ATT and NVIDIA Embrace the Open-Weight AI Wave Amid Cost and Demand Debates
Yahoo Finance ·
On August 12, 2026, AT&T Chief Data and AI Officer Andy Markus revealed that OpenAI models drive roughly 25% of the telecom's AI usage, with a long-term goal of 70% to 80%. Meanwhile, NVIDIA CEO Jensen Huang has spent the past month championing open-weight AI across a consortium of companies. By shifting to open models, AT&T cut costs by 80% to 90% in certain applications, processing an average of 45 billion AI tokens daily via a smart router. AT&T also developed its custom open model, OTel, delivering a fivefold return on investment this year. Cerebras CEO Andrew Feldman noted that chip stocks should not suffer from open-source releases, citing the Jevons paradox where lower costs drive higher overall consumption. As of Q1 2026, AT&T was held by 72 hedge funds, down from 77, while NVIDIA's backing rose to 275 hedge funds from 264. Both firms remain confident that cheaper AI will ultimately expand total usage rather than reduce infrastructure spending.
AI 시장 분석
AT&T is reducing costs by 80 to 90% in specific applications through the adoption of open-source AI models, processing an average of 45 billion tokens per day. NVIDIA CEO Jensen Huang argues that the open-weight AI trend will lead to expanded demand, sustaining infrastructure investment needs. Conversely, concerns also coexist that if compute demand per query decreases due to improved efficiency, chip demand growth could slow down.
상승 영향
- AI — AT&T cutting costs by 80 to 90% and achieving a 5x ROI through open-source AI adoption accelerates AI implementation speed, driving growth in the related ecosystem.
- Semiconductors — According to Jevons paradox, as AI operational costs become cheaper, overall usage will explode, steadily maintaining demand for chips and infrastructure.
하락 영향
- Semiconductors — If companies switch to cheaper open models requiring fewer computations per query, hardware performance demand may decline, slowing the growth of chip manufacturers like NVDA.
DYAX 전담 분석
AT&T's 80-90% reduction in AI costs and the proliferation of open-weight models increase the cost efficiency of the AI ecosystem while having a complex impact on computing infrastructure demand. According to Jevons paradox, if usage surges, semiconductor demand is sustained, but if compute per query decreases, growth momentum for hardware companies like NVDA could weaken.
The bullish scenario is that total semiconductor infrastructure demand continues to grow due to accelerated AI adoption driven by cost reductions, while the bearish scenario is that maximized efficiency directly translates to lower chip sales. Investors should closely monitor the adoption speed of open-weight models and actual hardware shipment indicators.
AI가 생성한 분석으로 투자 자문이 아닙니다.
DYAX Investor Sentiment
Bullish (Long) 62% · Bearish (Short) 38%
461 participants
Related News
- Faraday Future Outlines EAI Robotics Strategy at 2026 J.P. Morgan Conference
- Why ServiceNow Outperformed the Market on Thursday
- How DELL Stockholders Can Generate 21% Annualized Income
- Nebius Group Shares Surge 40% in Under Two Weeks Amid AI Boom
- Walmart Shares Slide Following Revenue Miss
- Coinbase Slumps 50% and 2X ETF Plummets 85%: The Brutal Math of Recovery
