Dev Barometer Q3 2025: AI-Driven Talent Meets Lagging Companies

BairesDev launched the Dev Barometer, a quarterly survey of the senior software engineers it works with as a nearshore partner. The first edition surveyed 1,129 developers and 50 project managers across 125 projects in 63 countries, most with 8+ years of experience. Developers have adopted AI deeply enough to measure real time savings and are actively reinvesting that time into AI-specific upskilling. The companies employing them mostly remain at early, pilot-stage AI maturity, with no clear strategy to capture the productivity already happening on their teams.

Fichamento

#

• Developers save an average of 7.3 hours per week using AI for coding; 48% say 10-50% of their code is now AI-assisted, with an average AI-assisted share of 23.5%.
• The saved time gets reinvested in upskilling, roughly 4 hours a week on average. The piece profiles Jorge García, a senior Ruby developer who moved into an AI Engineer role, and Ahmed Somarribas, who says AI helped him go from junior to near-senior level in JavaScript within a couple of months.
• 89% of developers report concrete personal benefits from AI (faster coding, accelerated learning, higher productivity); externally, the piece cites the Stack Overflow 2025 survey putting 51% of developers on AI tools daily and 84% using or planning to use them.
• AI and machine learning are the top upskilling priority for 44% of developers over the next quarter, ahead of every other technical area; 65% worry about falling behind on AI skills specifically.
• Learning happens informally more than formally: on-the-job learning (66%), YouTube (58%), paid courses (48%) rank above certifications (15%). Melissa Guzman, a full-stack engineer, describes short, targeted YouTube videos in her native language as especially useful.
• Organizations lag: 64% of projects sit at early AI maturity (average 2.2 of 5), only 3.2% have AI embedded at their core; large enterprises average 2.6 of 5, with 65% of enterprise projects citing a lack of AI/ML specialists as a barrier.
• Developers and project managers diverge on what’s blocking adoption: developers point to legacy infrastructure and restrictive policy, PMs to unclear ROI and limited resources. They agree only on data privacy and security as a shared concern.
• 87% of developers expect generative AI to disrupt business models within a year, ranking AI agents (75%) and private/fine-tuned models (55%) as the next most transformative forces after GenAI itself.
• On remote work: 76% of developers say it enabled their career in tech, citing global job access (59%), work-life balance (48%) and flexible hours (46%); Ronny Cortez, a BairesDev AI tech lead, frames the bigger opportunity as bringing lightweight, open-source AI models to low-resource communities rather than just access to frontier models.