Berlin is Germany's most progressive technology city by the measures talent cares about: distributed hiring, mission-driven equity structures, a gender pay gap of approximately 11 percent, seven points below the German national figure and six below the EU median. Yet when ENTRA's Pay Transparency Audit reviewed 47 Berlin AI startups in Q2 2026, only 25 percent posted salary bands on open roles. Germany's major corporates disclosed at 58 percent over the same period. The gap is 33 percentage points, running in the direction no one expects.
The paradox is behavioral, not legal. Berlin's startup corridor (DeepL, Zalando AI, N26, Trade Republic, Contentful, Aleph Alpha's Berlin team) operates under the same statutory framework as Siemens or SAP. The Entgelttransparenzgesetz, in force since January 2018, grants any employee at a 200-or-more-employee employer the right to request individual pay comparisons by gender. EntgTranspG does not mandate job-posting disclosure; that obligation arrives with Germany's transposition of EU Directive 2023/970/EU, which the BMFSFJ's Referentenentwurf of January 2026 projects for early 2027 after Germany missed the June 7, 2026 deadline. The right of request applies equally today to a 200-person AI startup and a DAX 40 employer with 80,000 staff. Salary transparency month closes with a specifically uncomfortable reading for Berlin's founders: the city that most loudly claims progressive employment values is the city whose startup sector most consistently withholds the information that makes those values testable.
Why startups lag corporates
Three dynamics explain the gap. The first is HR bandwidth: a 60-person startup managing a Series A hiring surge has no compensation benchmarking function. Enterprises like SAP run structured review cycles with external vendors; Berlin startups default to Gehalt nach Vereinbarung (pay by negotiation) because it transfers the anchoring problem to the candidate.
The second is competitive fear. Founders in ENTRA's Q2 2026 survey cited "competitor poaching" as the primary reason for withholding bands. That logic has a limited lifespan: the EU Directive's 100-plus-employee threshold will mandate disclosure by statute in 2027. The poaching argument becomes moot the moment the band is published by law.
The third is founder-era opacity. Many Berlin AI startup leaders built founding teams under informal, equity-heavy bilateral arrangements. Compensation spreads between a founding engineer and the most recent senior hire can reach €40,000 in base for the same functional role. Disclosing that spread requires either compressing it or defending it. Neither is comfortable. The EU Directive's mandate will force the conversation regardless.
Who is leading
Zalando AI leads the Berlin cohort at 78 percent disclosure, the highest rate among the AI-focused unicorns and growth-stage employers in ENTRA's audit. Senior Applied Scientist roles carried posted bands of €110,000 to €160,000 base (~$120K to $175K equiv at EUR/USD 1.09) in Q2 2026. Zalando's transparency posture predates the Directive negotiation: the company has maintained public band ranges on German roles since 2023, framing disclosure as a candidate-experience investment. Candidates enter first-round conversations with ceiling visibility; negotiation concentrates on scope and equity, not the base number.
DeepL posts bands on 65 percent of its Berlin AI roles. The Berlin and Cologne-based translation AI company, approximately 1,200 employees globally with a Berlin AI team of 200-plus, discloses Senior ML Engineer bands at €95,000 to €130,000 base (~$104K to $142K equiv), per ENTRA Q2 2026 job board monitoring. Research-adjacent roles carry lower disclosure rates than product-facing engineering roles, a pattern consistent with the research-prestige opacity visible further down the cohort.
Contentful, the Berlin-headquartered content infrastructure platform, posts bands on 55 percent of senior ML and AI engineering roles, at €95,000 to €140,000 base (~$104K to $153K equiv) for Senior ML Engineers, per ENTRA Q2 2026 monitoring.
The laggards
Aleph Alpha's Berlin team presents the clearest version of the research-prestige opacity pattern. The Heidelberg-headquartered sovereign AI lab maintains an active Berlin engineering satellite but posts bands on approximately 30 percent of open roles, the lowest rate in the ENTRA Q2 2026 Berlin sample. Research Scientist bands are estimated at €100,000 to €145,000 base (~$109K to $158K equiv), derived from ENTRA's triangulation of posted data and recruiter conversations rather than from consistently disclosed postings. The posture is structurally similar to Google DeepMind's London disclosure pattern: research organisations whose compensation advantage is partly informational withhold bands as a negotiation lever, banking on mission and prestige. "Wir kommunizieren Compensation im Kontext der Mission, nicht als abstrakten Benchmark," said one person familiar with Aleph Alpha's Berlin recruiting in a June 2026 ENTRA conversation. ("We communicate compensation in the context of the mission, not as an abstract benchmark.") That posture holds until the law requires the abstract benchmark.
N26 disclosed bands on 45 percent of ML Engineer roles in the Q2 2026 audit, at €80,000 to €110,000 base (~$87K to $120K equiv). Trade Republic carries estimated bands of €90,000 to €125,000 for comparable roles, with ENTRA finding band information in fewer than a third of active Q2 2026 postings.
The 2027 inflection
The BMFSFJ Referentenentwurf, published January 2026 after Germany missed the June 7 EU Directive deadline, is expected to become law in early 2027. When it does, three obligations apply to Berlin's startups: mandatory salary band disclosure in job postings for employers of 100 or more; individual pay information rights on request at the same threshold; and gender pay gap reporting for employers of 50 or more. The 100-employee marker is the critical line: Zalando AI, DeepL's Berlin team, N26, Trade Republic, and Contentful all sit at or above it. Aleph Alpha's Berlin satellite is counted toward the Heidelberg parent's headcount, not independently.
The EntgTranspG's request right already applies today to any Berlin startup with 200-plus employees. It is rarely exercised, in part because bilateral negotiation culture suppresses awareness and in part because the reputational dynamics of invoking the right at a 150-person company differ sharply from those at a 5,000-person enterprise. The 2027 law flips the architecture: disclosure becomes proactive at the posting stage rather than reactive at the employee's request. Founders who have deferred the compensation structure conversation will face a compressed implementation timeline compared to corporate peers who have been building disclosure infrastructure since 2023.
Forecast
ENTRA projects Berlin AI startup disclosure rates reaching 45 to 50 percent by Q1 2027, driven by BMFSFJ law passage and a cohort of startups crossing the 100-employee threshold in H2 2026. The compounding talent advantage for first movers is already visible in Q2 2026 application data: Berlin AI roles posted with salary bands generated 23 percent more qualified-applicant volume per role than undisclosed-band equivalents in the same employer tier. The 4,200-plus active AI roles tracked in ENTRA's Berlin Job Signal Index in Q2 2026 are increasingly filled by candidates who arrive with band expectations normalised from markets where disclosure is already standard. The competitive pressure runs faster than the legal clock.
ENTRA Pay Transparency Audit Q2 2026, Berlin-specific sample (n=47 AI startups), conducted May to June 2026 across LinkedIn Germany, Stepstone, and direct employer career pages; methodology available on request. German large-employer disclosure rate (58 percent) from ENTRA Germany industrial AI analysis published August 7, 2026. DeepL global headcount (approximately 1,200 employees) and Berlin AI team estimate (200-plus) per ENTRA LinkedIn headcount analysis, Q2 2026; not confirmed by DeepL. Salary band figures for all named employers are ENTRA Q2 2026 estimates derived from job board monitoring and recruiter-side conversations where noted; not confirmed by the respective employers. Berlin tech gender pay gap (approximately 11 percent) per ENTRA analysis of Berlin Senate Department for Economics, Energy and Public Enterprises wage data, 2025 reference year; German national figure (approximately 18 percent) per Statistisches Bundesamt, 2025; EU median (approximately 17 percent) per Eurostat, 2025. Berlin 2BR Prenzlauer Berg rental range (€1,650 to €2,100 per month) per Immoscout24 Q2 2026 market index. Berlin active AI roles (4,200-plus) per ENTRA Job Signal Index, Q2 2026. Germany EU Directive transposition timeline per BMFSFJ Referentenentwurf, January 2026, and European Commission DG EMPL transposition tracker, August 2026; "early 2027" projection is ENTRA assessment, not a confirmed government timeline. EntgTranspG (Entgelttransparenzgesetz) in force since January 6, 2018, BGBl. I S. 2152; 200-employee threshold per Section 12(1). EU Directive 2023/970/EU, May 10, 2023, OJ L 132/21; transposition deadline June 7, 2026; 100-employee threshold per Article 5. ENTRA Q2 2026 application volume data based on anonymised recruiter-reported inbound tracking across eight Berlin AI employers. All EUR/USD conversions at EUR/USD 1.09, canonical rate for this publication. Aleph Alpha quote translated by ENTRA from the German original.
For the Germany industrial AI pay transparency analysis published August 7, see Germany Industrial AI: Siemens, SAP, BMW, Bosch Pay Transparency Audit. For the EU Directive transposition picture across member states, see EU Pay Transparency Directive: August 2026 Enforcement Snapshot. For Berlin's graduate-stage compensation and Aleph Alpha's sovereign AI recruiting thesis, see Berlin's AI Talent Surge: Why Europe's Best CS Grads Are Staying Home.
Global Career PlatformFind AI talent. Find your next role.
Booking is hotels. · Airbnb is apartments. · ENTRA is global careers.