AI Optimization

Twitter

Facebook

Copy Link

Key Facts

Key PlayersNVIDIA, Hugging Face, Intel

Key TechniquesPruning, Quantization, Distillation

Top News

Nvidia DLSS 5 Coming to RTX 4000 GPUs After Achieving 5X Performance Gains on RTX 50 Series

Nvidia announced that DLSS 5 will extend to RTX 40 GPUs after initial optimization on the RTX 50 series. The AI-driven neural rendering technology has achieved a 5X performance gain since its March announcement. While no timeline was provided, the company plans to bring official support to RTX 4000 GPUs once RTX 50 series performance is fully tuned.

Technology2 sources

Nvidia DLSS 5 Coming to RTX 4000 GPUs After Achieving 5X Performance Gains on RTX 50 Series

Technology2 sources

DLSS 5 Launches in NBA 2K27 With Heavy Performance Cost, Modders Crack AMD Support

Nvidia's DLSS 5 officially debuts in NBA 2K27 under the name 'DLSS Neural Rendering,' delivering photorealistic AI-enhanced graphics but imposing a 50-73% performance penalty even on high-end RTX 50-series GPUs. Meanwhile, modders have successfully enabled the technology on AMD's RDNA 4 hardware, though performance remains limited.

Technology10 sources

DLSS 5 Launches in NBA 2K27 With Heavy Performance Cost, Modders Crack AMD Support

Technology10 sources

Nvidia proves AI harness, not model, drives agent performance with 100% ARC-AGI-3 score

Nvidia's research reveals that its AVO coding agent achieved a perfect 100% score on the ARC-AGI-3 benchmark using Claude Opus 5, while the model alone scored just 30%. The breakthrough demonstrates that the AI harness—the wrapper around the model—matters far more than the underlying AI for long-horizon tasks.

Technology2 sources

Nvidia proves AI harness, not model, drives agent performance with 100% ARC-AGI-3 score

Technology2 sources

AMD Exceeds AI Energy Efficiency Targets, Achieves 4X Improvement Ahead of 20X Goal by 2030

AMD reports its rack-scale AI systems achieved 4X better energy efficiency than its 2024 baseline, beating its own 3X interim target. The chipmaker's 20x2030 initiative aims to deliver 20X higher AI performance per watt by decade's end, with two 2030 racks matching the compute of 570 MI300X racks while slashing power consumption.

Technology7 sources

AMD Exceeds AI Energy Efficiency Targets, Achieves 4X Improvement Ahead of 20X Goal by 2030

Technology7 sources

Arrive Opens Bengaluru Innovation Center to Advance AI-Driven Urban Mobility Solutions

Arrive has launched a new innovation hub in Bengaluru, India, establishing a 20,000-square-foot facility that unites cross-functional teams in engineering, product development, AI, and customer care. The center will accelerate predictive AI features and engineering velocity to transform urban mobility worldwide.

Business2 sources

Arrive Opens Bengaluru Innovation Center to Advance AI-Driven Urban Mobility Solutions

Business2 sources

Snowflake Introduces Dynamic Model Routing to Cut Enterprise AI Costs by Up to 3x

Snowflake unveiled dynamic model routing within Cortex AI Gateway, enabling enterprises to automatically select the most cost-effective AI model for each task. Internal testing shows up to 3x greater token efficiency while maintaining quality. The capability addresses rising AI costs by routing simple queries to efficient models and complex tasks to frontier models.

Technology3 sources

Snowflake Introduces Dynamic Model Routing to Cut Enterprise AI Costs by Up to 3x

Technology3 sources

Agentic AI Costs Set to Balloon Fivefold by 2028 as Inference Paradox Takes Hold

Gartner predicts AI inference costs per agentic workflow will surge more than fivefold through 2028, driven by increasingly complex token consumption. Despite falling foundation model prices, the cost of implementing agentic AI systems continues to climb as businesses adopt resource-intensive multistep workflows that far exceed simple chatbots in expense.

Business3 sources

Agentic AI Costs Set to Balloon Fivefold by 2028 as Inference Paradox Takes Hold

Business3 sources

Rising Token Costs Force Enterprises to Rethink AI Economics in 2026

Enterprise AI spending has hit a critical inflection point as token costs become a boardroom priority. Organizations are rethinking AI economics, with mentions of token costs in corporate documents nearly tripling since January. Companies now face tough choices about balancing cloud-based AI with local inference to control escalating expenses while maintaining performance and security.

Business9 sources

Rising Token Costs Force Enterprises to Rethink AI Economics in 2026

Business9 sources

Pathway's 150M Model Proves AI Cost-Efficiency Doesn't Require Massive Scale

Pathway released BDH-CQ, a 150-million-parameter model scoring 29.5% on ARC-AGI-1 at just $0.0007 per task—11 times cheaper than GPT 5.6 Luna. The breakthrough stems from Post-Transformer architecture that reasons internally rather than generating token-heavy intermediate text, challenging the assumption that intelligence requires scale.

Technology3 sources

Pathway's 150M Model Proves AI Cost-Efficiency Doesn't Require Massive Scale

Technology3 sources

Cloudflare CFO: Humans Will Be 'Rounding Error' as AI Traffic Surges 1,000x in Five Years

Cloudflare CFO Thomas Seifert predicts machine-generated internet traffic could outnumber human activity by 1,000 times within five years. Machines already dominate internet traffic since May 2026, forcing infrastructure providers to prioritize efficiency and raising questions about how the web will adapt to serve AI-driven bots over people.

Technology2 sources

Cloudflare CFO: Humans Will Be 'Rounding Error' as AI Traffic Surges 1,000x in Five Years

Technology2 sources

Rippling Cuts AI Spending by 63% After Burning Millions on Tokens, Launches Control Tool

HR software provider Rippling was on track to spend 40% of its R&D headcount budget on AI tokens before its CFO sounded the alarm in March. The company has since slashed costs by 63% and built AI Spend Console, a new product that helps enterprises track and control AI spending while maintaining productivity.

Technology2 sources

Rippling Cuts AI Spending by 63% After Burning Millions on Tokens, Launches Control Tool

Technology2 sources

Naïve Secures $28.5M Series A to Let AI Agents Automate Business Creation and Cut Operational Costs

Palo Alto-based Naïve raised $28.5 million in Series A funding led by Nexus Venture Partners to expand its infrastructure for AI agents that automate business setup and operations. With over 30,000 developer customers and 10x revenue growth in six months, the startup is tackling rising inference costs while enabling autonomous businesses from AI automation agencies to rental car companies.

Technology2 sources

Naïve Secures $28.5M Series A to Let AI Agents Automate Business Creation and Cut Operational Costs

Technology2 sources

Uber Declares End of Tokenmaxxing Era as AI Spending Demands Proof of Value

Uber burned through its entire 2026 AI budget by April, prompting CTO Praveen Neppalli Naga to declare the end of the tokenmaxxing era. The rideshare giant is now demanding measurable returns on AI investments, shifting from unlimited experimentation to disciplined spending that prioritizes efficiency over raw consumption.

Business4 sources

Uber Declares End of Tokenmaxxing Era as AI Spending Demands Proof of Value

Business4 sources

Sapiom Raises $35M Series A to Slash AI Agent Costs as Companies Scrutinize Spending

San Francisco startup Sapiom secured $35 million in Series A funding led by Dragonfly to tackle escalating AI agent costs. The platform sits between AI agents and models, routing calls to the cheapest capable option and enforcing budgets before spending occurs. One customer, Polsia, slashed its monthly bill from $1.2 million to $100,000 after implementing Sapiom's cost optimization infrastructure.

Technology2 sources

Sapiom Raises $35M Series A to Slash AI Agent Costs as Companies Scrutinize Spending

Technology2 sources

MacPaw Partners with Liquid AI to Build On-Device AI Stack for Mac Developers

Ukraine-based app developer MacPaw has formed a strategic partnership with Liquid AI to develop on-device inference technology for Mac. The collaboration will power MacPaw's AI assistant Eney with local AI models and eventually make the tech stack available to developers building for its Setapp app store, which serves over 150,000 paying users.

Technology3 sources

MacPaw Partners with Liquid AI to Build On-Device AI Stack for Mac Developers

Technology3 sources

Prev12 3 4…10 Next