Technology thesis · Artificial Intelligence
high conviction matureNatural language processing
NLP as a distinct field has been largely absorbed by large language models; the remaining standalone NLP challenges are in multilingual, low-resource, and domain-specific applications.
Position maintained continuously · last reviewed Apr 22, 2026
The thesis
Core thesis
LLMs have effectively solved most classical NLP tasks (translation, summarisation, sentiment, NER) to human-competitive levels. What remains: low-resource languages, real-time speech processing, domain-specific terminology (medical, legal), and structured information extraction at scale. The standalone NLP market is being absorbed into the LLM platform market.
State of the art (2026)
By mid-2026 there is no longer a distinct NLP field to speak of: classical tasks – translation, summarisation, NER, sentiment, extraction – are handled directly by frontier models (Claude Opus 4.x, GPT-5.x, Gemini 3.x, plus open weights from DeepSeek, Qwen and Llama), with differentiation now on reasoning, tool-use and agentic quality rather than raw price. Input pricing for capable models has fallen into the sub-dollar-per-million-token range, commoditising text processing. The live standalone markets are voice (ElevenLabs raised $500M at an $11B valuation in February 2026, alongside Deepgram, AssemblyAI and Cartesia), translation (DeepL, ~$2B), document AI/IDP, and verticalised agents (Harvey, Sierra, Decagon). Specialty providers face structural compression from frontier labs above and open weights below.
Everything below is live inside CanaryIQ
The full analysis behind the verdict — the structure is real; the content unlocks when you log in.
Signal stack
Evidence stacked leading → lagging
Technology-native KPIs
Metrics that predict trajectory, tracked over time
Landscape map
Who builds what — and who depends on whom
Catalyst calendar
Dated events that will move the position
Technology roadmap
Milestones on the path to maturity
Watchlists
Companies, people and papers — each with a remove-by condition
Decision frameworks
The same call, framed for your desk
Thesis changelog
When our view changed, and why
Change our mind
3 disconfirming conditions
The rest is inside
You've read the verdict. The file is much deeper.
The full signal stack, technology-native KPIs tracked over time, the landscape of who depends on whom, the dated catalyst calendar, decision frameworks for every desk, live watchlists and the changelog of every time our call on Natural language processing has changed — all live inside CanaryIQ.