Cerebras Tries the IPO Window Again

Cerebras wants Nvidia’s lane

Cerebras says it made $510 million last year, which is enough to make this IPO filing real.

The AI chip startup filed to go public after shelving a 2024 IPO tied to a federal review of G42’s investment. Since then it raised fresh money at a $23 billion valuation and signed deals with AWS and OpenAI reportedly worth more than $10 billion.

This matters because most AI hardware challengers still sell a story. Cerebras can point to customers, revenue, and profit on one version of the books, even if adjusted numbers still show losses. That puts it in a smaller group.

The louder signal is customer concentration. If OpenAI and a few hyperscalers drive demand, investors will ask how durable that is once pricing gets tight. If this offering lands in mid-May, it becomes an early test of whether public markets want AI infrastructure beyond Nvidia itself. You can almost hear the trading desks at 6 a.m.

Plenty of companies can win a contract. Fewer can stay valuable after the first contract cycle ends.

ASUS is bundling the creator pipeline

ASUS did not show one product at NAB. It showed an entire production line.

At NAB Show 2026, ASUS unveiled a broad ProArt lineup: Adobe Premiere-validated displays, creator laptops, RTX 5090-powered desktops, GPU servers, and storage systems that scale to 10PB. The pitch runs from GoPro footage ingest to HDR finishing, with AI tools layered across each step.

This matters because creator hardware is crowded and margins are tight. ASUS is trying to sell a workflow instead of a laptop spec sheet. If you own the monitor, PC, server, storage, and bundled software, you become harder to replace.

The smarter move may be targeting prosumers and small studios who want enterprise-grade capability without stitching five vendors together. Expect more PC makers to package “end-to-end AI creation” instead of standalone devices. The bright trade-show lights in Las Vegas are full of that shift already.

If creators can buy one stack that mostly works, how much patience is left for custom setups?

Siemens is selling task-completing AI

Siemens says its new AI can write PLC code, configure devices, and build HMI screens inside live engineering projects.

The company launched the Eigen Engineering Agent, connected directly to TIA Portal, Siemens’ automation software used across factories and industrial systems. Siemens says more than 100 companies in 19 countries tested it, with workflows running 2 to 5 times faster and engineering efficiency up 50%.

The real shift here is not “AI for engineers.” That story is old. Siemens is pushing AI that works inside the actual system of record, where constraints, legacy code, and safety rules live. That is where most generic tools fail.

If the product works consistently, it could matter more than flashy coding demos because factory engineering budgets are large, sticky, and tied to uptime. This also raises the bar for every industrial software vendor. Customers will now expect AI that understands plant context, not a chatbot pasted onto the screen. You can picture the monitor glow on a late shop-floor desk.

Once AI starts shipping real configurations instead of suggestions, who becomes the reviewer and who becomes the operator?

Prohuman team

Free Guides

Explore our free guides and products to get into AI and master it.

All of them are free to access and would stay free for you.

Feeling generous?

You know someone who loves breakthroughs as much as you do.

Share The Prohuman it’s how smart people stay one update ahead.