Why AI Will Keep Disrupting Industries Through 2026
The AI wave isn’t a flash‑in‑the‑pan; it’s reshaping how enterprises create value, cut waste, and engage customers. Leaders from the International Olympic Committee (IOC), Moderna, and Sportradar recently shared their roadmaps at Reuters NEXT, highlighting three forces that will dominate the next three years: strategic choice between in‑house and third‑party models, the rise of low‑cost “small” models, and a cultural shift that puts innovation ahead of pure cost‑saving.
In‑House vs. Third‑Party: When to Build, When to Buy
Sportradar’s executive vice‑president Nicolo D’Ercole explained that the company leans on “foundation models” such as Gemini, OpenAI, and Anthropic for generic tasks—search, translation, or chat—while reserving custom model development for sport‑specific analytics. Sportradar’s proprietary models can predict the next play in a basketball game, a capability no off‑the‑shelf LLM can match.
Moderna’s Tracey Franklin echoed this split. The biotech giant partnered with OpenAI to democratize AI across its workforce, yet it still evaluates SAP and Workday roadmaps to decide which processes deserve a custom AI engine. “If it’s unique to our pipeline, we build it; otherwise we buy,” she said.
The Emergence of “Small” Models for Everyday Tasks
While giants race to push the limits of massive models, the next wave will be powered by lightweight alternatives like Gemini Flash, GPT‑Nano, and Claude 2. These models run on cheaper hardware, consume less energy, and are “good enough” for routine automation, chat support, and real‑time analytics.
Corna of the IOC pointed out that open‑source models are gaining traction for analytics on video streams—an area traditionally dominated by expensive proprietary AI. Smaller models enable the IOC to process thousands of event recordings without inflating the budget.
Cultural Foundations: Embedding AI Into the DNA of an Organization
Franklin warned that “cost‑first” arguments fall flat with employees. Instead, fostering a culture where AI is viewed as a tool for personal growth and innovation drives adoption. Moderna’s “AI for All” rollout gave every employee access to generative tools, sparking internal hackathons and cross‑departmental breakthroughs.
Similarly, Sportradar incentivizes its developers to experiment with AI‑powered features in sandbox environments, turning “play‑testing” into a formal part of the product lifecycle.
Real‑World Case Studies: Successes & Lessons Learned
Sports Analytics: Predictive Play Modeling at Sportradar
Using a custom model trained on 10 years of NBA play‑by‑play data, Sportradar reduced the latency of live‑prediction services from 2 seconds to 0.4 seconds, boosting partner satisfaction scores by 15% (source).
Pharma’s R&D Acceleration: Moderna’s AI‑Enhanced Drug Design
Moderna’s AI‑driven sequence analysis cut the average design iteration cycle for mRNA candidates from 45 days to 12 days, helping the company file 4 new IND applications in a single quarter (source).
What the Data Says
- 73% of C‑level executives plan to increase AI spend in 2025 (Gartner, 2024).
- Small model adoption is projected to grow at a CAGR of 42% through 2028 (IDC, 2024).
- Organizations that embed AI into their culture see a 2.5× higher employee retention rate (Harvard Business Review, 2023).
FAQ – Quick Answers to Common Questions
- When should a company build its own AI model?
- When the use case relies on proprietary data or requires performance that generic LLMs can’t deliver.
- What are “foundation models”?
- Large pre‑trained AI systems (e.g., Gemini, GPT‑4) that can be fine‑tuned for a wide range of tasks.
- Are small models safe for production?
- Yes—when properly validated, they offer comparable accuracy for many routine tasks while reducing cost.
- How can I start an AI‑first culture?
- Begin with accessible tools, celebrate early wins, and integrate AI learning into performance reviews.
Looking Ahead: The 2026 AI Landscape
By 2026, the AI market will be bifurcated: a handful of megamodels dominate high‑complexity domains, while a vibrant ecosystem of lightweight models fuels everyday automation. Companies that master the art of blending in‑house expertise, third‑party foundations, and a culture that prizes innovation over cost will capture the biggest share of AI‑enabled value.
Stay ahead of the AI curve – Subscribe for weekly insights
Have thoughts on AI strategy? Drop us a comment or join the discussion on our Twitter community.
Related reading