Jensen Huang took the stage at GTC 2026 and made a declaration: “The AGI era is here.” Whether you believe that framing or not, what NVIDIA announced underneath it was extraordinary on purely technical and commercial grounds.

Additional announcements: NemoClaw enterprise stack, $1T compute demand forecast (up from $500B), DLSS 5 for gaming, Feynman 2028 roadmap (including orbital data centers), 110 robots on the show floor, and the declaration that CUDA has now turned 20. Jensen’s forecast that every SaaS company becomes “agentic as a service” framed the entire week’s AI narrative.

📖 Full GTC 2026 coverage →

On March 17, CNBC confirmed what had been rumored for weeks: OpenAI has hired Cynthia Gaylor (former DocuSign CFO) as its investor relations head, with a Q4 2026 IPO target. The announcement crystallized several concurrent moves: Nick Turley now leads ChatGPT product, the company is rebranding ChatGPT as an “indispensable productivity tool,” and the enterprise expansion is explicit.

Bridge investors are being offered a 17.5% guaranteed return. GPT-5.3 Instant is now the default model. OpenAI Government is expanding. The pivot from “AI research lab” to “indispensable enterprise productivity platform” is complete — and the IPO is the formalization of that identity.

📖 Full OpenAI IPO analysis →

On March 11, pro-Iranian hackers attacked Stryker Corporation — one of the world’s largest medical device and equipment manufacturers. They wiped thousands of computers, triggering a global shutdown of ordering, shipping, and production systems. By March 18, Bloomberg confirmed what patient advocates had feared: surgeries were being delayed.

The Stryker attack represents a dangerous escalation: healthcare infrastructure as a deliberate geopolitical target. When medical device supply chains go down, patients waiting for implants, surgical tools, and emergency equipment are not abstract statistics — they are real people whose care is being interrupted by state-affiliated hackers. The 31% healthcare ransomware figure from BlackFog makes clear this is not an anomaly.

Apple shipped the MacBook Neo on March 11 at $599 — the most aggressive price point the company has hit for a full macOS laptop. The early reviews are landing this week, and the verdict is consistent: CNBC called it “the clearest attempt to challenge Google’s Chromebook dominance,” and the education market is taking notice.

The benchmark paradox worth noting: the iPhone 17e (A19 chip, same $599 price) scores slightly higher in Geekbench than the MacBook Neo’s A18 Pro. Apple has an internal chip segmentation strategy playing out in public — the Neo is intentionally positioned to not cannibalize higher MacBook tiers. For the education market and first-time Mac buyers, this matters not at all: $599 for a full Liquid Retina macOS machine is a watershed moment.

The Galaxy S26 shipped March 11 and landed in user hands this week. The real-world reports are trickling in, and the emerging consensus aligns with the spec sheet analysis: impressive upgrades, wrong upgrade cycle for most existing Samsung owners.

📖 Full S26 real-world review →

Every story this week orbits the same gravitational center: the scale of AI infrastructure investment is now so large that it is reshaping every adjacent industry simultaneously. NVIDIA forecast $1 trillion in compute demand. OpenAI’s $840B valuation reflects a bet that AI-native productivity tools are worth more than most sovereign technology companies. Apple’s $599 MacBook Neo is possible precisely because A-series chip yields and AI capabilities have matured to the point where the economics support mass-market pricing.

The $1 trillion figure isn’t abstract. It manifests in Stryker’s surgeries being delayed (healthcare infrastructure is under-hardened relative to the threat environment that $1T in AI investment creates). It manifests in TELUS Digital’s 700TB breach (cloud credential attacks scale with the surface area of cloud-first infrastructure). It manifests in OpenAI’s IPO (the financial markets are pricing in a world where AI productivity tools generate recurring enterprise revenue at scale).

The week of March 16 will be a chapter in the history of 2026. The $1 trillion compute demand number is the annotation that makes sense of the entire chapter.