SNI: WEEK 08

Feb 20
6 min read

Welcome to all the AI news that matters this week - across biopharma, medtech, life sciences, complex manufacturing and insurance.

tl;dr: AI goes live

The week's dominant signal across all four sectors was the same: AI moving from experimental to operational.

In biopharma, Merck's partnership with Mayo Clinic centres on a production data environment - where Merck will run AI models against de-identified clinical records at scale.

The deal signals that a competitive edge in AI-driven drug development may lie in access to high-quality, real-world clinical data, rather than in model sophistication alone.

Meanwhile, a PYMNTS analysis published on 13 February described 'a structural inflection point' as pharma deploys AI across operations - from patient recruitment and safety monitoring to regulatory submissions - rather than confining it to the discovery pipeline.

In medtech, Medtronic's Stealth AXiS clearance showed AI embedded in surgical hardware as part of an end-to-end workflow.

Copan's PhenoMATRIX system gained FDA 510(k) clearance for automated microbiology image assessment. Both are production systems clearing regulatory hurdles.

In manufacturing, Samsung began shipping commercial HBM4 units to customers - the physical infrastructure that AI accelerators depend on. And Intrinsic's CEO told CNBC that AI-powered robots could reshape global manufacturing through adaptable automation that goes beyond fixed-task repetition. A Deloitte analysis published the same day asked whether 2026 will be the year AI moves 'from possibility to production' on factory floors.

In insurance, the evidence that AI has already arrived at production scale was stark. AIG's CEO told analysts that AI outcomes have gone from 'aspirational' to 'beyond expectations', with Lexington Insurance processing over 370,000 submissions through its AI Assist tool. ERGO confirmed the workforce consequences of that productivity, announcing 1,000 job cuts as AI takes over telephony and claims. And Gallagher Re reported that two-thirds of all insurtech funding in 2025 went to AI-centred companies - $3.35bn across 227 deals.

The common thread is measurable output. Not prototypes, not proofs of concept, not conference demos. Production systems, processing real workloads, generating real returns.

Biopharma

The largest deals centre on securing access to the data infrastructure and proprietary biological datasets that make AI systems valuable over time. The FDA and EMA's joint publication of AI guiding principles signal that regulators are trying to keep pace - though whether principles translate into harmonised standards remains to be seen.

Data as the new platform

The Mayo Clinic's deal gives Merck direct access to advanced AI analytics tools and the ability to run AI models inside Mayo's secure computational environment.

The partnership will initially target inflammatory bowel disease, atopic dermatitis and multiple sclerosis. But the structural significance lies in the model: rather than extracting data, Merck brings its AI to where the data lives. Mayo's COO for the Platform, Maneesh Goyal, framed the opportunity: all clinical, omics and waveform data together will provide a view on disease progression, with AI as the accelerant.

This follows a pattern established in January. Chai Discovery's collaboration with Eli Lilly - one of pharma's largest AI software deals - trains a purpose-built AI model exclusively for Lilly, using Lilly's proprietary biologics data. Chai's zero-shot antibody design platform achieved double-digit experimental hit rates, compressing discovery timelines from months to weeks. Lilly representatives entered Chai's offices within days of seeing results from an early collaboration.

Pharma rewires development with AI operations

A PYMNTS analysis described a 'structural inflection point' in pharmaceutical development. AI is expanding beyond the discovery pipeline into operational infrastructure - patient recruitment from fragmented health records, trial site selection, safety signal detection and regulatory submission drafting.

The piece cited Reuters reporting that large pharmaceutical companies are deploying AI tools to speed clinical trials through automated workflows that reduce manual labour while maintaining scientific rigour. This operational shift aligns with new market projections forecasting the global AI-in-clinical-trials market will reach $20.1bn by 2033 at a 28.7% CAGR.

MedTech and digital health

Medtronic Stealth AXiS: AI meets robotic surgery

Medtronic received FDA clearance for the Stealth AXiS system - a next-generation spine surgery platform that integrates AI-based planning, real-time navigation and robotic assistance in a single workflow.

The system connects with Medtronic's AiBLE platform, combining AI, clinical data and services to support personalised planning, surgical execution and post-operative analysis. A key advance is the ability for surgeons to visualise anatomic motion and patient alignment in real time during spine surgery without repeated imaging. The modular design works across hospital operating theatres and ambulatory surgery centres, broadening the potential market for AI-assisted spinal procedures.

Copan PhenoMATRIX gains FDA clearance

Copan Group's PhenoMATRIX - an automated image assessment system for clinical microbiology - has received FDA 510(k) clearance as a Class II medical device. The system uses AI to assess culture plate images and flag results for microbiologist review. The clearance adds another AI-enabled diagnostic tool to the growing list of FDA-authorised devices - the agency cleared a record 258 AI devices in 2025, bringing the cumulative total past 1,300.

Complex manufacturing

Samsung ships the first commercial HBM4

Samsung announced that it has begun mass production and commercial shipment of HBM4 - the high-bandwidth memory that next-generation AI accelerators depend on. The new chips deliver a consistent processing speed of 11.7 Gbps, exceeding the industry standard of 8 Gbps by 46% and offering 2.7x the total bandwidth of HBM3E.

Physical AI moves onto the factory floor

Alphabet's Intrinsic told CNBC that AI-powered robots could reshape global manufacturing by making automation more flexible and easier to deploy.

CEO Wendy Tan White outlined a vision of SaaS-based robotics enabling reshoring and giving smaller manufacturers access to automation previously reserved for large-scale operations.

The same day, Automation World published a Deloitte analysis asking whether 2026 will be the year AI moves from 'possibility to production' on factory floors. Deloitte's survey of 3,200 global business leaders found 58% are already using physical AI to some extent, rising to 80% within two years. The investment pipeline supports this trajectory: a new report projects the global semiconductor equipment market will grow from $166bn in 2025 to $344bn by 2032.

AI-grown diamonds for next-generation chips

In a development bridging materials science and AI, Michigan State University researchers announced they are using AI and microwave plasma technology to grow high-quality lab diamonds for semiconductor manufacturing.

Backed by a $3m NSF grant, the research targets diamond substrates that can withstand the thermal demands of advanced chip architectures. Diamond's thermal conductivity is several times higher than silicon, making it a potential substrate material for chips that generate more heat as transistor densities increase.

Insurance

Gallagher Re: two-thirds of insurtech funding flows to AI

Gallagher Re's Q4 2025 Global InsurTech Report, published on 18 February, confirmed the scale of capital flowing toward AI in insurance. Global insurtech funding rose 19.5% to $5.08bn in 2025 - the first annual increase since 2021.

Two-thirds of that total - $3.35bn across 227 deals - went to AI-centred companies. Q4 alone surged 66.8% to $1.68bn. Re/insurers made 162 private insurtech investments in 2025 - more than any previous year. The industry that AI is reshaping has become its own largest source of insurtech capital.

AIG: AI outcomes go 'beyond expectations'

AIG CEO Peter Zaffino told analysts that the company's AI outcomes have shifted from 'aspirational' to 'beyond expectations'.

The most striking metric: AIG's Lexington Insurance subsidiary has processed over 370,000 submissions using its AI Assist tool - approaching a target originally set for 2030.

Zaffino described 'a massive shift in our ability to process a significant submission flow way beyond our expectations without additional human capital resources'. AIG deployed generative AI across underwriting and claims throughout 2025 and is now developing an orchestration layer to coordinate multiple AI agents across workflows.

ERGO: 1,000 jobs and the AI workforce question

Munich Re's primary insurance unit ERGO announced that it will cut approximately 1,000 positions in Germany - roughly 200 per year through to 2030 - as AI takes over telephony and claims processing.

ERGO employs around 15,000 people in the country. The company has ruled out compulsory redundancies, relying on natural attrition, phased retirement and severance packages. Some 500 employees will retrain through a Reskilling Academy over two years.

The cuts form part of Munich Re's programme to reduce annual costs by €600m by 2030. ERGO joins ING Groep and Allianz in signalling AI-driven headcount reductions across European financial services.

mea Platform raises $50m for production-grade insurance AI

UK-headquartered mea Platform raised $50m from Scottish Equity Partners - its first external funding after four consecutive years of profitable, bootstrapped growth. mea deploys a domain-specific language model and insurance knowledge graph to automate operations for carriers, brokers and managing general agents.

The platform is live in 21 countries and has processed more than $400bn in gross written premium. mea targets the estimated $2 trillion in annual industry operating expenses, claiming its AI can cut operating costs by up to 60%.