What’s New November 2025
More companies are shifting from experimentation to embedding AI into their business workflows in real ways. According to the latest McKinsey & Company survey, 88% of organizations now report regular AI use in at least one business function (up from 78% the previous year).
That means for Thoth AI, our mission of delivering scalable, reliable fine-tuning and evaluation solutions is getting an even stronger resonance.
Infrastructure And Compute Demands Are Surging
The tech world is seeing a sharp increase in demand for GPU fleets, memory, high-throughput storage, and orchestration systems. Per the latest article on macro-tech trends, with reference to the November edition of the Thoughtworks Ltd. Technology Radar, training large models or running inference at scale is no longer niche.
For Thoth AI, that means our infrastructure partnerships, model-deployment workflow optimization, and cost-efficiency practices are even more critical. We can help clients manage the shift from model research to real-world deployment.
Regulation Is Heating Up
In Europe, the Artificial Intelligence Act is closely watched. The European Commission is considering stretching or easing compliance timelines in reaction to industry and geopolitical pressures.
That means clients working globally need partners who understand regulatory risks and can embed governance, traceability, and compliance into their AI workflows. As Thoth AI, we are right in our core value, “safety and trust”.
Generative, Agentic, And Multimodal AI Are Advancing
Models that can reason, act (rather than just react), and handle text + images + audio are increasingly the norm. The Q3 2025 global report notes that “agentic AI” spend, and deployment are accelerating strongly.
For our work at Thoth AI, this means moving our services beyond “just fine-tune a language model” and offering “model + agentic workflow + monitoring + alignment” packages that deliver real business value.
The Investment Narrative Is Shifting
While investment in AI remains strong, there is a growing debate about whether we are entering a “bubble” given the hype, infrastructure costs, and unclear ROI for many.
For Thoth AI, our message to clients is this is about embedding AI responsibly, aligning with business strategy, preparing for global scale, and managing change. That is what separates long-term winners from short-term wins.
What’s To Come
Embedding AI across every vertical
Over the next 12-24 months, we expect to see more companies in sectors such as healthcare, finance, logistics, and manufacturing shifts from “we have a proof of concept” to “this model powers something real”. If you combine that with what McKinsey found — high-performing organizations redesign workflows and commit budget to AI — then you see a clear alignment with Thoth’s expertise.
We will therefore strengthen our offering in vertical-specific fine-tuning, evaluation frameworks, operational monitoring, and cultural change support.
More rigorous governance and traceability
As regulation catches up, as even internal enterprise risk units demand better audit logs and traceability of AI decisions, the need for embedded “human-in-loop / human-validation” systems will increase.
That means Thoth AI will expand its evaluation frameworks, documentation practices, and tooling for model governance, bias detection, and safe deployment. We’ll also scale our multilingual, cross-cultural evaluation capacity (which aligns with our global perspective value).
Focus on infrastructure, model optimization, and cost efficiency
Given the compute demands and cost pressures, optimizing models for latency, memory footprint, and energy usage will become more central. Also, edge deployments, hybrid cloud-on-prem pipelines, and federated learning will grow.
Thoth AI will continue investing in partnerships and expertise, advising clients on models that scale in their contexts.
From generative to agentic to autonomous workflows
The move is going from “model outputs text” to “model plans, executes, monitors, adapts”. That means we will see more agentic systems, models that handle multi-step workflows rather than single prompts.
For Thoth AI, our offering will evolve from fine-tuning chatbots to fine-tuning agent frameworks, workflow orchestration, multi-modal inputs and outputs, all wrapped in our safe, trusted, expert-led delivery.
Global scale with local nuance
With our global perspective pillar, we see more organizations demanding AI solutions that are not just English-language or US-centric. They want culturally adapted, multilingual models, evaluation across jurisdictions, and compliance with regional regulations. As regulations diverge (Europe, UK, US, Asia) and as AI products globalize, the demand for partners who bring human expertise plus technical muscle becomes crucial.
We will continue to build our multilingual SME network, evaluate across languages and markets, and create frameworks supporting global rollout.
Why This Matters for Thoth AI Clients
- Clients need to move from “cool proof of concept” to “mission-critical system,” and that means aligning model strategy with business strategy, not just building capability.
- The infrastructure curve means decisions about model size, deployment mode, monitoring, and cost matter just as much as “accuracy” or “capability”.
- Governance and safe deployment are no longer optional. They are central to risk management.
- Global scale is now a requirement for many companies; localized nuance wins.
- Hyperscale hype may fade, but value creation remains. The clients who win will embed AI in real workflows, redesign processes, and measure impact.
What We at Thoth AI Recommend Right Now
- Review your AI pilots and ask: “Which of these are ready to scale, which need redesign, which should be sunset?”
- Map your model-deployment path with infrastructure and cost in mind. Ask: What happens when I double users?
- Audit your governance framework: Are my model decisions traceable? Are human validators in place? Am I ready for regulation in Europe, the UK, or other markets?
- Think globally: If you serve multiple markets, how does your model perform in Spanish, Mandarin, and Arabic? Are cultural norms built in?
- Focus on value, not hype: Align your AI objectives with growth, innovation, or operational transformation (not just “we want AI”). According to McKinsey, the highest-performing companies do this.
