State of Tech Hiring 2026

Key facts

• As of 31 July 2026, we track 44,381 open direct-apply tech roles across 6,666 companies — postings pulled straight from company Greenhouse, Lever and Ashby boards.
• Full-stack is the biggest niche by far at 11,790 open roles — more than AI/ML (9,819) and DevOps (5,861) combined would need to shift to catch it.
• The market is barely moving on net: +545 roles in the 28 days to 31 July 2026 (15,948 opened, 15,403 closed).
• Only 17% of postings state a salary, and just 29% are remote — so treat every pay figure here as the disclosed slice, not the whole.
• AI/ML pays the most of any large niche: median band $158,797–$214,781 (n=295), and it's also the most remote-friendly of the big three at 34%.
• Neura Robotics added the most roles of any company — a net +364 in 28 days, its entire 364-role board effectively new.

This is the direct-apply slice of the market — jobs posted on companies' own boards, not aggregator reposts or staffing-firm listings. That's a deliberate choice: it strips out the noise and duplicates, so when I say 44,381 roles I mean 44,381 distinct openings you can apply to at the source. What follows is built entirely from that first-party dataset as of 31 July 2026. Where I don't have a number, I say so.

Which tech roles are most in demand in 2026?

Full-stack, and it isn't close. At 11,790 open roles, full-stack engineering accounts for more than a quarter of everything we track and outnumbers the second-largest niche, AI/ML at 9,819, by nearly 2,000 roles. If you build for both sides of the wire, the raw volume is on your side.

The middle of the pack is where the interesting story sits. DevOps and infrastructure comes in third at 5,861 — a reminder that all the model-training excitement still needs someone to keep the pipes running. Data (4,389) and product (4,039) are effectively tied for fourth and fifth. Then there's a real cliff: design (2,457), backend (2,247), security (1,827), and at the tail, frontend (1,007) and mobile (945).

A caveat worth being honest about: backend and frontend look small partly because so much of that work now lives under the "full-stack" label. The 2,247 backend and 1,007 frontend counts are roles companies explicitly titled that way. The real demand for those skills is buried inside that 11,790-strong full-stack pile. So read the niche sizes as a map of how employers label work, not a strict measure of which languages get used.

The headline for job-seekers: generalist web engineering is where the doors are, and AI/ML is the fast-growing challenger. Everything else is a specialist game with fewer seats.

What does tech pay in 2026?

Among roles that disclose salary, AI/ML sits at the top with a median band of $158,797–$214,781 — the highest of any large niche. That's the closest thing this dataset has to a clean signal that the market is paying up for machine-learning skill. Behind it, the disclosed medians cluster tighter than you might expect:

• AI/ML: $158,797–$214,781 (n=295)
• Backend: $152,500–$200,000 (n=52)
• Full-Stack: $150,228–$200,000 (n=365)
• Mobile: $150,000–$200,000 (n=21)
• Product: $150,000–$182,000 (n=112)
• Security: $150,000–$180,000 (n=55)
• DevOps & Infra: $139,944–$184,986 (n=192)
• Data: $135,000–$174,667 (n=126)
• Frontend: $130,000–$170,000 (n=15)
• Design: $120,000–$175,000 (n=91)

Two things jump out. First, the top-of-band spread is where AI/ML separates itself: its upper figure ($214,781) is roughly $15K clear of the next niches, which top out around $200K. Second, design and frontend anchor the bottom, both starting in the low $120K–$130K range — the only two large niches whose lower band dips under $135K.

Now the honest part. These medians come from the 17% of postings that actually state a salary. The sample sizes tell you how much weight to put on each: full-stack (n=365) and AI/ML (n=295) are solid; frontend (n=15) and mobile (n=21) are thin enough that I'd treat their bands as directional rather than gospel. Pay disclosure remains the market's weakest habit. If four in five roles still hide the number, any single median is a candle in a dark room — useful, but don't bet your negotiation on it alone. See how we build these figures on the methodology page.

Which tech jobs are the most remote in 2026?

Backend leads on remote share at 35%, with AI/ML close behind at 34% and product at 33%. That's a neat inversion of the usual assumption that infrastructure-adjacent roles anchor you to an office — the model and API work is going remote faster than most.

The overall remote figure across the whole dataset is 29%, so roughly seven in ten tracked roles still want you in a location. The niches dragging that average down are the ones you'd guess if you thought about where the hardware lives: DevOps at 22% and full-stack at 24% — the lowest two. Full-stack being both the biggest niche and one of the least remote is a big reason the market-wide number sits under a third. Design (28%), mobile (27%) and security (29%) land in the middle.

If remote is your non-negotiable, the volume-versus-flexibility trade is stark: full-stack has the most jobs but among the fewest remote ones, while backend and AI/ML give you a better remote hit-rate at the cost of a smaller (backend) or more competitive (AI/ML) pool.

Is tech hiring growing or cooling in 2026?

Flat, with a lot of churn underneath. Net movement over the 28 days to 31 July was +545 roles — a rounding error against a 44,381-role base, under 1.3% growth in a month. The market isn't shrinking, but it isn't booming either.

The churn is the real story. Companies opened 15,948 roles and closed 15,403 in the same window. More than a third of the entire tracked market turned over in four weeks. That tells job-seekers two useful things. One: freshness matters enormously — a board that looked one way a month ago looks materially different now, so apply to what's live today. Two: a near-net-zero market with heavy churn rewards speed. The roles are there; they just don't sit around.

The flat top-line also masks concentration. A handful of aggressive hirers are responsible for an outsized share of the new-role flow, which means "the market is flat" and "specific companies are scaling hard" are both true at once. You can watch that turnover on the movers page.

Which companies are hiring fastest in 2026?

Robotics and physical-AI companies dominate the growth leaderboard. The single fastest mover is Neura Robotics, which added a net +364 roles in 28 days — and with exactly 364 active roles, essentially its whole board is freshly opened. That's the profile of a company standing up hiring from close to zero, fast.

SpaceX is the volume story: +183 net new roles on top of an already large 1,726 active board — by far the biggest single employer in this growth cohort. Then a cluster of AI-infrastructure and frontier-model names: Fluidstack (+168 net, 314 active), Mistral AI (+145 net, 145 active — again, effectively an all-new board), and Cerebras (+120 net, 120 active). When the chip-and-compute layer and a leading model lab are all opening their entire boards in the same month, that's where the AI/ML volume is coming from.

Cssmerge posts a striking +169 net against 178 active. Nord Security (+103 net, 105 active) and Holder Construction (+107 net, 112 active) round out the triple-digit gainers, the latter a reminder that "tech hiring" now runs through companies that aren't nominally tech at all.

Below the century mark, the pattern holds: hard-tech and energy keep showing up. Doctolib (+100 net, 138 active) brings health-tech into the mix; 1X (+80 net) is a second humanoid-robotics name alongside Neura; TerraPower (+80 net) and Form Energy (+73 net, 172 active) put nuclear and grid-storage on the board; Allen Control Systems (+76 net) and Bakerhicks (+76 net) fill out the engineering-heavy tail; and Superhuman Platform (+73 net, 74 active) closes the list with — you guessed it — another near-fully-new board.

The through-line: the fastest hiring in mid-2026 is being driven by robotics, AI compute and frontier models, energy, and a few non-obvious names outside pure software. Many of these companies show net-new counts almost equal to their total active roles, which means they're new to our dataset or ramping from a standing start. Either way, they're the boards to watch if you want to join something in acceleration mode rather than steady state.

How do the niches stack up against each other?

Put the four dimensions side by side and each niche has a personality:

• Full-stack — the volume play. Most roles (11,790) by a wide margin, solid pay ($150K–$200K), but among the least remote (24%). If you want the most doors, this is it.
• AI/ML — the money-and-momentum play. Second-largest (9,819), highest pay ($158,797–$214,781), strong remote share (34%), and the niche most fueled by the fastest-growing companies. The catch is competition and specialization.
• DevOps & Infra — big (5,861) and well-paid, but the least remote of all (22%). Reliable demand, office-leaning.
• Backend — small as a labeled niche (2,247) but the most remote (35%) and second-highest paid. A quietly strong position if you can find the roles.
• Product & Data — mid-sized, mid-paid, mid-remote. The dependable center of the market.
• Design & Frontend — the tightest job pools and the lowest pay floors. Specialist grinds where the seats are scarce.

The cleanest summary I can give: AI/ML wins on pay and growth, full-stack wins on sheer opportunity, backend wins on remote flexibility, and design/frontend are the hardest rooms to get into.

What should a tech job-seeker do with this right now?

Play the concentration, not the average. The market as a whole is flat (+545 net), so "is hiring up or down" is the wrong question. The right one is who is hiring — and the answer is a specific, nameable set of companies opening large numbers of roles this month: Neura Robotics, SpaceX, Fluidstack, Mistral AI, Cerebras and the rest of the movers list. Aim there.

Match your target to what you actually want. Chasing the highest ceiling and remote flexibility? Point yourself at AI/ML and backend. Chasing the most interviews and fastest response? Full-stack has the volume, even if you'll likely be on-site. Want to join something on the way up rather than something stable? The robotics-and-compute cohort is where the net-new roles are concentrated, and several of those boards are almost entirely fresh openings.

Move fast and re-check often. With more than a third of the market turning over every four weeks, the fre