TL;DR
June 17, 2026
Vibe Coding went from a developer experiment to Collins Dictionary's Word of the Year in under twelve months. In this edition of Forward Lens, Saguna Consulting explores what it really means for enterprises- the productivity gains, the governance risks, and why the biggest shift isn't that AI writes code, but that software is becoming conversational.
.png)
This week's tech term explores Vibe Coding: why it went from a niche developer experiment to the Collins English Dictionary Word of the Year in under twelve months, and what it means for enterprises deciding where humans and AI each belong.
Most people think Vibe Coding is about AI writing code.
That's not the shift.
The shift is that software is becoming conversational.
For decades, if you wanted to build software, you had to learn the language of machines. Suddenly, machines are learning the language of humans.
A founder can describe an application in plain English and generate a working prototype in hours. A product manager can build internal tools without touching a development framework. An engineer can move from idea to deployment in a fraction of the time it once took.
The barriers that once separated ideas from execution are starting to disappear. And that's exactly why the conversation has become so heated.
Andrej Karpathy, AI researcher and former head of Tesla AI, coined the term "vibe coding" in early 2025 with a now-famous line: "fully give in to the vibes, embrace exponentials, forget that the code even exists."
The idea was simple: describe what you want in plain language, let the AI generate the code, and redirect your energy toward intent, direction, and outcome rather than syntax. Within weeks, the phrase had spread across developer communities worldwide. By the end of 2025, Collins English Dictionary named it their Word of the Year, not a tech publication, a dictionary. By 2026, the debate about whether it would catch on was already over.
The numbers followed quickly:
· 92% of US developers now use AI coding tools at least weekly
· 60% of all new code written in 2026 is AI-generated
· 87% of Fortune 500 companies have adopted at least one Vibe Coding platform
· In Y Combinator's Winter 2025 batch, 25% of startups had codebases that were 95% or more AI-generated
Earlier, AI made developers faster. Vibe Coding makes the idea-to-product loop near-instant.
The shift is structural. A developer working with a vibe coding tool no longer asks "How do I implement this?" They ask, "What do I want this to accomplish?" The role has moved from keyboard operator to product director, setting goals, defining intent, evaluating output. Teams are getting smaller while their output is increasing. Projects that previously took months are being completed in days or hours.
And it isn't limited to seasoned engineers. Vibe Coding has opened software creation to designers, product managers, and entrepreneurs who previously had no path from idea to functional application.
Tools like Cursor, GitHub Copilot, and Claude Code have professionalised rapidly. Cursor alone reached a $9 billion valuation and crossed $100 million in ARR. Between 2022 and 2025, AI coding tools attracted $9.4 billion in equity funding globally. This isn't experimental tooling anymore — it's infrastructure.
The developer of 2026 is part architect, part reviewer, part AI operator. The expectations for what a single engineer can handle have shifted, and that change isn't likely to reverse.
Enterprise adoption has moved faster than most expected. Vibe Coding platforms saw 340% growth in enterprise use between 2024 and early 2026.
Sundar Pichai confirmed that more than 25% of Google's new code is now AI-generated. Microsoft's Satya Nadella has indicated the same for significant portions of their codebase. Companies like Stripe and Shopify are running it in real production workflow, not pilots.
The fastest adoption, however, isn't happening inside large enterprises. It's happening among founders, startups, and product teams under pressure to move quickly. What once required weeks of development can now be prototyped in hours. Internal tools that previously sat in a backlog can suddenly be built by the teams requesting them.
The bottleneck is no longer can we build this. It's can we build this well enough to trust it.
Here's the uncomfortable reality. Writing code was never the hardest part of software development.
Understanding requirements is hard. Designing systems is hard. Security is hard. Governance is hard. Maintaining software over time is hard.
Vibe Coding accelerates creation. It doesn't eliminate complexity.
The security data is sobering. A recent research by fast company found that almost half of AI-generated code snippets contained security bugs across five major language models. A large-scale scan of 5,600 publicly deployed vibe-coded applications found:
· 2,000 critical vulnerabilities in live production environments
· 400 exposed secrets including API keys and access tokens
· 175 instances of sensitive personal data including medical records and payment information
AI-assisted commits expose secrets at more than double the rate of human-written code-: 3.2% versus 1.5%. And nearly 40% of AI interactions in enterprise settings involve sensitive company data, often flowing through personal accounts entirely outside corporate oversight.
The deeper risk isn't technical. It's behavioural. Developers trust AI-generated code more than manually written code — which reduces scrutiny at exactly the moment it's needed most.
That's why the firms seeing the greatest success aren't replacing engineering discipline. They're applying it differently. The bottleneck is shifting from creation to validation. The question is no longer "Can we build this?" It's becoming "Can we trust what we've built?"
Three forces are converging at the same time.
AI models have matured to the point where generated code is genuinely functional across a wide range of real-world use cases. The tooling ecosystem has professionalised at speed. And the productivity evidence is documented and consistent — teams using AI coding tools complete tasks 51% faster on average, with development timelines that once stretched months now compressing to days.
Forrester's report on Vibe Coding challenges frame the shift plainly: software development is expected to become the number one use case for AI. Vibe Coding will evolve into what they call "vibe engineering" a more structured discipline that pairs the speed of natural language development with the rigour of systems thinking.
The market for AI-assisted software creation is projected to reach $12.3 billion by 2027. The organisations that benefit most from this shift won't be the ones generating the most code. They'll be the ones that combine human judgment, engineering discipline, and AI capability most effectively — and build the governance layer that makes that combination trustworthy at scale.
At Saguna Consulting, we see Vibe Coding as a capability inflection point, one that rewards companies who treat it strategically rather than reactively.
The companies moving ahead won't simply adopt more AI tools. They'll define where AI-generated code is appropriate, build the review and scanning infrastructure to govern it safely, and retrain their engineering culture around intent and architecture rather than implementation alone.
Vibe Coding isn't the end of coding. It's the beginning of a higher-order craft, one where the constraint shifts from what you can build to what you can clearly envision.
We explored this topic in depth in our LinkedIn newsletter. Read the full Forward Lens edition here.
Next week in Forward Lens, we'll explore another emerging technology shaping the future of enterprise transformation.
Forward Lens by Saguna Consulting · forward together.
Follow the conversation on X · Explore more at sagunaconsulting.com/insights
.png)
