Submit a toolJoin free

Today in AI

AI’s expansion is now being shaped as much by finance, governance and security as by model capability. Meta is using outside capital to fund vast compute, while hedge funds are profiting from the wider AI supply chain. Industry workers are also asking governments to prepare tools for pacing frontier development. Meanwhile, a $1 billion agent-security deal and a major voice-AI round show where the next commercial layers are forming.

The latest signal

Top AI News

The five stories with the clearest impact, selected from today and yesterday.

AIInfrastructure

Meta and BlackRock form $14 billion AI data-centre venture

Reuters•28 Jul 2026, 10:39 BST

News Digest

Meta and BlackRock have formed a $14 billion venture to develop a one-gigawatt data-centre campus in El Paso. BlackRock-managed funds will own 80%, while Meta keeps 20% and leases the capacity instead of funding the site directly. The structure moves much of the construction cost and debt outside Meta as investors scrutinise returns from its vast AI spending.

Additional reporting from: Meta, BlackRock

Key takeaways

03
  1. BlackRock-managed funds will own 80% of the venture, while Meta retains 20% and contributes land and construction assets valued at $2.3 billion.
  2. The financing includes $12.5 billion of debt and $4.9 billion in BlackRock cash, letting Meta secure capacity through long-term lease agreements.
  3. The one-gigawatt El Paso campus is expected to begin operating in 2028 and forms part of Meta’s planned $600 billion US AI infrastructure build-out.

Why it matters

The deal shows how the AI infrastructure race is moving beyond technology budgets into project finance. Meta gains substantial compute without carrying the full construction cost, while outside investors and lenders take a larger role in funding the boom. That creates flexibility, but it also ties future capacity to debt and long-term leases.

Read original story
AIFinance

AI trade puts hedge funds on course for another strong year

Reuters•28 Jul 2026, 15:01 BST

News Digest

Hedge funds returned an average 7% in the first half of 2026, comfortably above their 4.1% ten-year average, according to Goldman Sachs data seen by Reuters. Managers benefited by rotating through the AI trade, from chips to power, data centres and memory stocks. Nearly half of surveyed allocators plan to increase hedge-fund exposure during the second half.

Additional reporting from: Goldman Sachs

Key takeaways

03
  1. Average first-half returns reached 7%, beating the ten-year norm of 4.1% and marking a sixth consecutive half-year above the long-term average.
  2. Managers shifted exposure across the AI supply chain, moving from semiconductors into power, data centres and, most recently, memory stocks.
  3. Of 341 allocators overseeing more than $1.5 trillion, nearly half plan to add hedge-fund exposure in the second half and only 3% plan cuts.

Why it matters

AI is influencing capital allocation far beyond technology shares. The strongest funds are rotating through the infrastructure, power and memory layers that support the boom rather than relying on one crowded stock. That broadens the opportunity, but it also spreads the consequences of an AI correction across more sectors and portfolios.

Read original story
AIPolicy

AI workers call for international tools to pace frontier development

Reuters•28 Jul 2026, 23:03 BST

News Digest

More than 1,100 employees and leaders from OpenAI, Anthropic, Google, Meta and other labs have urged the US government to support international tools for pacing frontier AI. The statement warns that automated AI research could accelerate capability gains beyond current oversight. OpenAI said development may eventually move so fast that the world will need to slow it deliberately.

Additional reporting from: Pacing the Frontier, The Verge, TechCrunch

Key takeaways

03
  1. The initiative includes employees from OpenAI, Anthropic, Google and Meta, with support from the nonprofits Guidelight AI Standards and Encode AI.
  2. Signatories want the US to support international monitoring and governance tools that could deliberately slow automated frontier development.
  3. OpenAI responded that future model development may accelerate so sharply that the world will need mechanisms for pacing the rate of advancement.

Why it matters

The policy debate is shifting from whether frontier AI should ever be slowed to what practical pacing mechanisms could look like. Support from researchers across rival laboratories gives the proposal unusual weight. The harder question is whether governments can coordinate internationally without freezing competition or giving established firms more control.

Read original story
AICybersecurity

Cyera agrees $1 billion Oasis deal to secure enterprise AI agents

TechCrunch•29 Jul 2026, 01:09 BST

News Digest

Data-security company Cyera has agreed to acquire Oasis Security for about $1 billion, mostly in cash. Oasis specialises in controlling non-human identities, including AI agents that need permission to access enterprise software. Cyera plans to combine the technology with its data-security platform as businesses deploy more autonomous systems across sensitive corporate environments.

Additional reporting from: Cyera, Oasis Security

Key takeaways

03
  1. Oasis monitors non-human identities and controls which systems AI agents can access, addressing a security layer that traditional tools often miss.
  2. The target has raised about $195 million since 2022, while Cyera has raised roughly $2.3 billion and recently exceeded $150 million in annual revenue.
  3. Cyera plans to combine Oasis with its data-security platform, creating one layer for governing sensitive information and autonomous software access.

Why it matters

AI agents are becoming enterprise users in their own right, with credentials, permissions and access to sensitive systems. This deal values the security layer around those agents at roughly $1 billion. It signals that identity control, monitoring and data access may become some of the most valuable infrastructure in the agent economy.

Read original story
AIBusiness

Fish Audio raises $52 million to scale expressive AI voice models

TechCrunch•28 Jul 2026, 15:00 BST

News Digest

Fish Audio has raised a $52 million seed round to expand its AI voice models for creators and enterprises. The company says more than eight million people use its open-source or hosted tools, generating $21 million in annual recurring revenue. Funding will support advanced speech models, including audio understanding and speech-to-speech systems planned for this year.

Additional reporting from: Fish Audio, Coreline Ventures

Key takeaways

03
  1. Fish Audio reports more than eight million users and $21 million in annual recurring revenue, unusually strong traction for a seed-stage company.
  2. Its platform offers over 15,000 natural-language voice controls; three speech-generation models are open source and the newest is available by API.
  3. The company plans audio-understanding and speech-to-speech models this year, moving beyond generation into fuller conversational AI systems.

Why it matters

Voice AI is moving from novelty demonstrations into customer service, games, avatars and real-time agents. Fish Audio’s funding, revenue and open-source adoption show that challengers can still gain ground in a crowded market. Consent, ownership and fast takedown systems will remain central as voice cloning becomes easier to deploy.

Read original story