hiring process doesn’t move with it, capable candidates get filtered out simply because they’re described differently than the keyword list expects. A sourcing strategy still anchored to last year’s terms isn’t looking at a smaller pool, it’s looking at a differently-named one, and that’s a fixable search problem, not a strategic one.
• Keeping vocabulary current pays, literally. Scarce, fast-moving skill categories like Agentic AI and Vector Database tend to command premium rates. Updating sourcing language isn’t just about filling more roles, it’s about filling higher-margin ones.

Insurance, Consulting, and Manufacturing are outgrowing Tech and Software. Insurance postings are up 69% and Consulting up 34% over the last six months. Clients outside the traditional tech industry are where hiring is heating up fastest.

Growth is geographic too, not just industry-specific. New Jersey (+37%), Maryland (+28%), Michigan and Massachusetts (+25%) posted some of the strongest six-month growth in the country. Widen the map before writing off a market.

Volume still lives in the fundamentals. Software Engineers, Data Engineers, and Systems Engineers remain the highest-volume titles in tech. The percentage-growth stories are real, but they’re happening on a much smaller base. Don’t chase headlines and overlook where the bulk of fillable roles actually sit.

A few early indicators are worth tracking as this market continues to evolve:

• Which industries start hiring their own internal tech recruiters first. Industries usually ramp up internal tech hiring before that demand overflows to staffing firms. Watching where internal recruiting teams are growing can offer an early read on where outside placement demand is headed next.
• Whether foundational skills keep pace with emerging AI skills. When infrastructure and integration skills, such as Enterprise Integration, grow alongside headline-grabbing skills like Agentic AI, it can signal that companies are moving beyond AI experimentation and into real-world deployment.
• Whether application volume is starting to outpace real signal. As more candidates lean on AI tools to mass-apply, a rising number of applications won’t necessarily mean a rising number of genuine candidates. Building the gap between application volume and real signal into how a pipeline gets evaluated is worth doing now, before it gets noisier.

The gap this report opened with, demand is real but candidate confidence hasn’t caught up, doesn’t close on its own. Postings are climbing, but only in specific industries, titles, and states, not across the board. Candidates are responding rationally to a process that’s gotten harder to read, not overreacting to a market that’s actually fine. And the bar for what counts as “qualified” keeps shifting as AI reshapes the vocabulary and the roles themselves. Closing that gap takes action: getting specific about where the growth actually is, helping candidates speak today’s vocabulary, and keeping people in the loop on screening decisions. Here’s what that looks like in practice.

• The market didn’t stop. It shifted. Postings are up, but growth is concentrated in specific industries, titles, and skills. Read the specifics, not just the headline number.
• AI fluency is table stakes, not a specialty. Help candidates layer AI vocabulary onto existing expertise, and don’t let good candidates get screened out over keyword mismatches.
• Trust is the differentiator. Human review, transparent timelines, and realistic job descriptions matter more than ever. This is where thoughtful, human-backed hiring out-performs AI-only approaches.
• The edge is specificity, not size. The organizations that win the next few quarters won’t be the ones with the most headcount. They’ll be the ones who can talk specifics, industry, title, state, skill, and move on them while the rest of the market is still reading the headline number.