Lenovo's third quarter earnings report offered a glimpse into how high memory prices are going to impact consumer electronics. PCs and smartphones will take a hit on price increases, but Lenovo expects to make up the sales on higher-end servers and other gear.

Yang Yuanqing, CEO of Lenovo, said:

"The supply shortage and rising cost situation is precedented and still not finished. The DRAM cost increased by 50% last quarter, but the current quarter versus last quarter almost doubled again even with the contract price. So this structural imbalance between supply and demand is not simply a short-term fluctuation. It's likely to have a prolonged impact on the industry throughout this year.

Looking ahead, high material cost will probably constrain the demand for PC smartphone. But that's just from a unit volume point of view. But given the higher pricing and the market shifting to the premier segment because of AI PC, AI phone, we believe overall PC revenue will still grow year over year."

Anthropic said it will cover the electricity price increases to consumers from its data centers. In a post, Anthropic said "AI companies shouldn't leave American ratepayers to pick up the tab."

It's a nice thought, but paying for grid upgrades, bringing net new power generation, reducing strain on the grid and investing in local communities seems like a roundabout way to lower my bill that’s already gone up.

Zoom out: Anthropic's missive follows a similar post from Microsoft. These efforts from AI giants are designed to head off NIMBY as communities realize that having a massive AI data center in your back yard may not be so swell. NIMBY was a key acronym in my 2026 predictions.

Charlie Guo read the system cards for OpenAI's GPT-5.3-Codex and Anthropic's Claude Opus 4.6 so you didn't have to. Thanks for your service Charlie! The post is worth a read, but what was most notable was this:

Both models also exhibited what researchers call “evaluation awareness” - they can often “tell” when they’re in a test environment, and they behave differently as a result...

Which, of course, creates a problem. If AI systems behave better under observation and worse when they think they’re not being watched, how do we accurately measure their true capabilities and alignment?

Good question.

Meta said its variable interest entity (VIE) partnership with Blue Owl Capital to develop its Hyperion data center campus was flagged by its auditor Ernst & Young.

The company's joint venture with Blue Owl Capital was announced in October. Under the partnership, Blue Owl Capital will own 80% of the joint venture and Meta will own 20%. The deal covered $27 billion in total Hyperion development costs.

At the time, the deal made me go hmmm since it was fairly obvious Meta didn't want the costs on its balance sheet. The Wall Street Journal spotted the Ernst & Young disclosure in Meta's 10-K filing. "Auditing the Company’s determination of the primary beneficiary of the VIE was especially challenging due to the significant judgment required in determining the activities that most significantly affect the VIE’s economic performance based on the purpose and design of the entity and assessing whether the Company has the power to direct those activities," said Meta.

Yes, off balance sheet arrangements are complicated.

Meta said it addressed the VIE in its audit. "We obtained an understanding, evaluated the design and tested the operating effectiveness of controls over the Company’s determination of the primary beneficiary of the VIE, including controls relating to the determination of the activities that most significantly affect the VIE’s economic performance and assessing which party has the power to direct those activities," said Meta.

Bottom line: VIE arrangements are already under scrutiny. You can expect more with AI capex surging. And while we’re at it depreciation expenses are hotly debated too.

Classiq, a quantum software company, said it has released Classiq 1.0, the milestone release for its quantum computing engineering and development platform.

Features in Classiq 1.0 include hardware aware execution across simulators and QPUs, debugging transparency, support for generative quantum function and correct-by-construction enforcement.

Classiq last year raised $110 million in funding.

Salesforce said it will acquire Cimulate, a company that specializes in product discovery for commerce. Salesforce said the acquisition will bolster its agentic commerce efforts as well as Agentforce Commerce. Cimulate provides "an intent-aware context engine purpose-built for retail."

Cimulate's flagship product is CommerceGPT, an LLM-based operating system for retail.

Cloudflare posted fourth quarter revenue of $614.5 million, up 34% from a year ago, with adjusted earnings per share of 28 cents, a penny ahead of estimates.

The company projected first quarter revenue of $620 million to $621 million, well above the $614 million Wall Street estimate. For 2026, Cloudflare projected revenue of $2.785 billion to $2.795 billion, above the $2.74 billion consensus estimate.

Matthew Prince, CEO of Cloudflare, said in a statement that the company's network is seeing more traffic due to AI. "If agents are the new users of the web, Cloudflare is the platform they run on and the network they pass through. This creates a virtuous flywheel: more agents drive more code to Cloudflare Workers, which fuels demand for our performance, security, and networking services," said Prince.

Teradata reported fourth quarter earnings of 38 cents a share on revenue of $421 million, up 3% from a year ago. Non-GAAP earnings were 74 cents a share, 18 cents a share above estimates. CEO Steve McMillan said the company's Autonomous AI and Knowledge platform is seeing strong demand.

For 2025, Teradata reported earnings of $1.35 a share on revenue of $1.66 billion, down 5% from a year ago. For 2026, Teradata projected revenue to fall 2% or be flat.

OpenAI said it is beginning to test ads in ChatGPT for logged-in adult users of its Free and Go plans.

In a blog post, OpenAI said: "Ads do not influence the answers ChatGPT gives you. Answers are optimized based on what’s most helpful to you. When you see an ad, they are always clearly labeled as sponsored and visually separated from the organic answer."

QAD | Redzone inked a strategic partnership with Tata Consultancy Services (TCS). The partnership is focused infusing AI into manufacturing operations in the mid-market. TCS will also provide QAD | Redzone with engineering services.

Kyndryl's third quarter earnings missed estimates, and the company named a new CFO. Separately, Kyndryl said its accounting is being reviewed by the Securities and Exchange Commission. Kyndryl shares were halved.

The company reported third quarter net income of $57 million, or 25 cents a share, on revenue of $3.9 billion, up 3% from a year ago. Non-GAAP earnings were 52 cents a share compared to Wall Street estimates of 60 cents a share.

In addition, Kyndryl said it has named Harsh Chugh interim CFO. Kyndryl also replaced its general counsel and corporate controller.

In an SEC filing, Kyndryl said it couldn't file its quarterly report on time. "The Company, through the Audit Committee of its Board of Directors, is reviewing its cash management practices, related disclosures (including regarding the drivers of the Company’s adjusted free cash flow metric), the efficacy of the Company’s internal control over financial reporting, and certain other matters following the Company’s receipt of voluntary document requests from the Division of Enforcement of the Securities and Exchange Commission (“SEC”) relating to such matters," said Kyndryl. "Due to this review, the finalization of the Quarterly Report, including the Company’s assessment of internal control over financial reporting, requires additional time to complete."

Siddhant Khare riffed on AI fatigue from a software developer's view. The upshot is he shipped more code than ever, but is also more drained than ever.

Before AI, my job was: think about a problem, write code, test it, ship it. I was the creator. The maker. That's what drew most of us to engineering in the first place - the act of building.

After AI, my job increasingly became: prompt, wait, read output, evaluate output, decide if output is correct, decide if output is safe, decide if output matches the architecture, fix the parts that don't, re-prompt, repeat. I became a reviewer. A judge. A quality inspector on an assembly line that never stops.

This is a fundamentally different kind of work. Creating is energizing. Reviewing is draining.