There’s a specific kind of meeting that happened a lot in 2023. Smart people gathered around a whiteboard, someone wrote “AI use case” at the top, a pilot was approved, and everyone felt good about the future. Then the pilot finished. A report was written. And… not much changed.
2024 felt different. Not universally, not evenly — but noticeably. The question in the room shifted from should we try this? to why isn’t this in production yet?
That’s a bigger deal than it sounds.
The Threshold Nobody Announced
There wasn’t a press release. No one rang a bell. But somewhere around the middle of this year, enterprise AI crossed a quiet threshold.
McKinsey’s State of AI survey found roughly 71% of organisations using generative AI in at least one business function — a sharp jump from the year before. Gartner confirmed generative AI as the most frequently deployed AI solution in enterprise. The language inside organisations shifted too: from “proof of concept” to “deployment timeline,” from “let’s explore” to “what’s blocking us.”
That said, the honest read of the data has a useful asterisk attached. Gartner also found that only about 48% of AI projects actually make it into production — and the ones that do take, on average, eight months to get there. So “AI went mainstream” and “AI is still harder than it looks” are both true simultaneously. Which is a more interesting observation than either one alone.
The Funding Story Nobody Quite Expected
If enterprise AI’s maturation was the year’s quieter story, the venture capital comeback was the louder one.
Q4 2024 posted a multi-year high in global VC investment — the strongest quarter since the peak of 2022, according to KPMG’s Venture Pulse. The numbers were striking not just for their size but for their shape. Five US-based AI companies — Databricks, OpenAI, xAI, Waymo, and Anthropic — collectively raised over $32 billion in that single quarter alone. That’s not a diverse ecosystem getting funded. That’s a small number of very large bets.
The pattern worth sitting with: deal volume fell to its lowest annual level since 2016, even as total capital reached multi-year highs. More money chasing fewer deals. Megarounds — $100 million and above — accounted for over half of all global VC investment in 2024. The barbell that the capital concentration post from earlier this year flagged has widened further. The middle is thinner.
What Startups Actually Learned
For founders who’ve been around since the capital-abundant years, 2024 probably felt like the final exam of a course nobody signed up for.
The companies that navigated it well share a recognisable profile: real revenue, defensible unit economics, a clear answer to “when does this make money?” They found that discipline — which felt like a constraint in 2021 — turned out to be the thing investors wanted to see most. There’s something almost poetic about that. The era of growth-at-all-costs bequeathed its opposite: a funding environment that rewards the boring virtues.
The exits remained slow. IPO timelines stretched to a median of 7.5 years from first funding — two years longer than in 2022. M&A picked up slightly, but mostly in specific pockets: AI talent acquisitions from Microsoft, Alphabet, and Amazon stood out as a distinct pattern. When traditional acquisitions are complicated by regulatory scrutiny, buying the team and the technology quietly, without buying the whole company, turns out to be an elegant workaround.
The Enterprise Execution Gap
Here’s the tension that will define the next chapter, and it’s worth naming honestly.
Widespread adoption and successful scaling are not the same thing. Gartner estimates only around 9% of enterprises have meaningfully scaled AI beyond experimental stages. The same organisations that enthusiastically deployed AI tools in 2024 are now sitting with a harder question: how do we build operating models, governance structures, and measurement frameworks that make this sustainable?
The AI ethics post from June flagged this gap between policy and practice. It’s still the live conversation. And as agentic AI moves into enterprise operations — which the most recent post explored — the governance question gets more urgent, not less.
2025 will likely be remembered less for what gets launched and more for what gets operationalised. The organisations that figure out the scaling problem — not just the adoption problem — are the ones that will look back on 2024 as the year that mattered.
The year’s most honest summary might be this: AI moved from the slide deck to the systems, but the hard work of making it work at scale is still mostly ahead of us. Which, if you think about it, is exactly where the interesting problems live.
What was the defining AI moment for your organisation in 2024 — the one that made it feel real?
Let’s keep learning — together.
harish.balasubramanian 
Share your thoughts