12 percent. Statistics Sweden's (SCB) 2025 lönestatistik estimates Sweden's unadjusted gender pay gap in private-sector AI and technology roles at approximately 12 percent. France carries 22 percent; Germany, 18 percent; the Netherlands, 13 percent. Sweden reaches 12 percent not because its AI labour market is structurally different from its European peers but because Swedish employment law has required employers to actively audit and correct pay differentials since 2008, 17 years before any EU Directive compelled the same.
The paradox at the centre of this briefing: Sweden, the EU member state with the most rigorous mandatory pay equity audit framework in force, did not transpose EU Directive 2023/970/EU by the June 7, 2026 legal deadline. It sits alongside France, Germany, the Netherlands, and 19 other non-transposing member states. The reason explains precisely where Sweden's pay transparency lead is real and where it remains structurally incomplete.
Diskrimineringslagen: The Audit Obligation That Precedes the Directive
Sweden's Diskrimineringslagen (Discrimination Act, SFS 2008:567) imposes an annual lönekartläggning on every Swedish employer with 10 or more employees. That threshold is the most consequential number in the European pay transparency landscape: EU Directive 2023/970/EU requires annual pay gap reporting only from employers with 250 or more staff. Sweden's audit obligation applies to firms that fall 240 employees short of that trigger.
The lönekartläggning process goes beyond passive reporting. Employers must categorise all roles into equivalent-work groupings, calculate pay differences between male and female employees within each grouping, and document the result. Where the analysis produces an unexplained differential — one not attributable to objective, gender-neutral criteria — the employer must take corrective action within three years. The process must be conducted with union representatives under Medbestämmandelagen (Co-Determination Act) provisions. At Stockholm's tech sector, where IT&Telekomföretagen collective agreements cover a substantial share of employers, a union counterpart reviews the pay mapping annually alongside management.
The Diskrimineringsombudsmannen (DO), Sweden's Equality Ombudsman, is the enforcement authority. The DO does not receive annual reports from each employer; it enforces on complaint. The lönekartläggning is an internal audit, not a public database. That distinction is the structural gap that Directive 2023/970/EU addresses.
The Transposition Paradox: Leading Without Full Compliance
Of the 27 EU member states, only Slovakia, Italy, Lithuania, and Malta completed Directive 2023/970/EU transposition by the June 7, 2026 legal deadline, per the European Commission DG EMPL transposition tracker, July 2026. Sweden is among the 23 non-transposing states, with the Arbetsmarknadsdepartementet (Ministry of Employment) projecting implementing legislation for 2027.
The gap between Sweden's existing framework and Directive compliance is specific. The Diskrimineringslagen does not require salary ranges to appear in job postings before interviews. Article 5 of Directive 2023/970/EU does. The Diskrimineringslagen does not require the publication of company-level pay gap reports to a competent national authority. Article 9 of the Directive does. Sweden's annual lönekartläggning is more rigorous than anything in the Directive's own text. But it is an internal audit; the Directive's additions are external and candidate-facing.
What Sweden's 2027 transposition will add: salary bands in job advertisements before any candidate interview, pay data available on individual request at companies with more than 50 employees, and annual gap reports filed with the Diskrimineringsombudsmannen for employers with 250 or more staff. For a market already conducting annual pay audits at 10-employee scale, the substantive compliance burden is modest. The addition is public architecture on top of existing institutional discipline.
Klarna and Spotify: Voluntary Disclosure Runs Ahead of Law
Klarna, which returned to net AI hiring in H1 2026 following its 2024 to 2025 automation-driven headcount reset, discloses salary bands in an estimated 88 percent of active ML and AI roles, per ENTRA Q2 2026 monitoring of Klarna's careers portal and LinkedIn postings. Senior ML Engineer roles carry posted ranges of SEK 1,200,000 to SEK 1,700,000 annually, approximately €107K to €152K at SEK/EUR 11.2, or approximately $117K to $166K at EUR/USD 1.09 (Q2 2026 mid-market). Klarna publishes an annual gender pay report for its Swedish entity, citing an unadjusted gap of approximately 9 percent in its 2024 disclosed figures — below the SCB tech sector estimate and the lowest in this article's employer set.
Spotify's Stockholm AI and ML function discloses salary bands in 100 percent of Sweden-based postings, per ENTRA Q2 2026 monitoring. ML Engineer roles run SEK 950,000 to SEK 1,350,000 annually, approximately €85K to €121K (~$92K to $132K equiv); Staff ML Engineer roles carry posted ranges of SEK 1,350,000 to SEK 1,750,000, approximately €121K to €156K (~$132K to $170K equiv). Spotify publishes an annual Equity Report disaggregating gender pay data globally; its Sweden-specific disclosure aligns with Diskrimineringslagen requirements but, under the 2027 transposition, will require formal DO reporting that currently does not exist.
The structural point: both employers are disclosing ranges voluntarily before the Directive requires it. In the markets covered earlier in this August series where voluntary compliance was lowest — Thales at 52 percent in France, BMW at 52 percent in Germany — non-disclosure clustered at employers least exposed to candidate-market pressure. In Stockholm, Klarna and Spotify recruit from a talent pool that treats a posted range as a baseline expectation before the first screening call. The expectation precedes the law.
Ericsson and the Teknikavtalet Layer
Ericsson's Kista AI function, running approximately 350 AI and ML roles in Sweden as of Q2 2026 per ENTRA LinkedIn headcount analysis, operates under the Teknikavtalet: a suite of collective agreements between Teknikföretagen (the Swedish engineering employer association) and Unionen/Sveriges Ingenjörer for white-collar employees, with a separate agreement covering IF Metall for blue-collar workers. The Teknikavtalet does not require salary bands in external job postings. It requires biannual pay reviews and transparency to union representatives on pay-setting criteria, creating a calibration layer that constrains the range of discretionary pay decisions before any candidate negotiation begins.
Ericsson's posted disclosure rate for active Sweden AI roles stands at approximately 91 percent in ENTRA Q2 2026 monitoring. Senior AI Research Engineer roles carry stated bands of SEK 1,350,000 to SEK 1,750,000, approximately €121K to €156K (~$132K to $170K equiv); AI Research Lead roles reach SEK 1,600,000 to SEK 2,100,000, approximately €143K to €188K (~$156K to $205K equiv). Ericsson's general-purpose AI model obligations under Article 53 of Regulation (EU) 2024/1689 have added model documentation and transparency roles to its Kista posting stack since H2 2025, with those roles carrying bands at the upper end of the Senior AI Engineer tier.
WASP and KTH: The Academic Anchors
WASP (Wallenberg AI, Autonomous Systems and Software Program), Sweden's national AI research programme with a SEK 5.8 billion funding base through 2031, anchors the academic salary layer that calibrates Stockholm's PhD-to-industry pipeline. WASP-funded doctoral researchers receive standard Swedish doctoral employee salaries, progressing from approximately SEK 370,000 annually (~€33K; ~$36K) in the first research year to approximately SEK 430,000 (~€38K; ~$42K) by the fourth year, per doctoral salary agreements at KTH, Chalmers, and Stockholm University. WASP postdoctoral researchers earn approximately SEK 490,000 to SEK 560,000 annually (~€44K to €50K; ~$48K to $55K), depending on host institution and research area.
KTH Royal Institute of Technology publishes faculty salary schedules under Sweden's offentlighetsprincipen, the constitutional principle established in Tryckfrihetsförordningen (Chapter 2) that makes public-sector employer pay data accessible to any member of the public on request. KTH Associate Professor (Universitetslektor) salaries in AI-adjacent departments run approximately SEK 700,000 to SEK 850,000 annually (~€63K to €76K; ~$68K to $83K); KTH Professor of Machine Learning roles carry published schedules of approximately SEK 880,000 to SEK 1,150,000 (~€79K to €103K; ~$86K to $112K).
No equivalent transparency exists at any private AI employer in Sweden. For a junior researcher choosing between a WASP doctorate and a direct industry role at Klarna or Ericsson, the salary gap at the early-career band is approximately SEK 800,000 to SEK 1,000,000 annually. WASP industrial partner agreements attempt to offset that gap by embedding doctoral researchers in company settings throughout their programmes.
The Kollektivavtal Advantage: Structural Transparency Without Posting
Sweden's kollektivavtal (collective agreement) system covers approximately 80 percent of Swedish employers by employee count, per Medlingsinstitutet (the National Mediation Office) 2025 data. For Stockholm's AI employers, the operative agreements are Unionen (covering white-collar IT workers at Klarna, Spotify, and King) and Teknikavtalet (covering engineering roles at Ericsson, Saab, and Hitachi Energy). Both require that pay criteria be disclosed to union counterparts and that salary reviews follow documented, negotiated processes.
The transparency advantage this creates is not candidate-facing in the same way as a job posting band. It is institutional: union counterparts at these companies review role-level pay data that candidates never see in any posting. When a candidate negotiates a Stockholm AI offer, they negotiate with an employer whose pay-setting decisions have already been calibrated against collective norms. The floor is higher and less discretionary than in a market with no equivalent collective review mechanism. Where London produces a 52 percent AI posting rate and Munich approximately 60 percent, Stockholm's tech-sector posting rate at its largest employers runs at 88 to 100 percent. The kollektivavtal is part of why employers choose to disclose: the internal discipline is already documented, posting the number adds no meaningful exposure.
What Candidates Negotiating Stockholm AI Roles Should Know
Three things are true simultaneously in Stockholm's AI compensation market in August 2026. First, the posted range is almost always present, and it is backed by a lönekartläggning process that has constrained arbitrary pay-setting at scale for 17 years. Second, the posted range is gross SEK: for international hires eligible for Expertskatt — Sweden's Expert Tax Relief scheme, which exempts 25 percent of income from tax for qualifying foreign specialists in their first five years — the after-tax figure differs from gross-to-gross comparisons. Third, the posted range is a starting point, not a ceiling. Swedish employment law does not prohibit negotiating above a posted band, and the lönekartläggning means employers hold internal data on where equivalent peers sit.
The 2027 Directive transposition will add two tools not currently available to candidates: the right to request data on what comparable peers earn in the same role and the right to a joint pay assessment where a disclosed company-level gap exceeds five percent. Sweden's AI employers are already operating well above the standard the Directive sets. What 2027 adds is not the data — the data exists — but the legal right for candidates to reach it.
Sources and methodology: SCB (Statistics Sweden) gender pay gap data per Lönestatistik 2025 (private sector, information and technology sector); approximately 12 percent estimate for AI and technology roles is ENTRA's estimate based on SCB sector data and ENTRA Q2 2026 job market monitoring; SCB does not publish a standalone "AI sector" gender pay gap sub-segment; France (22%), Germany (18%), Netherlands (13%) gender pay gap figures per EU Commission benchmarking data and national statistics offices as reviewed in this August series (August 7, August 9, August 10, 2026). EU Directive 2023/970/EU of the European Parliament and of the Council of May 10, 2023; OJ L 132, 17.5.2023; member-state transposition deadline June 7, 2026. EU Directive transposition status: four member states completed transposition by June 7, 2026 (Slovakia, Italy, Lithuania, Malta) per European Commission DG EMPL transposition tracker, July 2026; Sweden among the 23 non-transposing member states as of article date. Arbetsmarknadsdepartementet transposition timeline (2027) per ENTRA Europe Bureau monitoring of Swedish government publications; not confirmed as a formal legislative commitment. Diskrimineringslagen (SFS 2008:567): annual lönekartläggning obligation, Chapter 3; 10-employee threshold per 2017 amendment (prop. 2015/16:135); union review obligation under Medbestämmandelagen (SFS 1976:580). Diskrimineringsombudsmannen (DO) enforcement is complaint-driven; DO does not receive annual employer pay reports. Directive 2023/970/EU Article 5: pre-interview salary range requirement; Article 9: annual pay gap reporting; Article 10: right of access to pay data; 250-employee annual reporting threshold and 50-employee individual request threshold per Directive text. Klarna AI posting disclosure rate (approximately 88%) and salary bands (SEK 1,200,000 to SEK 1,700,000, Senior ML Engineer) per ENTRA Q2 2026 monitoring of Klarna careers portal and LinkedIn; not confirmed by Klarna. Klarna gender pay gap (approximately 9%, 2024 Swedish entity) per Klarna published annual reporting; ENTRA has not independently verified the calculation methodology. Spotify 100% AI posting disclosure rate and salary bands (ML Engineer SEK 950,000 to SEK 1,350,000; Staff ML Engineer SEK 1,350,000 to SEK 1,750,000) per ENTRA Q2 2026 monitoring of Spotify careers portal and LinkedIn; not confirmed by Spotify. Spotify Equity Report per published Spotify global equity and inclusion reporting; Sweden-specific gender pay disaggregation is an ENTRA estimate. Ericsson Sweden AI/ML headcount (approximately 350 roles, Q2 2026) per ENTRA LinkedIn Talent Insights analysis; Ericsson does not publish granular Sweden AI headcount separately. Ericsson disclosure rate (approximately 91%) and salary bands (Senior AI Research Engineer SEK 1,350,000 to SEK 1,750,000; AI Research Lead SEK 1,600,000 to SEK 2,100,000) per ENTRA Q2 2026 monitoring of Ericsson Careers; not confirmed by Ericsson. Ericsson GPAI obligations per Regulation (EU) 2024/1689, Article 53. Teknikavtalet is a suite of collective agreements: Teknikföretagen–Unionen/Sveriges Ingenjörer (white-collar) and Teknikföretagen–IF Metall (blue-collar); Unionen coverage at Klarna and Spotify per ENTRA employer research; not publicly confirmed by named employers. Medlingsinstitutet 2025 kollektivavtal coverage (approximately 80% by employee count) per Medlingsinstitutet published data. London AI posting rate (approximately 52%) and Munich (approximately 60%) per ENTRA Q2 2026 monitoring. WASP doctoral salary range (SEK 370,000 to SEK 430,000) and postdoctoral range (SEK 490,000 to SEK 560,000) reflect standard Swedish doctoral employee salary scales at KTH, Chalmers, and Stockholm University per 2025 to 2026 doctoral salary agreements; WASP does not set a separate doctoral salary rate; rates vary by year of programme and host institution. WASP funding base (SEK 5.8 billion through 2031) per WASP published programme information. KTH faculty salary ranges (Universitetslektor SEK 700,000 to SEK 850,000; Professor SEK 880,000 to SEK 1,150,000) reflect ENTRA review of KTH published salary data accessible under offentlighetsprincipen; specific figures are estimates based on published KTH salary schedules and are subject to individual academic merit criteria; Tryckfrihetsförordningen (TF, SFS 1949:105), Chapter 2. WASP-to-industry salary differential (approximately SEK 800,000 to SEK 1,000,000 annually at early-career band) is an ENTRA illustrative estimate; not a precisely calculated figure. Expertskatt scheme per Forskarskattenämnden (Expert Tax Board) published 2026 guidance; see ENTRA Stockholm briefing July 28, 2026 for full effective-rate calculations; this article provides only a summary reference. SEK/EUR conversion at 11.2 (Q2 2026 mid-market approximation); EUR/USD at 1.09 (Q2 2026 mid-market approximation); all conversions are approximations. Salary and disclosure data are ENTRA estimates based on job board monitoring through June 2026; not independently confirmed by named employers. This article does not constitute tax or financial advice.
Global Career PlatformFind AI talent. Find your next role.
Booking is hotels. · Airbnb is apartments. · ENTRA is global careers.