Jan 28, 2026

AI News Roundup: DeepMind Open-Sources AlphaGenome as Amazon Cuts 16,000 Workers in Bungled Layoff

The contrast couldn’t be starker: on the same day Google DeepMind opens up its Nobel-worthy AI to accelerate cancer research, Amazon bungles the layoff of 16,000 workers by accidentally emailing them condolences before they’d been told they lost their jobs.

Meanwhile, Big Tech’s AI spending binge faces its moment of truth as Microsoft and Meta report earnings, Illinois becomes the first state to mandate AI transparency in hiring, and new research suggests companies aren’t even using the AI capabilities they’re paying billions to build.

Here’s everything that matters from January 28, 2026.

The Big Story: DeepMind Opens AlphaGenome to the World

Google DeepMind has released AlphaGenome’s source code and model weights to the scientific community for noncommercial use, a significant step toward democratizing cutting-edge AI for biological research. The announcement, detailed in a Nature paper published today, comes seven months after the model’s initial launch.

AlphaGenome, the successor to DeepMind’s Nobel Prize-winning AlphaFold, can analyze sequences of up to one million DNA letters and predict how genetic mutations affect gene expression. This capability is particularly valuable for investigating what scientists call “DNA’s dark matter” - the vast stretches of genetic code whose functions remain poorly understood.

The impact is already measurable: nearly 3,000 scientists from 160 countries have been using AlphaGenome to advance research into cancer, neurodegenerative disorders, and infectious diseases. By open-sourcing the model, DeepMind is effectively multiplying that research capacity, allowing labs worldwide to customize and build upon the technology rather than simply accessing it through an API.

This follows the playbook that made AlphaFold transformative for protein science - release the model, let researchers iterate, and watch the breakthroughs compound. For genomics, where understanding mutation effects could unlock treatments for thousands of diseases, the timing couldn’t be more consequential.

Today’s Top Stories

Amazon Bungles 16,000-Worker Layoff

Amazon’s “Project Dawn” layoff of 16,000 workers turned into a PR disaster when AWS prematurely sent a commiseration email hours before employees were supposed to be informed. The message, signed by Colleen Aubrey, SVP of Applied AI Solutions at AWS, stated that impacted employees in the U.S., Canada, and Costa Rica had already been told they lost their jobs - except they hadn’t.

The cuts affect workers across multiple divisions and represent one of the largest single-day layoffs in Amazon’s history. The company has not commented on the premature notification.

Sources: Yahoo Finance, The Street

Microsoft and Meta Earnings Put AI Spending Under the Microscope

Both Microsoft and Meta are reporting earnings after market close today, with Wall Street laser-focused on whether their massive AI infrastructure bets are paying off.

Microsoft is navigating a $100 billion annual capital expenditure run-rate for AI infrastructure. Consensus estimates project revenue of approximately $80.3 billion (15% YoY growth) with EPS between $3.86-$3.97. The company plans to increase AI capacity by more than 80% throughout fiscal year 2026 and roughly double its data center footprint over the next two years.

Meta’s capex trajectory is even steeper: analysts project nearly 57% growth in capital expenditures in 2026 to over $110 billion. Goldman Sachs forecasts capex of $125 billion this year, rising to $144 billion in 2027.

OpenAI Research Reveals “Capability Overhang”

OpenAI has released new research on what it calls the “capability overhang” - a fundamental disconnect between what AI systems can do right now and what companies are actually doing with them.

The research, as characterized by The Decoder, suggests the AI industry has found its 2026 narrative: users, not models, are the bottleneck. If this framing takes hold, it could reshape how enterprise AI investments are evaluated - shifting scrutiny from model capabilities to organizational readiness.

OpenAI also announced new finance leadership, with Ajmere Dale joining as chief accounting officer and Cynthia Gaylor as business finance officer.

Apple Launches Creator Studio Pro

Apple unveiled Creator Studio Pro, positioning AI explicitly as a collaborative assistant rather than a replacement for human creativity. The software automates organizational tasks like finding clips, building slides, and managing production workflows - deliberately steering clear of content generation.

It’s a philosophical stake in the ground: Apple is betting that creators want AI to handle the tedious parts of their work, not the creative parts.

Quick Hits

Illinois AI Disclosure Law: Starting today, Illinois requires companies to disclose when AI influences employment decisions - the first state law mandating transparency for AI-driven hiring practices.
AI Infrastructure Boom: ASML’s record orders for semiconductor manufacturing equipment show the AI infrastructure boom isn’t slowing. As the sole supplier of EUV photolithography equipment for cutting-edge AI chips, ASML is a bellwether for industry demand.
AI Healthcare Market: The U.S. AI in Healthcare market is projected to reach $99.77 billion by 2033, up from $10.26 billion in 2025 (CAGR of 32.88%).
Fed Watches AI: Federal Reserve officials are considering AI-related productivity themes at their January FOMC meeting, including potential labor market and inflation impacts.
Prompt Injection Alert: New research on PromptForest tackles malicious prompt detection, addressing what OWASP identified as the #1 critical vulnerability appearing in 73%+ of production AI deployments.
Google Gemini for JEE: Google is launching JEE Main practice tests within Gemini, targeting students preparing for India’s most competitive college entrance exam.
Defense AI: Odysight.ai (NASDAQ: ODYS) received two pilot orders from a major defense customer for AI-based monitoring on combat helicopters and airborne weapons systems.
AI CEO Statements: Anthropic’s Dario Amodei and OpenAI’s Sam Altman both issued statements addressing ICE enforcement actions in Minneapolis, condemning violence while navigating politically sensitive waters.

What This Means

January 28 captures the growing tension at the heart of the AI industry: the technology is advancing rapidly (DeepMind just handed the world a tool to decode DNA), but the organizational and human elements aren’t keeping pace. OpenAI’s “capability overhang” research suggests that enterprise AI’s biggest problem isn’t the models - it’s that companies don’t know how to use what’s already available. Meanwhile, Amazon’s layoff disaster and the $200+ billion in quarterly AI capex from Microsoft and Meta raise uncomfortable questions about whether Big Tech’s AI investments are building sustainable value or just moving fast and breaking things - including their own workforce. The Illinois disclosure law hints at the regulatory response to come: if AI is making consequential decisions, people deserve to know. As earnings season forces a reckoning with AI spending, the industry’s narrative is shifting from “what can these models do?” to “why aren’t we doing more with them?”