AT&T adopts open AI models to cut costs, improve security
AT&T says it is moving more of its AI workloads to open models to reduce costs and improve security. According to Chief Data and AI Officer Andy Markus, open models are used in about a quarter of AI today, with a target of 70% to 80%. The company reports expense reductions of up to 90% in some cases and says the approach supports 1,000+ internal AI uses.
How this was made
The 30-second read
Why it matters
If the open-model shift is executed broadly, it could reduce AI-related operating costs and improve data control, but the article does not provide consolidated financial impact or formal guidance.
Market read
Traders may view this as incremental positive for cost discipline and security, but without quantified consolidated effects it is unlikely to drive a major repricing.
What to watch
The article does not address model performance, compliance outcomes, or whether savings are sustainable across all AI workloads, which could temper the market’s extrapolation.
Background
AT&T is increasing the share of its AI workloads using open models and running them in its own data centers rather than relying on external cloud services.
Ticker impact
AT&T says it is shifting AI work to open models, targeting 70% to 80% usage and cutting AI expenses by up to 90% in some cases.
Near-term impact likely limited unless traders extrapolate meaningful, recurring opex savings; otherwise it reads as an operational modernization story.
The piece includes specific internal targets (70% to 80% open-model usage) and a large internal expense reduction (up to 90% in some cases), but lacks revenue, capex, or consolidated cost guidance, limiting direct valuation impact.
Market effects
Highlights telecoms’ push toward running AI in-house and using open models, which may influence peers’ AI build-vs-buy and cloud spend narratives.
No specific regional market linkage beyond general telecom operations.
Limited; the story is operational and company-specific with no cross-border regulatory or procurement event.
Counterpoint
Internal expense cuts and open-model adoption may not translate into consolidated savings if integration, governance, or performance costs rise.
Key entities
- companyAT&T
Telecom operator shifting AI workloads to open models to cut costs and improve security, targeting 70% to 80% open-model usage.
- executiveAndy Markus
Chief Data and AI Officer at AT&T, quoted on expected open-model adoption and cost/security rationale.
