TickerTain — Make Analysis Fun Again

▶ 0:26 · Open Source AI · Enterprises shift 40-70% of AI workloads to open-source models to cap foundation model spend

now playing · Open Source AI

Open Source AItailwindscore 8/10aaron holmes

Enterprises shift 40-70% of AI workloads to open-source models to cap foundation model spend

AT&T processes 45B tokens/day internally and uses open-source models (Nvidia NeMo, Meta Llama, Google Gemma) for ~40% of workloads, targeting 60-70%, to keep OpenAI/Anthropic spend flat des…

Open Source AItailwindscore 8/10aaron holmes

Enterprises shift to open-source models to cut AI costs 40-60%

Large enterprises like AT&T are systematically replacing 40-70% of frontier model API calls with self-hosted open-source models (Llama, Gemma, NeMo) to control costs, with the performance g…

Enterprise AI Adoptiontailwindscore 7/10aaron holmes

Model routing architectures emerge as standard for cost-performance optimization

Enterprises are building internal routers that automatically direct tasks to open-source or frontier models based on complexity (e.g., code generation → Anthropic/OpenAI; summarization → op…

AI Infrastructuretailwindscore 7/10aaron holmes

Self-hosted models on own data centers make inference 'trivial cost'

Running open-source models on existing enterprise data centers eliminates per-token API fees, making marginal inference cost near-zero after fixed infrastructure investment. However, hidden…

AI Geopolitics & Export Controlsmixedscore 6/10aaron holmes

US enterprises evaluate but reject Chinese open-source models (DeepSeek, Kimi) over political risk

AT&T analyzed DeepSeek and Moonshot's Kimi models — which are technically competitive — but declined deployment due to White House signaling on scrutiny of Chinese models. Political risk cr…

AI Regulation & Policyheadwindscore 6/10aaron holmes

US enterprises avoid Chinese models despite performance edge

AT&T and other large US companies are evaluating but not deploying Chinese open-source models (DeepSeek, Kimi) due to political backlash risk and White House signals of increased scrutiny,…

AI Economics & Business Modelsheadwindscore 7/10aaron holmes

OpenAI and Anthropic launch privacy features in direct response to open-source competitive threat

Both OpenAI and Anthropic announced enterprise privacy/confidential computing features within days of each other, a reactive move to counter the growing enterprise pitch that self-managed o…

←

▶ 0:26 · Open Source AI · Enterprises shift 40-70% of AI workloads to open-source models to cap foundation model spend

now playing · Open Source AI

Open Source AItailwindscore 8/10aaron holmes

Enterprises shift 40-70% of AI workloads to open-source models to cap foundation model spend

AT&T processes 45B tokens/day internally and uses open-source models (Nvidia NeMo, Meta Llama, Google Gemma) for ~40% of workloads, targeting 60-70%, to keep OpenAI/Anthropic spend flat des…

Open Source AItailwindscore 8/10aaron holmes

Enterprises shift to open-source models to cut AI costs 40-60%

Large enterprises like AT&T are systematically replacing 40-70% of frontier model API calls with self-hosted open-source models (Llama, Gemma, NeMo) to control costs, with the performance g…

Enterprise AI Adoptiontailwindscore 7/10aaron holmes

Model routing architectures emerge as standard for cost-performance optimization

Enterprises are building internal routers that automatically direct tasks to open-source or frontier models based on complexity (e.g., code generation → Anthropic/OpenAI; summarization → op…

AI Infrastructuretailwindscore 7/10aaron holmes

Self-hosted models on own data centers make inference 'trivial cost'

Running open-source models on existing enterprise data centers eliminates per-token API fees, making marginal inference cost near-zero after fixed infrastructure investment. However, hidden…

AI Geopolitics & Export Controlsmixedscore 6/10aaron holmes

US enterprises evaluate but reject Chinese open-source models (DeepSeek, Kimi) over political risk

AT&T analyzed DeepSeek and Moonshot's Kimi models — which are technically competitive — but declined deployment due to White House signaling on scrutiny of Chinese models. Political risk cr…

AI Regulation & Policyheadwindscore 6/10aaron holmes

US enterprises avoid Chinese models despite performance edge

AT&T and other large US companies are evaluating but not deploying Chinese open-source models (DeepSeek, Kimi) due to political backlash risk and White House signals of increased scrutiny,…

AI Economics & Business Modelsheadwindscore 7/10aaron holmes

OpenAI and Anthropic launch privacy features in direct response to open-source competitive threat

Both OpenAI and Anthropic announced enterprise privacy/confidential computing features within days of each other, a reactive move to counter the growing enterprise pitch that self-managed o…