As new data center buildouts hit planning walls and AI inference providers hike costs, is the future of AI to roll your own models?

Following the Trump administration's block on Anthropic's Mythos 5 and Fable 5 models, world leaders have raised concerns that without direct access to frontier models, they may need to develop their own national alternatives.

Oregon State University researchers have developed a brain-inspired phototransistor that combines light sensing, memory, and signal processing in one device. The hardware can electronically control how long optical memories persist or fade, potentially improving energy efficiency in future AI vision systems.

Nvidia showcases agentic robots that can teach themselves high-precision and dexterous tasks - like PC building - in the real world.

Qualcomm is in talks to buy RISC-V-based AI accelerator and CPU developer Tenstorrent for $8 billion - $10 billion.

Two jurisdictions in Tennessee just passed a data center moratorium as three more a set to vote on bills that delay these projects. These temporary bans have gained widespread support, especially in rural regions where developers are increasingly looking to for building their massive projects.

Marvell shares its vision for optically connected data centers, connecting devices across hundreds of kilometers, and the company already has hardware to build them.

State attorneys general have opened a broad investigation into OpenAI, subpoenaing documents on ads, user retention, data handling, minors, health data, model behavior, and safety policies.

David Sacks said the US government warned Anthropic that Claude Fable 5 had been jailbroken and that CEO Dario Amodei refused to fix the flaw.

A preprint claims Pearl’s AI mining network consumes 320,000 GPU-equivalents and 112 MW while producing no verified useful AI computation.

A research firm says the number of blocked data centers in the first quarter of 2026 already matches the number of projects stopped in 2025. The opposition also comes from both sides of the aisle, despite President Trump's push for more AI development inside the U.S.

Companies look for cheaper alternatives as token costs for frontier AI models skyrocket, potentially impacting OpenAI and Anthropic's bottom lines. Subscriptions also take a bite out of these startup's profitability, as utilization rates higher than 5.7% could lead to losses.

Anthropic disabled its two most capable AI models, Claude Fable 5 and Claude Mythos 5, for every customer worldwide on Friday.

OpenAI says it has banned two clusters of ChatGPT accounts it believes are operating from China, and that used its models for covert influence campaigns targeting U.S. tech and policy debates.

Meta has finished separating its operations from Manus, the Chinese-founded agentic AI startup it acquired for roughly $2 billion in December.

After first announcing its scarily capable Mythos Preview model back in April, Anthropic is releasing a public version of Mythos, called Fable 5, that it says is "safe for general use."

The company that just a few weeks ago told us that its Mythos model was much too powerful to be released is now saying that we might need to hit the pause button.

A growing number of CEOs expect AI-driven layoffs, but economic data paints a more complex picture as companies cut junior roles before proving AI delivers meaningful productivity gains.

Google's $920-million-a-month deal with SpaceX will let it secure 110,000 Nvidia GPUs starting October 2026. This is the second data center deal that SpaceX has secured in a matter of weeks, especially as it's quickly approaching its IPO on June 12, 2026.

A research group that includes Huawei Technologies says it completed full-parameter post-training of DeepSeek's V4-Pro, a 1.6-trillion-parameter model.