The Market Has Already Voted on AI Productivity Tools
The numbers are striking. AI applications are collectively generating an estimated $150–200 billion in annual recurring revenue, according to Northzone analysis. Yet Sanjot Malhi, Partner and Head of the Growth Fund at Northzone, argues that the largest share of that capital is sitting on a foundation of sand.
'The world does not need any more AI productivity tools,' Malhi wrote in a commentary published Thursday. His firm evaluated tools 'running well into triple digits in the past 12 months alone' and concluded that 'the vast majority of those are destined for the graveyard.'
The indictment is precise. Tools that search, summarize, automate and take notes deliver what they promise — but they carry the weakest economic moats in the AI ecosystem. Open-source models threaten AI applications from below. The foundation models themselves threaten to absorb the use cases from above. The plain-vanilla productivity layer, Malhi writes, is caught in the middle.
$1 Trillion in New Revenue, Uncertain Longevity
Northzone's analysis estimates that roughly $1 trillion in net new AI ecosystem revenue has been added since ChatGPT launched in November 2022 — a rate Malhi calls unprecedented. The problem is quality, not quantity. 'The quality of that revenue is amongst the riskiest it's ever been,' he writes.
The coding vertical illustrates both the ceiling and the path forward. AI Coding accounts for 20–30% of total AI application ARR — the largest single vertical — and it did not stay a productivity tool. It evolved from GitHub Copilot into action systems like Cursor and Claude Code, and then into fully autonomous platforms such as Blitzy and Factory, which can ingest hundreds of millions of lines of code and deliver solutions over weeks of independent work. Malhi notes that 'very early signs of recursive superintelligence are already appearing' in that vertical.
His thesis: every vertical will follow the same arc. AI doctors and lawyers will eventually 'deliver autonomous value superior to any single human being.' The winners will either be companies that do not yet exist, or the rare productivity tools that convert their proprietary data sets and embedded workflows into genuine systems of action before the window closes.
Where Northzone Is Deploying Capital Now
Northzone sat on what Malhi calls a 'then-contrarian thesis' for more than a year, deliberately not deploying capital into productivity tools even as competitors poured money in. 'The technology just didn't exist,' he writes.
Since the beginning of 2026, the firm has 'actively led rounds in excess of several hundreds of millions of dollars' as that technology convergence arrived. Current portfolio bets include Tandem Health, cited as an example of the evolution from productivity tool to system of action, and XBOW and Blitzy, described as 'true autonomous systems of work, from day one.'
Looking ahead, Malhi flags three categories as the next frontier: AI for science (drug and materials discovery), autonomous AI for defense, and physical AI — which he suggests 'might be larger than all of Digital AI put together.'
---
The free-market lesson here is straightforward: capital is self-correcting when investors do their homework. Northzone's framework — moat durability, revenue quality, trajectory from tool to autonomous system — is exactly the kind of disciplined underwriting that separates durable value creation from venture-funded noise. The firms that chased AI Notetaker #25 will learn what every bubble teaches: novelty is not a moat, and the market prices that distinction eventually. The next cycle rewards builders who solve hard problems, not those who wrap a large language model in a tidy UI and call it a business.
