The Morning Download: Remember AI

What's up: Anthropic is raising $10 billion at a $350 billion valuation; Snowflake is set to acquire Observe; Character.AI and Google will settle lawsuits over teen suicides.

Good morning. Samsung expects its operating profit to triple to a record on surging memory chip demand. “The world’s largest memory-chip maker has been benefiting from a surge in semiconductor prices over the past year amid increasing AI adoption and as companies pour billions of dollars into AI infrastructure,” WSJ reports.

Memory prices are expected to surge 40% to 50% in the current quarter, after a similar gain in the final quarter of 2025, Barron’s says, citing Counterpoint Research. That is good news for memory and storage companies such as SanDisk, Seagate Technology and Western Digital, although rising prices for NAND flash memory and DRAM memory chips could have an impact on consumer markets, Barron’s warns.

New architectures. That’s why companies such as Nvidia are developing new approaches to long- and short-term memory. On Monday at CES, Nvidia launched its Inference Context Memory Storage Platform, a new class of storage infrastructure.

The enterprise impact. The soaring demand for memory and storage is more than just a supply and demand problem. New architectures will facilitate the deployment of AI at scale by improving functionality and price performance, especially when it comes to agents. Nvidia said the new platform “extends AI agents’ long-term memory and enables high-bandwidth sharing of context across clusters of rack-scale AI systems—boosting tokens per seconds and power efficiency by up to 5x.”

As the broader AI tech stack races to catch up with models and GPUs, a solid foundation for AI adoption across the enterprise is being constructed.

The Wall Street Journal Leadership Institute’s Belle Lin writes for the newsletter on a $1 billion acquisition by Snowflake.

Snowflake plans to acquire the application observability startup Observe, aiming to improve its monitoring capabilities for AI agents and shore up its own AI-based offerings. The deal is worth about $1 billion in cash and stock, according to a person familiar with the matter.

The acquisition is expected to close in January.

Snowflake, a cloud-based data-warehousing company, said a key component of the acquisition is to help customers directly manage the performance of their software, including AI agents. Observe, which was built on the Snowflake platform, also offers its own AI-based agent to make the process of app-monitoring easier.

AI agents have become a critical piece of the puzzle for businesses to achieve a return on their technology investments. And for tech vendors, the autonomous AI bots are expected to power a new generation of products and capabilities.

While it’s still early days for AI agents, according to Jeremy Burton, Observe’s chief executive, the technology itself is more complex than software of the past. And that makes the need for IT teams to monitor it even more important, he said.

The majority of San Mateo, Calif.-based Observe’s roughly 180 employees will be joining Snowflake. The startup, founded in 2017, has about 100 customers, including major banks like Capital One, Burton said. He declined to share the company’s annual recurring revenue. In 2024, Burton said annual recurring revenue was $21 million.

Snowflake’s acquisition mirrors a larger trend in which data infrastructure companies are buying up smaller firms that own pieces of the corporate data pie, said Sanjeev Mohan, a tech analyst and founder of advisory firm SanjMo. Taken together, they point to the need for centralized corporate data sources that can be tapped by large language models, he said.

Derek Hernandez, a senior emerging tech analyst at PitchBook, said companies like Observe are viewed by investors as “essential pick-and-shovel players for the AI era.” Observability tools are part of how businesses will scale generative AI by helping manage their data loads and the cost of graphics processing units, Hernandez added.

New financing kicks off what is likely to be another banner year for AI startup funding. 

Anthropic plans to raise $10 billion at a valuation of $350 billion before the new investment, nearly doubling its valuation from four months ago, WSJ reports. GIC, Singapore’s sovereign-wealth fund, and Coatue Management plan to lead the new financing, according to people familiar with the matter. The Claude developer has raised tens of billions to date and is expected to go public this year.

Anthropic also expects to break even for the first time in 2028, putting it on track to turn a profit quicker than rival OpenAI.

Not every story concerning AI in the courtroom has to be about attorneys submitting briefs riddled with citations to fictional cases and factual errors.

Increasingly, judges are citing the technology's ability to help summarize legal filings, prepare for hearings and map out decisions, WSJ reports.

A 'case study.' To test AI's capabilities, Federal judge Xavier Rodriguez ran a case involving hundreds of thousands of exhibits and testimony from more than 70 witnesses through an AI tool. It took the tool minutes to produce a first draft that included the findings of fact that had taken his team weeks to produce.

A high-profile case involving AI and mental health is resolved.

Startup Character.AI and Google have agreed to settle several lawsuits with families of teenagers who killed or harmed themselves after interacting with the startup’s chatbot, WSJ reports. The families had sued the companies in states including Florida, Colorado, Texas and New York for hurting their children and negligence. OpenAI and Meta Platforms are facing similar lawsuits.

Meanwhile Grok is facing mounting criticism from authorities in the European Union, France and India.

A researcher tells Bloomberg that Elon Musk’s X is a top source for images of people that have been undressed by AI. During a single one-hour session, X’s Grok chatbot generated some 6,700 images that were sexually suggestive, according to Genevieve Oh, a social-media and deepfake researcher. Victims tell Bloomberg that getting X to remove images has been difficult. 

JPMorgan Chase’s asset-management unit is swapping out proxy-advisory firms for an internal AI-powered platform it is calling Proxy IQ, WSJ reports.

The bank will use the platform to assist in casting shareholder votes on U.S. companies and also analyze data from more than 3,000 annual company meetings and provide recommendations to portfolio managers, according to an internal memo.

JPMorgan Chief Executive Jamie Dimon has been one of the most outspoken critics of the industry, calling (human) proxy advisers incompetent.