# Tech Career Report 2026 · Manfred

2026 edition · November 2026 · https://www.getmanfred.com/tech-career-report

Every year at Manfred we've shared an X-ray of Spain's tech talent: what it knows, what it earns, and where it works or wants to work from. Our view of the market has always been somewhat limited, so this year, for the first time, we've put what we know about talent in front of an equally big mirror, the market, crossing the Manfred community with the tech offers collected by TheirStack. That lets us bring you the most complete report we've ever made. Spoiler: the interesting part is everything that doesn't match.

## Key figures

- **127.665** professionals in the Manfred community
- **167.645** tech job offers in the last 12 months
- **10,3%** of development offers show a salary
- **100%** of Manfred offers are published with a salary

## Quotable facts

- 4.4% of tech job offers in Spain look for junior profiles, while one in five professionals has less than two years of experience.
- Only one in ten development job offers in Spain is published with a salary: 10,560 of 102,792 in the last 12 months.
- The median offer for backend in Spain is €37,500 gross a year; among talent, a backend with 10 years or more asks for around €57,700.
- 67% of tech offers allow working from home (33% remote, 34% hybrid); 52.5% of talent prefers fully remote.
- Kubernetes appears in 17.4% of tech offers and is used by 8.8% of talent: a 2.0× tension index.
- Tech offers with AI in the title went from 1.5% in 2023 to 10.5% in 2026, and offers mentioning AI agents grew 8.3× in a year.
- One in three tech offers in Spain is written in English; asking for English in a Spanish-language offer lifts the median from €33,000 to €36,000.
- 39% of tech offers are posted by consultancies and IT services firms, and another 15% by recruitment agencies.
- By company size, the offered median ranges from €31,500 to €39,000; by company type, senior base pay ranges from about €40,000 to about €90,000, and at big tech, with variable pay, the total package ranges from €110,000 to €250,000.
- Only Product Design / UX reaches parity (50% women); Software Architect is furthest (5%) and Tech Lead stays at 6%.
- 52% of tech offers don't mention any specific benefit; the most cited is health insurance (26.5%).
- Two in three AI offers in Spain ask for Python (66%), 36% talk about AI agents and nearly one in three ask for LLMs or RAG (31%).
- Security is the most pressured role after AI: it appears in 6.3% of tech offers in the last 12 months while only 1.2% of talent holds it (5.3×).
- Half of the Manfred community's tech talent earns €40,000 gross a year or more, 1.6 times the country's median salary (€24,497, INE 2024).
- Half of tech talent lives outside Madrid and Catalonia, yet only 21% of offers are there.
- Judgement is what's wanted, but it's measured in years: 55.5% of tech offers that state years of experience ask for 2 to 4, and only 3 in 100 accept under 2.
- Starting out, women and men ask for the same; with 10+ years, women's median is 13.4% lower (about €8,300 a year), and it is lower in 19 of 19 roles.
- 98% of surveyed talent use AI in their daily work, but only 20% have been trained by their company and 21% use what it generates almost untouched.

## How to read it

- Talent: the Manfred community (127,665 professionals; percentages on Spain-based profiles with a tech role; talent salaries are what talent asks for).
- Offers: the market, 167,645 tech offers from the last 12 months (TheirStack). Salaries are calculated on the 102,792 development offers.
- Amounts are gross annual salaries in euros. "†" marks small samples.
- Talent salary has two readings: what talent earns today (only used for the headline: half earns €40,000 or more) and what it asks for to change jobs (everything else). Experience brackets don't overlap: under 2 years, 2–4, 5–9 and 10+.
- Talent percentages are shares, not headcounts: don't turn them into numbers of people.

## Data is great. What you do with it is better

We've spent years on the other side of the mirror, talking every day with tech talent and the companies looking for it. If this report stirred something, let's talk.

## 01 · Supply & demand

*Who is out there, and who is wanted*

The starting picture: talent and offers by role, year-on-year change, experience and our take.Spoiler: it's not the same picture, and those just starting out get the worst of it. We said it months ago: it's “no country for juniors”.

### Talent and offers by role

*Source: talent (Manfred community) + offers (market)*

**One role to rule them all**

How much each role weighs among talent and among the last 12 months of offers. Offers with a generic title are split according to what their description asks for. If it doesn't fit anywhere, it stays in “Other”. (We're all entitled to be a little fussy.)

- Each role's rank by its share of tech offers in Spain (1 = the most offers), with the share next to it. 2023 to 2026 (January to September). Each offer counts in one role only; development offers with generic titles (from 41% in 2023 to 28% in 2026) are not shown. We compare shares, not volumes: how many offers are collected depends on the channels TheirStack tracks each year (LinkedIn joins in September 2023 and Tecnoempleo is complete from 2025), so volume doesn't say whether the market is growing.
- "Pool" is the sum of the talent weight of the source roles. One profile can feed several roles, so pools don't add up to 100. The "measurable today" figures are real TheirStack offers from the last 12 months; the rest is our editorial reading and is marked as signal or hypothesis.
- Offers: each role's share of the 167,645 tech offers in the last 12 months (1 Oct 2025 to 30 Sep 2026) by title. The 29.5% with generic titles ("developer", "programmer") is assigned by what the description asks for: 56% to backend, 9% to full stack, 2% to frontend, and the 33% that mentions no stack stays as "generic, no stack"; those four figures are estimates. Talent: Manfred community; if a person has several tech roles, their weight is split among them, so that column adds up to 100. "Other" (generic titles, no stack) has no equivalent in talent.

#### Our take

**The future isn't fewer roles: it's broader roles, with AI built in.**

Almost 6 in 10 professionals in the community come from frontend, full stack, backend or mobile, the natural pool for the product engineer: someone who touches back and front, understands the product and works with AI's help. Today you only catch glimpses of it, in offers asking for a plain "developer".

What can already be measured points to AI: AI Engineer grows twentyfold and the Forward Deployed Engineer appears. The rest, like the agent platform or product with AI, are our own signals or hypotheses, drawn from talking to professionals and companies in the sector. We mark them so you don't mistake them for data.

**Our bet.** The title that will grow most doesn't exist on job boards yet: someone who builds product end to end with agents alongside. Today they're called a plain "developer"; in two years they'll have a name of their own.

- **If you're job hunting:** If you come from backend or full stack, the shortest path is AI Engineer: it's the role with the most measurable offers. If you're QA or DevOps, look at the agent platform and evaluation.

- **If you're hiring:** Don't look for "an AI Engineer" in the abstract: decide whether you need someone who builds models into the product, who runs them, or who deploys them at the customer.

#### Our take

**Backend fits what companies ask for; Frontend doesn't.**

Backend is the case that fits: almost one in four professionals and one in five offers. Frontend is 14.7% of tech talent and only 3.8% of offers: it weighs four times more in talent than in offers.

That demand hasn't vanished; it has moved to a role with no surname: one in ten offers asks for a plain "developer", no stack.

Where it doesn't match is at the edges. After AI (5.5×), Security is the most pressured role: 6.3% of offers against 1.2% of talent. Product Manager is also asked for above supply (1.3×). Tech Lead and Engineering Manager go the other way: there are more profiles than offers.

**Our bet.** The line between frontend and backend will keep fading. The profile that fits in 2027 will be the one who builds product end to end; whoever only writes code or builds screens will find it harder every year.

> The AI row counts offers with AI or Machine Learning in the title, which covers everything from training models to integrating an API. We know, and that's a can of worms for another day. Even so, we've dared to break it down in Roles being born.

- **If you're job hunting:** If you're frontend, also look at developer, software engineer, product engineer, full stack and AI Engineer offers: that's where part of the demand is going.

- **If you're hiring:** Your next full stack is already working as a frontend. There's almost four times as much frontend talent as offers: if a full stack or product engineer is hard to fill, don't wait for the perfect profile. Hire a frontend keen to touch the back end and give them product context.

### What experience the market asks for

*Source: talent (Manfred community) + offers (market)*

**Judgement is what's wanted, but it's measured in years**

Offers don't know how to ask for judgement, so they ask for years. Here's how many, and what each band pays.

- Whole market (TheirStack): 83,623 of the 167,645 tech offers in the last 12 months (1 Oct 2025 to 30 Sep 2026) state years of experience, about half. We count the minimum (“3 to 5” counts as 3). Medians need at least 30 offers with a salary; below that we flag them as indicative. Brackets here are years asked, not job title (junior, senior…), so their medians differ from the by-title chart.

#### Our take

**The market mainly wants someone who already knows the job but doesn't cost much yet: 2 to 4 years.**

Judgement and years don't always go together, though we tend to treat them as the same thing. Over half of the offers stating years ask for 2 to 4 (55.5%), and that band's median matches the whole market: €34,500.

For newcomers it's even harsher than by title: when an offer states years, only 3 in 100 accept under 2.

And at the top something odd happens: the published salary almost flattens after five years. Experience doesn't stop paying: the companies that pay seniors best just don't show the salary.

**Our bet.** When the directive forces companies to publish ranges, that apparent ceiling after five years will vanish from the data. And many seniors will find out they were being paid below market.

- **If you're job hunting:** If you have more than five years, the published median is not your ceiling.

- **If you're hiring:** If you're looking for a mid-level profile with 3 to 5 years of experience, you compete with over half the market for the same profile.

### Junior talent

*Source: talent (Manfred community) + offers (market)*

**The junior generation at a crossroads**

One in five tech professionals has less than two years of experience. Many chose their career in the middle of the pandemic, with tech booming and the promise of good salaries. Now they find that only 4.4% of offers look for a junior, counting each role once per company and month. With "junior" in the title, the median is €25,500.

> ★ The most open door 14% of Data Analyst offers stating years accept fewer than 2. In Software Architect, Tech Lead and Engineering Manager, over 60% ask for 5 or more.

- Offers published on Manfred that state years of experience, 2022 to 2026 (2026, January to October). We take the minimum ("3 to 5 years" counts as 3). These are offers from Manfred clients, not the whole market, but they track the sector trend closely.
- In frontend, only 2.9% of offers are junior. Junior medians come from few offers with salary (16 to 46 per role): treat them as a guide. Junior: titles with junior, trainee, intern or graduate, counting each role once per company and month.
- Proportionally, consultancies and staffing look for juniors least: 3.3% of their offers, vs. 5–6% for the rest.

#### Our take

**Every year bootcamps, vocational schools and universities turn out thousands of people eager to get into tech. The market increasingly looks for people who already know the ropes.**

Junior offers are few and pay little: their median, €25,500, is barely above Spain's median salary (€24,497). Seven in ten don't even say what level they want. And only 4.4% look for someone just starting out. So beginners end up sending CVs blind. The bottleneck isn't talent: it's the first opportunity.

And it's getting worse. Among the offers we publish at Manfred, those open to someone with under two years were 7.8% in 2022; in 2026, under 1%. Those asking for five years or more went from 24% to 60%. The ladder is still there, but the first rung is missing.

And the door is badly placed. Consultancies post more junior offers than anyone because they post more of everything, but proportionally they ask for the fewest. And where most juniors are (in Andalusia, Castilla-La Mancha and Extremadura, more than one in four profiles has under two years) is where the fewest offers appear.

**Our bet.** AI won't shut the junior door: it will move it. With assistants writing the boilerplate, what a junior brings is no longer coding fast but understanding the business early and how their work affects it. Companies that get this first will have a pipeline; the rest, an ever pricier war for seniors.

- **If you're job hunting:** Today, Python and JavaScript are the most open doors to get started, and cybersecurity is the role that opens the most junior positions. But stacks change: what will serve you is understanding where the sector is heading.

- **If you're hiring:** Whoever doesn't train people today won't have seniors tomorrow. Every year thousands of people come out of bootcamps, vocational training and universities, and you rarely find them through a job ad. You find them at hackathons, meetups and communities: showing up pays off.

## 02 · Pay

*Show me the money (and the range, please)*

Talking about salaries isn't taboo. Here we set the salary of each role, level and company type against the market, where only 1 in 10 offers shows it. At Manfred, every offer has a range.

### Pay by role and experience

*Source: talent (Manfred community) + offers (market)*

**There's no single median backend salary. There are four, depending on experience**

Two views. Your role: the salary band each experience level asks for, with the market median in green. All roles: talent versus market, role by role.

- Talent: what the Manfred community asks for, in gross yearly salary bands by role and years of experience; each column adds up to 100%. Market: median of offers with salary by level (junior, mid, senior, lead), all roles.
- Talent: median of what the Manfred community asks for, estimated within the salary bands, by role and experience. Market: median of offers with salary over the last 12 months; management, product, architecture, systems and data use a broader filter. † Few offers with salary. Level row: approximate match (under 2 years ≈ junior; 2 to 4 ≈ mid; 5 to 9 ≈ senior; 10 or more ≈ lead or staff).

#### Our take

**Context: the Backend offer median is €37,500; talent asks for between €26,500 (under 2 years) and €57,700 (over 10). If you're hiring, look at your cell, not the average.**

The market median for backend is €37,500. Sounds reasonable until you put it on the map: it's what a backend with 2 to 4 years asks for (€37,900). With 5 to 9 years it asks for around €51,200, and with 10 or more, around €57,700.

We see it in many of the searches we open: the company arrives with "the market average" in mind and the profile it needs has six years. That's why every Manfred offer goes out with a range and an associated level. This chapter doesn't give you a number: it gives you a cell with much more accurate ranges.

**Our bet.** The conversation will stop being "what does the market pay" and become "what does it pay for this role, this level and this type of company". Whoever keeps setting pay with a single number will be late to every hiring process.

- **If you're job hunting:** Compare what you ask for with your cell (role and bracket), not with the market median.

- **If you're hiring:** Look at your cell, not the median. Offering the market median to someone with five years or more means offering what a 2 to 4 year profile asks for.

### Pay by company size and type

*Source: talent (Manfred community) + offers (market)*

**Size barely changes median pay; company type more than doubles it**

Left: median salary by company size. Right: Manfred's six salary tiers and where they fall. Heads-up: only one in ten offers shows a salary.

- All levels. Only offers that publish a salary count, one in ten (10,560 of all the development offers this table was built on).
- Source: The hexamodal distribution of tech salaries in Spain (Manfred, March 2026, in Spanish), checked in November 2026 against levels.fyi and public salary guides. For reference, on Levels.fyi the median total compensation of a software engineer in Spain is €56,225 (1,441 salaries), and at Microsoft, Datadog or Revolut it exceeds €100,000 (November 2026; includes variable pay). Indicative senior base salary; in tiers 5 and 6, variable pay lifts total pay a lot. Public sector (SP): career civil servant in the central administration with 5 to 9 years, groups A2 and A1 (the typical value is close to A1), using the official 2026 pay tables (Ministry of Finance, RDL 14/2025) and the specific allowance for ICT posts published in the BOE; no productivity bonus.

#### Our take

**Company size matters less than it seems: what changes pay is the type of company.**

By size, the market looks flat: from companies with up to 10 people (€31,500) to those with over 5,000 (€39,000), the median rises €7,500.

But a large consultancy and a big tech company can be the same size and pay in different worlds. According to our hexamodal analysis, which splits the market into six company types: around €40,000 versus around €90,000 base salary for a senior.

> They’re not the same measure: on the left, medians of offers at all levels; on the right, senior profiles.

And published salaries aren’t the whole market. From what we see every day at Manfred, the ones who publish salary most are consultancies and traditional companies (39% of offers come from consultancies); international scaleups and big tech rarely show it.

**Our bet.** Tech pay in Spain will be read more and more by company type than by city or size. Knowing which league you play in will be the first question in any job move.

- **If you're job hunting:** Big companies usually pay more, but not all of them, and they almost never say so. If it isn't published, the EU pay transparency directive requires them to give you the range before the first interview and bans asking about your current or past salary.

- **If you're hiring:** Publishing the salary sets you apart and supports internal pay equity and development plans. Several US states already require it by law: it's only a matter of time before we have to do it here too.

### Offered pay over time

*Source: offers (market)*

**Offered salaries: nearly frozen since 2023**

Median offered salary by role and year.

#### Our take

**Offered salaries have been almost stagnant for three years; only a few roles are breaking through that barrier.**

A backend developer was offered €39,000 in 2023 and €37,500 today. Counting only 2024 and 2025 inflation (2.8% and 2.7% on average, per INE), that €37,500 buys what €35,500 did in 2023: almost 9% less than three years ago. Only AI, Data Scientist and Data Engineer rise by more than €1,000 on average.

We've said for years that talking about pay shouldn't be taboo, and this year the data proves us right, but not for the reasons we'd like. The money hasn't vanished: it has moved to experience (from junior to lead the median doubles, from €25,500 to €55,000) and to a handful of companies that pay (and play) in another league.

**Our bet.** For most tech professionals in the country, adding AI to their title will become practically mandatory if they want to find opportunities. Companies can't yet tell the wheat from the chaff, so AI Engineers will start popping up everywhere. Senior profiles and early adopters are already seeing that these roles tend to pay more. Likewise, the scarcest roles (security and platform) will have more room to negotiate and improve their pay.

- **If you're job hunting:** The biggest pay rise comes from moving up a level, not from waiting for the market to rise. There are several levers, from languages to using AI, and all of them depend largely on how far you're willing to keep training and learning.

- **If you're hiring:** With the market standing still, applications per offer have shot up. Being able to pay above the median for the role you're hiring sets you apart more than ever to attract the talent you need, not just whoever is available.

### Signed salary vs. published range

**1 in 7 hires is signed above the offer's maximum**

A study of the hires closed by Manfred between 2022 and 2026, with the range published in the offer and the salary that was signed.

- **14%** sign above the published maximum.
- **74%** sign within the range.
- **12%** sign at the bottom of the range or below it.

- 0% is the published minimum and 100% the maximum. 2026, year to date.
- While the market's offered median has been nearly flat for three years, what actually gets signed is rising.
- Manfred hires with a published range and a signed salary (salaries between €20,000 and €200,000).

#### Our take

**The published range isn't a ceiling: it's a starting point.**

Until 2023, half of hires were signed below the midpoint of the range. In 2026, the median is signed at 67%. And one in seven goes above the published maximum: when the profile fits, the range stretches.

With the EU pay transparency directive, publishing the range will stop being optional. What we see in our processes is that publishing it doesn't freeze negotiation: on the contrary, it puts it where it should be for both sides.

- **If you're job hunting:** We usually ask for a 10% to 30% raise when we change companies, but it depends a lot on the context of the move.

- **If you're hiring:** Publish a range you can honour. The maximum you publish should be something you can actually pay if you find the right person.

### Offers in English

*Source: talent (Manfred community) + offers (market)*

**One in three tech offers in Spain is in English**

The language of the offer, what knowing English pays, and what talent asks for depending on their English level.

- † Only 2.4% of offers written in English publish a salary: handle with care. In Spanish-language offers, asking for C1 or above lifts the median to €37,500.
- Basic = A2, intermediate = B1-B2, advanced = C1, native = C2. Talent: median of what the Manfred community asks for by stated level.

#### Our take

**English is no longer a CV extra: a third of offers aren’t even written in Spanish.**

And it pays on both sides. In offers, a Spanish-language one that asks for English offers a median of €3,000 more; if it asks for C1, €4,500 more.

In talent, the gap grows with experience: with 10 years or more, a native or bilingual profile asks for 65% more than one with basic English (€31,300 more).

**Our bet.** The English-speaking part of the market will keep growing, because it's the part hiring remotely for abroad. For many profiles in Spain, English will make the difference between the stalled market and the one that moves.

- **If you're job hunting:** If you're wondering what to study this year besides Kubernetes: English.

- **If you're hiring:** If you ask for English, pay for it. The market already does, and not realising that this talent can work for any company in the world is holding you back.

## 03 · Where, who and how

*Remote is turning hybrid. Madrid and Barcelona versus the rest*

What talent is looking for and what companies offer. Where remote is heading. Where companies hire, who posts the offers and much more.

### Remote, hybrid and on-site

*Source: talent (Manfred community) + offers (market)*

**Hybrid offers win, but watch the trend**

- **52.5%** of talent wants fully remote. Adding hybrid, 88% want to work remotely, fully or partly (with hybrid).
- **67%** of offers are remote or hybrid: 33% fully remote and 34% hybrid.
- **21% vs 33%** accept on-site, vs. offers that are on-site or don't state a work model.
- **85% → 57%** fully remote offers at Manfred, 2021 to 2025. Hybrid climbs from 12% to 38%.

- In the market, "on-site" includes offers that don't state a work model. Talent: Manfred community preferences (more than one could be chosen; normalised to 100). Only 0.6% want on-site only. Manfred: 2026 offers (January to September).
- Offers with a full salary range. Hybrid and on-site figures are approximate: TheirStack corrected InfoJobs' work-model label.

#### Our take

**Remote wins: 52.5% of talent wants it fully and 67% of offers let you work from home, fully remote or hybrid.**

Talent is clear: 52.5% want fully remote. The most offered model is hybrid (34%), ahead of fully remote (33%), and one in three offers is on-site or doesn't say. 67% allow working from home, but half of those only on some days. Only 0.6% of talent wants office only.

We see it in our own offers: at Manfred, fully remote has fallen from 85% to 57% in four years, and hybrid takes almost everything it loses. Going to the office isn't paid better: the three models offer similar pay, from €34,500 remote to €36,000 hybrid.

It depends on where you live: outside Madrid and Catalonia, over 90% want remote work and hybrid convinces far fewer people. We break it down below, region by region.

**Our bet.** Hybrid is already the sector standard, but with a high cost: every office day narrows your market to people living an hour or less from your office. Curiously, more companies are asking for 100% on-site and benefiting from the state of the sector to attract talent that, very recently, wouldn't have accepted that model. Companies that keep real remote will have half of Spain to choose from.

- **If you're job hunting:** Remote doesn't mean being a lone wolf. Teamwork is critical in this model, especially now that so many interactions are with an LLM's context window rather than with a colleague to debate which solution is best or how to solve a problem.

- **If you're hiring:** Flexibility is the cheapest lever for attracting talent, and every office day you add is a filter.

### Map: talent, offers and companies

*Source: talent (Manfred community) + offers (market) + companies (Manfred Spain Tech Map) · only at Manfred*

**Half of talent lives outside Madrid and Catalonia**

Three layers on one map: where talent lives (by role and city, with what it asks for), where offers are and where companies with a tech team are. 49% of tech talent lives outside Madrid and Catalonia, yet only one in five offers are there.

#### Our take (offers)

**Madrid takes half of the offers that state a region, but Barcelona pays €4,500 more at the median with half the volume.**

The surprise is Málaga: it’s already the third city by number of offers, ahead of Valencia and Seville. Málaga as a tech hub is a reality.

**Our bet.** Málaga could be an example for the future; it's only the first case. Wherever product offices and talent that wants to stay come together, the same will happen.

- **If you're job hunting:** Only one in three offers is fully remote; the rest are hybrid or don't say, so you'll often need to live nearby or be willing to commute. If you can't move, your opportunities shrink a lot.

- **If you're hiring:** Madrid and Barcelona compete for the same talent with far more offers than the rest of Spain. That drives more turnover than anywhere else: keeping your talent is critical.

### Talent and offers by region

*Source: talent (Manfred community) + offers (market)*

**Talent is spread out. Offers aren't**

Distribution of talent and offers by region. On the right, what talent in each area asks for and how it wants to work.

- Talent: Manfred community, Spain-based profiles with a tech role. The split by region is calibrated against INE employment in IT activities (Madrid, Catalonia, Andalusia and Valencia region; the rest as a block). Offers: split of the 137,437 tech offers from the last 12 months that state a region. Another 30,305 only say "Spain": 80% are remote, although over half of remote offers do state a region. What it asks for, work model and juniors are calculated within each region, with 30 or more profiles; "Open to hybrid" is the share that ticks hybrid among their options.

#### Our take

**Madrid has almost two offers per point of talent. The Canary Islands, one per seven.**

Half of tech talent lives outside Madrid and Catalonia, yet only one in five offers are there. Andalusia, the Valencia region and Galicia hold 28% of talent and 13% of offers.

That talent knows it: outside Madrid and Catalonia, over 90% want remote work and hybrid convinces far fewer people (from 37% in the Canary Islands to 62% in the Valencia region, vs. 72% in Madrid and Catalonia). Every office day added to an offer weighs more the further from Madrid the candidate lives.

And it asks for less: the median goes from €45,100 in Madrid to €37,000 in Andalusia and €33,900 in Extremadura. With 10 years or more, from €63,200 to €55,000. Each area also has its specialty: Madrid concentrates product, design and data; elsewhere backend, full stack and QA weigh more; Galicia has almost twice the average share of SysAdmins and the Basque Country stands out in AI and Data Science.

**Our bet.** There's a clear opportunity to hire outside Madrid and Barcelona: plenty of professionals and few offers.

- **If you're job hunting:** If you live outside Madrid or Barcelona, your market is remote: benchmark against what's asked in Madrid, not just in your area.

- **If you're hiring:** If you only hire in Madrid, you compete for 28% of talent with half of the country's offers. Remote gets you the rest, and in Andalusia, the Valencia region or Galicia people ask for 13% to 18% less.

### Benefits in offers

*Source: offers (market)*

**One in two offers doesn't mention a single benefit**

What share of offers mentions each benefit. On the right, its rank among benefits in Manfred offers in 2025.

> ★ Half of them, blank 52% of tech offers don't mention any specific benefit. Health insurance, flexible pay and flexible hours appear most. Bonus, in 5%; stock options, in 1.5%.

- There's no standard: specific expressions are searched in the description. 48% mention at least one. "Not ranked": not in Manfred's ranking.

#### Our take

**If an offer states neither salary nor benefits, candidates have no reason to apply.**

Health insurance and flexible pay are now the expected minimum, and still only appear in one in four offers.

In our offers, a training budget ranks third, behind team-building and health insurance. In the market, training appears in fewer than one in five offers.

**Our bet.** With pay stalled, benefits will be where companies compete. And training, with AI changing the job every few months, will go from extra to essential.

- **If you're job hunting:** If the offer doesn't say, ask: health insurance and flexible pay are now the minimum.

- **If you're hiring:** The cheapest things to mention are the ones most often forgotten. Flexible hours cost nothing and only 21% of offers mention them.

### Who posts the offers

*Source: offers (market)*

**One in two offers comes from a consultancy or agency**

Who is behind tech offers in Spain.

- **39%** posted by consultancies and IT services firms. Half of them have 201 to 1,000 employees.
- **15%** posted by recruitment and staffing agencies.
- **21%** gets reposted: same company, same title, in at least three different months.
- **1.7%** are duplicates that slip through after removing the same offer repeated within 30 days.

#### Our take

**More than half of tech offers are posted by a consultancy or an agency, not by the product company where you’d be coding.**

One in five reappears in at least three different months of the year. It could be a role that isn’t filled, serial hiring, or an offer to keep candidates on hand before a project is signed.

We can’t yet tell one from the other: that requires knowing how long each offer stays open, and TheirStack has only tracked that since this summer. The first snapshot is just below.

**Our bet.** Once we have a full year of data on how long each offer stays open, we'll be able to separate live vacancies from “ghost” offers. We suspect there are more of them than it seems.

- **If you're job hunting:** Take a close look at offers that keep reappearing or have been up for months: they may be “ghost” offers.

- **If you're hiring:** Your competition isn't just the startup across the street; it's also the consultancy that has been posting your exact role for three months.

### How long it takes to fill a role

*Source: offers (market) · only at Manfred*

**Filling a tech role takes a median of over 40 days. Leadership and architecture roles take the longest**

Companies are finding it ever harder to hire, largely because of the huge volume of offers in the sector, even if they aren't real. Applications have shot up this past year and the time teams spend interviewing keeps growing, largely because we have all had to adapt to AI on both sides. More on that later.

A tech hiring process takes a median of 42 days to close successfully, and most of that time is lost to internal bureaucracy and technical tests. Companies rarely hire the first candidate they interview, even if that person is perfect.

- **42 days** median, from active search to hire.
- **28–72 days** is what half of processes take.

**Market · about 6,900 LinkedIn and Indeed offers (August–September 2026)**

| Role | % still open after 30 days |
|---|---|
| Lead / Architecture | 52% |
| Backend | 51% |
| Cybersecurity | 50% |
| **All** | **45%** |
| Mobile | 40% |
| Full stack | 39% |

#### Our take

**Filling a lead, architecture, backend or cybersecurity role takes longer: half of those offers are still open after a month.**

In lead, cybersecurity and data engineering, one in four is also reposted for months. Full stack and mobile close sooner: about ten days at the median.

> Take it as a first snapshot: the sample is small, only covers LinkedIn and Indeed, and an offer closing doesn’t mean it was filled.

**Our bet.** The profiles that take longest to fill (lead, architecture, cybersecurity) are exactly the ones AI makes more necessary, because they review and decide. Those timelines will grow, not shrink.

- **If you're job hunting:** An offer reposted for months isn't always a live vacancy.

- **If you're hiring:** If you're looking for a lead, an architect or a cybersecurity profile, expect more than a month, and expect someone else to have been looking for the same for a while.

## 04 · Stack

*Plenty of JavaScript. Not enough Kubernetes.*

Which technologies talent knows, which ones the market demands and which are rising or falling. The big shift: tech offers with AI in the title went from 1.5% in 2023 to 10.5% in 2026 (to September).

### Technologies: talent vs. offers

*Source: talent (Manfred community) + offers (market)*

**Demand for Kubernetes is twice its talent supply**

The tension index compares how much talent uses each technology with how many offers ask for it. Above 1×, more is asked for than exists; below it, there's talent to spare.

> ★ Hated because it's everywhere Java 25.2% of offers ask for it and 26.5% of professionals flag it as a technology they don't want to work with. Not a contradiction: it's one of the most widespread stacks out there, so plenty of people have already battled with it (and with some 2009 monolith). Not wanting to go back doesn't mean you won't: 37.9% of talent work with Java. If you have to, you have to.
> ★ Oversupplied 0,36× JavaScript: 45% of talent uses it and 16.4% of offers mention it, a share that also dropped 3.8 points in a year.

- Talent: % of professionals working with the technology, among those who state their stack. Offers: % of offers mentioning it. Not the same base, so the index shows direction and order, not an exact ratio.

#### Our take

**Spanish tech talent was trained on the web, and the market has moved to the cloud.**

There are plenty of hands in JavaScript, React and CSS, because that’s where almost everyone starts programming. There’s a shortage in Kubernetes, Azure and Terraform, which you don’t learn in a twelve-week bootcamp but by wrestling with production systems at three in the morning.

**Our bet.** The gap between web and cloud won't close with more bootcamps. It will close inside companies, by training the people they already have.

- **If you're job hunting:** If you're thinking about what to learn next year, look at Kubernetes, Azure or Terraform: offers ask for 1.7 to 2 times what exists.

- **If you're hiring:** In cloud you compete for scarce talent. Training people on your team may pay off more than waiting for the perfect candidate.

### Talent community by technology

*Source: talent (Manfred community)*

**What the market wants most is in veterans' hands**

What share of the community works with each technology, and with how much experience.

- Talent: Manfred community, % of professionals working with each technology (among those who state their stack) and how the people using it split by years of experience.

#### Our take

**Veterans dominate systems technologies, and those are exactly the ones the market asks for beyond what exists.**

In Kubernetes, fewer than one in ten have under two years. In Terraform, two in three have five or more. And in Ansible, 40% have ten or more.

At the other end, Pandas and React are the entry point: almost a third of Pandas users and one in four React users are juniors.

**Our bet.** Without succession now, in a few years the cloud bottleneck will be people, not technology: today's systems veterans will be managers or close to retiring.

- **If you're job hunting:** If you come from frontend, the road to the cloud exists, but it's a long one.

- **If you're hiring:** If you're hiring for systems, you're hiring seniors. Set salary and expectations accordingly.

### Stack trends

*Source: offers (market)*

**The market switches stacks: cloud up, JavaScript down**

How much weight each technology has gained or lost in offers: the last 12 months versus the previous 12, in points.

- 2026, January to September: that's why AI shows 16.9% here and 15.7% over the last 12 months.

#### Our take

**AI is what grows most in offers: it gains almost 10 points and now appears in 15.7%.**

Next come CI/CD, Python, Docker and Kubernetes. And the cloud isn’t a one-year fad: Kubernetes has almost doubled since 2023, and so has Terraform, from 3.7% to 7.2%.

What falls is exactly where talent is most crowded: JavaScript, CSS and Java.

**Our bet.** AI and cloud aren't this year's fad: they've been rising together for three years. The stack most in demand in 2027 is the one that combines both.

- **If you're job hunting:** If you're choosing what to learn, AI, CI/CD and cloud are rising in offers; JavaScript, CSS and Java are losing weight.

## 05 · Women

*Women in tech: the gap won't close on its own*

How many women are there in tech, and what salary do they ask for? In the Manfred community, women make up 15.6% of tech profiles. And the share varies a lot depending on where you look: 18% in Madrid, 8% in Murcia. To understand the pay gap, we look at something job offers don't capture: how much each person wants to earn in their next job.

### Women in tech by role

*Source: talent (Manfred community)*

**Only one role reaches parity**

Share of women in each role. The line marks 50%.

### Pay gap

*Source: talent (Manfred community)*

**Early in their careers, women and men ask for the same. From ten years of experience on, what women ask for is 13% lower at the median.**

We compare, by role or by years of experience, the median salary women and men ask for to change jobs. Women are 15.6% of the Manfred community's tech talent.

- The data comes from the Manfred community and reflects the salary its professionals ask for to change jobs, not their current salary.

### Gender pay gap: analysis

*Source: talent (Manfred community)*

**Early in their careers, women and men ask for the same salary. With 10+ years of experience, the salary women ask for is 13.4% lower at the median.**

What women and men ask for by years of experience and by role, and how to read it.

- **0.4%** difference with under 2 years of experience.
- **13.4%** median gap with 10+ years: about €8,300 a year.
- **19 of 19** roles where women's median is lower.

#### Our take

**The gap isn't there at entry: it opens with experience and shows up in every role.**

With under 2 years of experience, women and men ask for practically the same (€27,000 vs €27,100: a 0.4% difference). From there the distance grows at every step: 4.9% at 2–4 years, 10.5% at 5–9 and 13.4% at 10+, about €8,300 a year.

By role, women's median is lower in all 19 positions. The difference is largest in Security (22.5%), Full Stack (22.3%), AI (19.0%), and Backend (18.6%), and smallest in Product Design / UX (4.9%), Data Analyst (6.8%), and Data Scientist (7.8%).

To compare what women and men ask for, we start from the salary professionals in the Manfred community want to earn when changing jobs, which may differ from their current salary. When looking at the data by role, bear in mind that each role brings together people with different levels of experience and that, in seven of them, women are under 10% of the talent. In those cases, the median of what women ask for rests on few answers. So rather than holding on to a single figure, it's worth looking at how the gap changes with years of experience. The more years of experience, the bigger the difference between what women and men ask for.

That the gap shows up with experience and not at entry points to what happens during a career, not at the starting point. These data can't separate the causes, but they do confirm that the gap exists and can be measured: measuring it every year is the first step to closing it.

- **If you're job hunting:** With 5+ years, women's median is 10% to 13% lower. If you negotiate, start from the median for your role and bracket, not from your last payslip.

- **If you're hiring:** Set pay by role and level and publish the range: that way the gap isn't carried from one job to the next. The EU pay transparency directive points the same way.

## 06 · AI

*Companies ask for it. People use it. But hardly anyone teaches it.*

AI gets its own chapter this year: how the community uses it, how technical interviews are changing, what offers ask for and what new roles it creates.

### AI in everyday work

*Source: talent (Manfred community)*

**Almost everyone uses it, but only one in five has been trained**

We asked our community which tools they use, how much freedom they have, how the code turns out and what they fear.

- **98%** use AI in their daily work.
- **86%** say AI makes them more productive.
- **5.8%** of tech offers ask for working with code assistants.
- **20%** have received AI training from their company. 45% learned on their own.

- Cada persona que usa IA maneja de media 3,1 herramientas.

#### Our take

**Almost every developer uses AI every day, but their company hardly ever teaches them how.**

98% use AI and 86% say it makes them more productive. But only 20% have received training from their company: 45% learned on their own and 35% haven’t been trained.

Company size changes almost everything. At small companies, 81% have total freedom to use AI and Claude leads (64%). At companies with over 1,000 people, only 48% do and GitHub Copilot leads (50%), even though that's where most training happens (32% vs 11%).

At Manfred we see it every day, and it matches the survey one hundred per cent. Installing the tool is easy; knowing how to use it wisely is another story. That's why few people trust it, and only 21% use AI-generated code untouched. There's fear too: 45% to 51% think AI will wipe out pure coding, whether you're just starting out or have been writing code for over twenty years.

> A word on the sample: this is the Manfred community, which already works with AI and is very experienced (56% have over ten years). It doesn’t represent the whole market.

**Our bet.** Within two years, AI training will be negotiated the way remote is today. Companies that don't offer it will lose the people who get the most out of it.

- **If you're job hunting:** Don't wait for your company to train you: learning on your own is the most common route.

- **If you're hiring:** Offering training and freedom to use AI is an advantage: at companies with over 1,000 people, one in five can't use it.

### The technical interview in the age of AI

*Source: talent (Manfred community) + companies (conversations with companies and public examples)*

**Coding unaided matters less; directing and reviewing the AI matters more**

84% of the community would allow AI in a technical test, and companies are already moving. These are the four interview formats we see emerging.

- **Before: Unaided algorithms.** Whiteboard, LeetCode-style exercises and a take-home. They measured pattern memory and speed.
- **Now: AI allowed, in plain sight.** Rounds where the candidate works with an assistant. They measure how you break things down, what you ask and how you verify.
- **Trend: Judgement and ownership.** You defend what you delivered: fundamentals, architecture and decisions. Exactly what is hardest to find.

- **84%** of the community would allow AI in a technical test.
- **46%** of the companies we talked to have redesigned their technical test because of AI.
- **67%** of those companies have dropped the take-home or the algorithmic exercise.

Would you allow AI in technical selection tests? Yes, it's part of daily work 52%, Yes, with limits 32%, No: it makes spotting talent harder 16%.

1. **Live coding with an AI assistant.** The candidate works on an existing project with an AI assistant at hand and is observed while working and deciding. Measures: Breaking down the problem; What you ask it; Verifying what it returns. Used by: Meta; Canva.
2. **Reviewing and debugging AI-generated code.** You get code written by an AI and have to find its flaws, fix them and bring it up to production standard. Measures: Reading someone else's code; Spotting errors that look right; Production standards. Used by: Canva; Shopify.
3. **Take-home with AI and a live defence.** You use the tools you would use on the job, AI included, and then explain your decisions and your prompts in a conversation. Measures: Ownership of the result; Explaining prompts and decisions; Real understanding. Used by: Shopify; Companies we talked to.
4. **A problem that changes mid-session, and systems design.** The brief changes halfway through or you're asked to design a system. Here AI helps less and you can tell whether the fundamentals are there. Measures: Fundamentals; Architecture; Judgement under change. Used by: Manfred data: companies we talked to describe it.

It isn't universal: some companies keep AI-free rounds to measure fundamentals, and some allow AI in one round and ban it in another. The named examples are public and policies change. Sources and method, in “How it's made”.

#### Our take

**The interview no longer asks whether you can write code unaided: it asks whether you can direct, check and defend what AI writes.**

84% of professionals would allow AI in a technical test, and 52% see it as part of daily work. Of the companies we talked to, 46% have redesigned their test and, of those, 67% have dropped the take-home or the algorithmic exercise.

It's the hottest topic this year in our conversations with companies. Technical tests no longer look for speed but for judgement: if an AI solves it in two minutes, you're measuring the tool. If it forces you to break down, review and defend, you're measuring the person.

**Our bet.** The classic take-home test's days are numbered. It will be replaced by something much closer to real work: a problem, an AI at hand and someone asking why.

- **If you're job hunting:** Practise with the AI in front of you and analyse what you ask it, what you verify and why you discard what doesn't convince you. "The AI suggested it" is not an answer.

- **If you're hiring:** Decide what you want to measure. If an assistant solves your test in minutes, you're measuring the assistant: move to evaluating how a problem is broken down, reviewed and defended.

### AI, from the inside

*Source: talent (Manfred community) + offers (market)*

**One in six offers asks for AI, but only one in sixty profiles holds an AI role**

We wanted to see how much AI weighs in job offers, what they really ask for and how many people show it on their profile. The problem is that many profiles aren't up to date (especially among people not looking for a job), which makes it hard to tell who already uses AI day to day. That's why we ran the AI survey.

- Offers from the last 12 months. An offer can mention several things. "AI in the description" (15.7%) is not the same as "AI in the title" (10.5% in 2026): the former includes backend or data offers asking for some AI.
- Offers with AI in the title, October 2025 to September 2026. One offer can ask for several things, so it doesn't add up to 100.

#### Our take

**AI is no longer a niche: one in six offers mentions it and those asking for agents have multiplied eightfold in a year.**

What they ask for isn’t the classic data scientist: two in three AI offers ask for Python, almost one in three asks for LLMs or RAG and more than a third talks about agents. LangChain is now requested more than TensorFlow.

Talent hasn't kept that pace: in the community, only 1.7% hold an AI or machine learning role, and one in twenty already works with frameworks such as TensorFlow or PyTorch. What is spreading is using AI as a tool: 5.8% of offers already ask for working with code assistants.

**Our bet.** The most in-demand AI profile won't come out of a machine learning master's. It will come from product teams who already know backend and have learned to put models into production.

- **If you're job hunting:** You don't need to retrain as an ML engineer; understanding how AI fits into what you already do is enough.

- **If you're hiring:** There are very few pure AI profiles. Look for people who already build product and want to bring AI into it.

### Roles being born

*Source: offers (market)*

**AI Engineer grows twentyfold. And the Forward Deployed Engineer arrives**

We split AI into three profiles: who builds models into products (AI Engineer), who trains them (ML Engineer) and who deploys them with the customer (Forward Deployed Engineer).

- Each bar is the share of that role's offers that mention the technology. In bold, the role that asks for it most.
- Offers in the last 12 months within "Development, other titles".
- Trend: offers from January to September of each year, per 1,000 tech offers, so it doesn't depend on how many offers TheirStack collects. AI Engineer includes AI Developer and LLM Engineer; "AI/ML Engineer" counts as ML Engineer. Bending Spoons posts 406 AI Engineer offers repeated across cities: we count each position once per company and month. The FDE median comes from 23 offers with a salary: treat it as indicative.

#### Our take

**We've now put a name to the elephant in the room. Today there are almost three times as many AI Engineer offers as ML Engineer ones.**

The market asks for things that don't look alike: agents, RAG and LangChain on one side; MLOps with PyTorch on the other. The AI jobs being hired in Spain are no longer for people who train models, but for those who help build them into the product.

A profile that didn't exist two years ago also appears: the Forward Deployed Engineer, half engineer, half consultant, who sits with the client to deploy. Offers are few, almost all in English and with high salaries.

We notice it in the searches that reach us: more and more companies ask for "an AI Engineer" and, when you dig, want very different things. The title has become a wildcard.

**Our bet.** The Forward Deployed Engineer will arrive in Spain with the companies that sell AI to other companies. Today there are few offers, but in two years it will be one of the most fought-over profiles, because it combines two things that rarely go together: knowing how to build and knowing how to talk to the client.

- **If you're job hunting:** If you come from backend, AI Engineer is closer than it looks: what's asked for most is integrating, not training.

- **If you're hiring:** Before posting "AI Engineer", decide whether you want someone who trains models or someone who integrates them. They are different profiles.

### AI first

*Source: talent (Manfred community) + offers (market)*

**Would you rather ask your AI?**

You can read this report with your AI too. Take all of it to your favourite LLM and ask whatever you like: it has the report's figures, so it doesn't need to make any up.

- Try: "What should a backend developer with 6 years earn?" · "Which technology should I learn this year?" · "How long will it take to hire a tech lead?"

## 07 · What the mirror tells us

*Two markets under one name, and neither is easy to get into*

We put talent in front of the market and they don't recognise each other. The mirror shows two markets under one name: a big, nearly frozen one, and a small one that hires with AI and pays more than double in base, and up to three and a half times with variable pay. Between them, the way in narrows.

### 1. Two markets under one name

*Market* · **×3.5** is what a senior profile is paid at big tech, counting variable pay, versus a small consultancy (more than double in base alone)

More than half of all offers are posted by a consultancy or an agency, more than half of those asking for years of experience want 2 to 4, and only one in ten shows the salary publicly, which has barely moved in three years. At the other end, international scaleups and big tech pay that senior profile more than double a small consultancy's base, and up to three and a half times counting variable pay. That's why the median misleads: what changes pay isn't company size, it's company type.

> “The market has parted like the Red Sea: there are those who pay a lot and those who pay very little. The middle is disappearing.” (Co-founder of an AI SME)

### 2. The way in isn't just narrowing: it's moving

*Entry* · **7.8% → 0.5%** of Manfred offers accepted someone with under two years of experience (2022 → 2026)

One in five tech professionals has less than two years of experience and only 4.4% of offers look for a junior profile. In Manfred offers, those open to someone with under two years of experience went from 7.8% in 2022 to 0.5% in 2026. The companies using AI the most put it bluntly: if you used to hire ten and now hire one, better make it a senior. Yet the same conversations raise another idea: the "junior" label matters less and less, and what companies want is hungry people who can show what they've built. The way in is no longer the title or the years worked: it's what you're capable of doing, plus fundamentals AI won't give you. And whoever doesn't train people today won't have seniors tomorrow.

> “If you used to hire ten and now you hire one, I'll go for a senior.” (Head of Engineering at a European e-commerce company)

### 3. Talent trained on the web; the market moved to cloud, AI and security

*Stack* · **5.5×** what is asked for in AI vs. the talent available: the biggest mismatch in the report

Frontend-focused profiles make up 14.7% of talent and 3.8% of offers; 45% of talent uses JavaScript while 16% of offers ask for it. Meanwhile AI (5.5×), security (5.3×) and Kubernetes (2×) are in demand far beyond supply, and what's scarce is in veterans' hands: in Kubernetes, fewer than one in ten has under two years. The problem isn't how much talent there is, but where it sits: too much on the web, too little in cloud, AI and security.

> “Good electronics or PLC engineers are really hard to find. Full stacks, you'll find fourteen under any rock.” (CTO at an industrial company)

### 4. Using AI no longer sets you apart; reviewing it does

*AI* · **20%** has had AI training from their company, although 98% use it

98% of talent uses AI and 86% say it makes them more productive, but only 20% have been trained by their company and only 21% use what it generates almost untouched. In the teams using it most, we're told that far more code gets shipped and, therefore, far more gets reviewed: the bottleneck has moved from writing code to reviewing it. Amazon has just said so in its guideline for engineers: the role isn't defined by how much AI you use, but by impact on customers and the business, and by being able to explain where it doesn't help.

> “The bottleneck is now in the review.” (Engineering lead at a software scaleup)

### 5. The classic technical test is dead; the new one measures judgement

*Interviews* · **46%** of the companies we talked to have redesigned their technical test because of AI

84% of the community would allow AI in a technical test, and hiring processes are changing: of the companies we spoke to, 46% have redesigned their technical test because of the arrival of AI and 67% of them have dropped the take-home or the algorithm exercise. Instead: live sessions with AI, reviewing flawed AI-generated code, explaining the prompts, or solving a problem that changes as you go. What gets assessed now is fundamentals, architecture and judgement: exactly what's hardest to find.

> “These days an offline technical test makes no sense, because it tells you nothing.” (CTO at a data scaleup)

### 6. Remote isn't ending: it charges a toll in office days, and half the country pays it

*Remote* · **88% vs. 67%** of talent wants remote (fully or partly), vs. 67% of offers that are remote or hybrid

88% of talent wants to work from home at least part of the time and 67% of offers allow it, but half of that 67% is hybrid, and hybrid means living nearby. With half of all talent outside Madrid and Catalonia and only one in five offers there, every office day is a geographic filter. Companies that open up to remote reach more people.

> “We've had to open up remote again for technical profiles.” (People lead at a 20-person tech company)

### 7. Publishing the salary is no longer a gesture: it's an advantage, and soon an obligation

*Pay* · **1 in 10** tech offers show the salary

Only one offer in ten makes the salary public, and the deadline to transpose the EU directive that requires it before the interview passed in June. The companies we spoke to that already publish bands see it as an edge: less time wasted, and hardly anyone drops out over money. The gap makes it urgent: in backend profiles, women's median ask is 19% lower, and setting pay by what each person asks for means inheriting that gap.

> “Nobody has dropped out on me over money.” (CTO at a data scaleup that publishes its salary bands)

These takeaways cross the report's data with thirteen conversations with CTOs, engineering leads and People leads at companies hiring tech talent in Spain (October 2026). Quotes are anonymous, by role.

## Spain vs. the world

*Your commit is worth double in Berlin*

Median offered in Berlin: €76,000. In Madrid: €34,900. Development offers from the last 12 months, in euros. (The filter differs from chapter 03, where Madrid comes out at €34,500.)

*Source: offers (market)*

- In the US many more offers publish salaries due to state pay transparency laws; in the UK, by market convention. Dollars to euros at 0.8608.

## The questions you ask us every year

### How do you find a tech job in 2026?

Keep four facts in mind. Over half of offers are posted by a consultancy or an agency, not the company you'd be coding for. One in five gets reposted for months, so ask how long it has been open. One in three is written in English. And only one in ten shows the salary (at Manfred, all of them do). If you're starting out, the funnel is even narrower: only 4.4% of offers look for a junior.

### How long does a tech hiring process take?

On the company side, a median of 42 days from activating the search to hiring, and half of processes take between 28 and 72 days (Manfred processes in 2025 and 2026). On the candidate side it's usually a bit shorter: expect five to six weeks from applying to accepting an offer (a 40-day median for technical roles, according to Ashby). An offer disappearing sooner doesn't mean it was filled: on LinkedIn and Indeed, half of full stack and mobile offers close within about 10 or 11 days, and among lead and architecture offers that do close, the median is 19 days (over half are still open after a month).

### Are there ghost job offers?

One in five tech offers (21%) reappears in at least three different months of the year, with the same company and title. It may be a role that isn't getting filled, serial hiring, or an offer to keep candidates on hand before a project is signed. In lead, cybersecurity and data engineering, it's one in four.

### What do technical tests look like today?

They're changing because of AI. Of the companies we spoke to, 46% have redesigned their technical test and, of those, 67% have dropped the take-home or the algorithm exercise. Instead: live coding with an AI assistant, reviewing flawed AI-generated code, defending an AI-assisted take-home live, or solving a problem that changes as you go. Less memory, more judgement. And 84% of professionals would allow AI in the test.

### How do tech professionals use AI?

98% use it daily and 86% say it makes them more productive, but only 20% have been trained by their company: 45% learned on their own. The most used is Claude (56%), followed by Claude Code (42%) and GitHub Copilot (31%). And hardly anyone trusts it blindly: only 21% use the code it generates without changes. Survey of 756 professionals in the Manfred community.

### Which tech roles are being born?

AI Engineer has grown 21-fold since 2023 (3,311 offers from January to September 2026), and companies now ask for almost three times as many AI Engineers as ML Engineers: the AI being hired is the kind that goes into product, not the kind that trains models. The Forward Deployed Engineer also appears, half engineer and half consultant, deploying with the client: still few offers (273), almost all in English.

### Which tech roles pay the most?

By offered median, management and architecture: Engineering Manager, €79,000; Software Architect, €48,000; Product Manager, €47,500; Tech Lead, €45,000. Then AI and Data Scientist (€42,500). At the other end, Data Analyst (€30,500) and QA (€31,500). Few offers publish pay, so treat it as a guide.

### How much is a backend developer offered in Spain in 2026?

The median of offers with a published salary is €37,500 gross per year, across all levels. Among talent, a backend with 2 to 4 years asks for around €37,900, and with 10 or more, around €57,700.

### Which technologies have more demand than talent?

Kubernetes (2.0×), Azure (1.7×), Terraform (1.7×), Google Cloud (1.6×) and AWS (1.2×) are asked for in offers more than they appear among tech talent.

### How many tech job offers in Spain are remote?

33% of tech offers are remote and 34% hybrid: in total, 67% allow working from home. 52.5% of tech professionals prefer fully remote; on-site is only 11.7% of preferences, and just 0.6% want on-site only.

### Do big companies pay more?

A little: the offered median goes from €31,500 at companies of up to 10 people to €39,000 at those over 5,000. Company type matters far more: per Manfred's hexamodal analysis, a senior ranges from around €40,000 base salary at small consultancies and SMEs to around €90,000 at big tech, and around €140,000 counting variable pay.

### How much does English pay?

One in three tech offers in Spain is written in English. In Spanish-language offers, asking for English lifts the median from €33,000 to €36,000, and asking for C1 or above, to €37,500.

### How much is a junior offered?

The median for offers with "junior" in the title is €25,500. Only 4.4% of offers ask for juniors; one in five tech professionals has less than two years of experience.

### Which tech role is growing fastest?

AI: from 1.5% of tech offers in 2023 to 10.5% in 2026. Followed by Security (+2.6 points) and Software Architect (+1.3 points).

### Who posts tech job offers?

39% are posted by consultancies and IT services firms, and 15% by recruitment agencies. 21% are reposted with the same title in at least three different months.

### Where are most tech offers and where is pay highest?

Over a third of development offers are in the city of Madrid (36,868 of 102,792). Barcelona has the highest median among large cities: €39,000, vs. €34,500 in Madrid.

### How much does tech talent in Spain earn?

Half of the Manfred community's tech talent earns €40,000 gross a year or more, 1.6 times the country's median salary (€24,497, INE 2024). What it asks for to change jobs ranges, at the median, from €35,600 in frontend to €69,500 for Engineering Manager.

### Where does tech talent in Spain live?

Half of it lives outside Madrid and Catalonia. Madrid holds 28% of talent and Catalonia 23%, but together they take 79% of offers that state a region.

### Which tech roles are hardest to fill?

AI and Security. AI is 9.6% of tech offers and only 1.7% of talent (5.5×); Security, 6.3% of offers and 1.2% of talent (5.3×). Among technologies, Kubernetes is in demand at twice its supply.

### How many women work in tech?

Women are 15.6% of the Manfred community's tech talent. Only Product Design / UX reaches parity (50%) and Software Architect stays at 5%. In backend, the median woman asks for 19% less than the median man (€34,600 vs. €42,500).

### What is the Tech Career Report 2026?

It's Manfred's annual report on Spain's tech job market. The 2026 edition is the first to cross the Manfred community (127,665 professionals) with 167,645 tech offers from the last 12 months, analysed with TheirStack: pay by role and experience, juniors, remote work, English, technologies, AI, territory and the gender gap. It's free and its data is published under CC BY-NC 4.0: you can reuse it for non-commercial purposes, always crediting the source.

### How much does a developer earn in Spain in 2026?

In tech offers that show a salary, the median is €25,500 for junior, €34,500 when the offer sets no level, €40,500 for senior and €55,000 for lead, staff or architecture. Half of the Manfred community's tech talent earns €40,000 gross a year or more. Only one in ten offers publishes the salary.

### How much does Big Tech pay in Spain?

For a senior profile, the usual base salary at Big Tech runs from €80,000 to €130,000 (typical, about €90,000), and the wide range goes from €75,000 to €200,000 depending on whether pay is national or international. With variable pay, total pay runs from €110,000 to €250,000. That is more than twice the base at small consultancies and SMEs.

### How much does an IT professional earn in Spain's public sector?

It depends mostly on the civil-service group and the post, more than on years. Group C1 (assistant technician, e.g. TAI): about €22,500 on entry and about €30,000 with 10 years or more. Group A2 (management, e.g. GSI): about €31,000 on entry and €41,000 with 10 or more. Group A1 (senior corps, e.g. TIC): about €44,000 on entry and €55,500 with 10 or more, up to €80,000 in senior posts or better-paying administrations such as Madrid City Council. Gross annual civil-servant pay in 2026, excluding productivity pay.

### Do remote jobs pay more than office jobs?

Almost the same. In tech offers with a salary, the median is €34,500 remote, €36,000 hybrid and €35,000 on-site or unspecified: going to the office barely pays. 52.5% of talent prefers fully remote and 67% of offers allow working from home.

### How do I know if I'm underpaid?

With "Your place in the tech market", the report's free tool. You pick your role, experience, the type of company you work at and your salary, and it tells you what share of profiles like yours earn less than you, what each company type pays and which move would get you closer to the next step. It doesn't store your data.

### If you are a woman working in tech, what should you know?

That the gap isn't there at the start: with under two years, women and men ask for the same. It opens with experience, up to 13% less with ten years or more, and in backend women's median ask is 19% lower. Today women are 15.6% of the Manfred community's tech talent, and only Product Design reaches parity. The good news: more and more companies ask us for hiring processes with gender-balanced shortlists, and the EU pay transparency directive will make them explain their pay differences. Bet on it: ask for what a male colleague with your experience would ask for. The range is there to be negotiated.

## How it's made

### Two sources, one question

Talent: the Manfred community, 127,665 professionals, with data exported on 28 Oct 2026. Percentages are calculated on profiles in Spain with at least one tech role, and each profile is weighted by how recent its data is: one not updated for years counts less, and its experience is not recalculated. Before calculating, we remove gross errors (salaries impossible for the stated experience or off the scale). Talent salaries are what talent asks for, the salary it would move for, in bands; the median is estimated within each band and reweighted by region to match the spread of offers. Current salary is only used for the headline figure in the pay chapter. Experience in four non-overlapping brackets: under 2 years, 2 to 4, 5 to 9 and 10 or more.

Territory: each profile counts in the region and city where it says it lives (Manfred hides the city when fewer than 10 people live there) and splits its weight across its tech roles. Because the Manfred community isn't spread like tech employment, we calibrate each region's weight against INE employment in IT activities (Labour Force Survey): Madrid 28%, Catalonia 23%, Andalusia 12.7%, Valencia region 9.6% and the rest, 26.7%. What talent asks for and how it wants to work are calculated within each area, uncalibrated, and we only publish areas with 30 or more profiles. Barcelona includes L'Hospitalet de Llobregat.

Offers: 167,645 tech job offers published in Spain across 17 roles plus generic offers, TheirStack data as of 12 Oct 2026. "Last 12 months" runs from 1 Oct 2025 to 30 Sep 2026; in yearly series, each year covers January to September so they can be compared. Salaries by role, level, size and region, English, work model, technologies, benefits and the international comparison were computed on the 102,792 development offers from the same 12 months, and we keep them as a reference for the whole.

### What counts as a tech offer

Offers whose title includes developer, programmer, software engineer, frontend, backend, full stack, DevOps, data engineer or data scientist. Role charts, seniority and the 167,645-offer base add QA, security, mobile and AI, and remove non-technical roles with "developer" in the title (Business Developer). Offers with generic titles are assigned a role based on their description. Engineering Manager, Software Architect, Product Manager, Tech Lead, Product Design / UX, SysAdmin and Data Analyst use a broader filter (October 2025 to September 2026). Duplicates: same company with the same title within 30 days counts once; 1.7% slip through. Junior: titles with junior, trainee, intern or graduate, once per company and month. Language: detected from the most frequent words in the description; "asks for English" if it mentions it, and C1 if C1, C2, native or bilingual appear. Benefits: expression search in the description. Closings: closing date on LinkedIn and Indeed, tracked since summer 2026. Salary: midpoint of the yearly range, between €12,000 and €300,000, with medians from 30 offers with salary. Exception: junior medians by role (16 to 46 offers), flagged as indicative. Junior, language and benefits: offers from October 2025 to September 2026.

Yearly series: total volume depends on the sources tracked at each point (LinkedIn joins in September 2023, Jooble leaves that same month, Tecnoempleo is complete from 2025 and company websites grow in 2026). So we don't compare volumes across years: we compare shares within each year. In the only source collected without gaps since 2023, InfoJobs, development offers fall between 2023 and 2025.

### How to cite this report

Manfred × TheirStack. Tech Career Report 2026. November 2026. https://www.getmanfred.com/tech-career-report. Talent: Manfred community, 127,665 professionals (exported 28 Oct 2026). Offers: 167,645 tech job offers published in Spain between 1 Oct 2025 and 30 Sep 2026.

Licence: the data and charts in this report are published under CC BY-NC 4.0. You can use, adapt and share them for non-commercial purposes, always crediting the source: "Tech Career Report 2026, Manfred × TheirStack", with a link to the report. For any commercial use, get in touch first.

### Glossary

- **Median:** The middle value: half ask for (or offer) more and half less. We use it instead of the average because a few very high salaries don't move it.

- **Percentile:** Your position within a group. Being at the 70th percentile means your salary is above what 70% of profiles with your role and experience ask for.

- **Salary range:** The minimum and maximum published in an offer. Every Manfred offer has one; in the market, one in ten.

- **Tension index:** Share of offers asking for something divided by the share of talent that has it. Above 1×, demand exceeds supply; it shows direction and order, not an exact ratio.

- **Tier (company type):** Six groups of companies by how much they pay, from small firms and national consultancies to big tech. At Manfred we call it the hexamodal distribution of salaries.

- **Generic offer:** An offer whose title names no specific role ("developer", "programmer"). In the supply and demand table we assign it by what its description asks for; if it mentions no stack, it stays as "generic, no stack".

- **Junior:** Offers with junior, trainee, intern or graduate in the title, counting each role once per company and month. In talent, under two years of experience.

- **On-site or unstated:** Offers that require the office or don't state a work model. We group them because only about 2 in 100 explicitly say on-site; the rest don't state a work model.

- **Talent:** Professionals in the Manfred community (127,665). Percentages are calculated on profiles in Spain with a tech role, weighted by how recent their data is.

- **Offers:** Tech job offers published in Spain and collected by TheirStack from job boards, company sites and aggregators, deduplicated (1.7% slip through).

### The AI survey

Survey “AI in development 2026”: form sent to the Manfred community, with 756 responses between 7 September and 18 October 2026. It is a self-selected sample: 56% have over ten years of experience and 98% already use AI, so it doesn’t represent the whole market. Company data (training, freedom and licences) is calculated over the 722 people who are working. For multiple-answer questions, percentages add up to more than 100%. Experience of over 20 and over 30 years is grouped together. Only aggregated data is published; the email addresses some respondents left are not used.

### Technical interviews with AI

“The technical interview in the age of AI”: the 84% and the 52% / 32% / 16% breakdown come from the “AI in development 2026” survey (see “The AI survey”). 46% of the companies (6 of 13) have redesigned their technical test and 67% of those (4 of 6) have dropped the take-home or the algorithmic exercise; both figures come from Manfred's conversations with companies in the sector, a small, non-representative sample. The named examples (Meta, Canva, Shopify) are public and are not Manfred data. Company policies change and can vary by round; the formats are described in general terms. Sources consulted in November 2026: Canva engineering blog, Business Insider (Meta) and Hello Interview (Shopify and others).

### Pay transparency directive

EU pay transparency directive: it requires employers to state the salary or range before the interview. The deadline to bring it into Spanish law passed on 7 June 2026.

## Sources and licence

Talent: Manfred. Offers: TheirStack. AI survey: Manfred community, 7 Sep to 18 Oct 2026. Please cite as "Tech Career Report 2026, Manfred × TheirStack" with a link to https://www.getmanfred.com/tech-career-report
