April 02, 2026 – AI Daily Recap

From OpenAI’s record-shattering $122 billion raise to Nvidia’s strategic $2 billion bet on Marvell, Tuesday’s headlines signal that AI infrastructure is now being financed—and restructured—at nation-state scale. Here’s everything that matters.

OpenAI Closes $122 Billion Round at $852 Billion Valuation

OpenAI has officially completed the largest private funding round in history, raising $122 billion at an $852 billion post-money valuation. The bulk of the financing came from three tech giants: Amazon invested $50 billion, while Nvidia and SoftBank each contributed $30 billion. Microsoft also participated, though the size of its investment wasn’t disclosed. The company is now generating $2 billion in monthly revenue and approaching 1 billion weekly active users—numbers that position frontier AI development less like a software business and more like critical global infrastructure. Notably, $35 billion of Amazon’s commitment is contingent on OpenAI going public or achieving artificial general intelligence, fueling intense IPO speculation.

Nvidia Invests $2 Billion in Marvell, Launches NVLink Fusion

Nvidia is taking a $2 billion stake in Marvell Technology and opening its system to allow Marvell to integrate custom AI chips and networking equipment directly into its platform. The centerpiece of the deal is NVLink Fusion, a new platform designed to integrate Marvell’s semi-custom AI accelerators into Nvidia’s proprietary high-speed interconnect fabric. The companies will also collaborate on silicon photonics—optical interconnect technology that moves data using light instead of copper—and will leverage Nvidia’s Aerial AI-RAN technology to turn 5G/6G telecom networks into AI-capable infrastructure. Marvell shares surged 11% on the news, validating its leadership in the custom ASIC space at a time when AI infrastructure spending is expanding well beyond GPUs alone.

Oracle Cuts Up to 30,000 Jobs to Fund AI Infrastructure Push

Oracle is cutting an estimated 20,000 to 30,000 workers from its 162,000-strong workforce as it aggressively redirects capital toward AI data center infrastructure. The restructuring costs are expected to reach $2.1 billion, largely driven by severance expenses. Analysts at TD Cowen estimate the cuts could unlock $8 to $10 billion in incremental free cash flow—capital that Oracle intends to pour into expanding cloud capacity for AI customers including Nvidia, Meta, OpenAI, AMD, and xAI. Oracle’s stock is down roughly 25% this year, but shares rose nearly 6% following the announcement as investors welcomed the cost rationalization. The pattern is becoming familiar across enterprise tech: shrink headcount, grow data centers.

Huawei’s 950PR Chip Wins ByteDance and Alibaba Orders

Huawei’s new 950PR AI chip has cleared customer testing and is attracting large orders from ByteDance and Alibaba, according to Reuters. The chip is designed to excel at inference workloads—the process of running trained AI models—and is now significantly more compatible with Nvidia’s CUDA software ecosystem, which has been a key barrier for adoption. Huawei plans to ship around 750,000 units this year, with mass production beginning next month and full-scale shipments in H2 2026. The standard DDR version is priced at roughly $6,900, while a premium HBM variant will cost about $9,600. With Nvidia’s most advanced chips banned from sale in China by Washington, the 950PR positions Huawei as the leading domestic alternative for Chinese tech giants.

AI Security Startup Tenex Hits Unicorn Status With $250M Raise

Sarasota-based cybersecurity startup Tenex.ai has raised $250 million at a valuation exceeding $1 billion, achieving unicorn status just seven months after its $27 million Series A. Crosspoint Capital led the round, with participation from Shield Capital and DeepWork Capital. The company, staffed by former CISOs and financial services professionals, offers an AI-powered managed detection and response (MDR) service that blends automation with human oversight to identify and neutralize threats. Tenex has partnered with Google, Microsoft, and Amazon to deliver their security products, and plans to use the new capital for global expansion. The leap from $27M to $250M reflects intense investor demand for AI-native cybersecurity at scale.

Google Ships Gemini 3.1 Flash-Lite, Its Cheapest Model Yet

Google has released Gemini 3.1 Flash-Lite in preview, its most cost-efficient model to date. Priced at just $0.25 per million input tokens and $1.50 per million output tokens, Flash-Lite delivers 2.5x faster time to first answer and 45% faster output generation compared to Gemini 2.5 Flash—while matching or exceeding its quality across key benchmarks. Available now through the Gemini API in Google AI Studio and Vertex AI, the model is optimized for high-volume, low-latency use cases like content moderation, translation, and UI generation. At one-eighth the cost of Gemini Pro, Flash-Lite signals Google’s aggressive push to win the high-throughput inference market where margins are razor-thin but volume is enormous.

The Big Picture

Tuesday painted a vivid portrait of where the AI industry stands in early April 2026: capital is concentrating at unprecedented scale, infrastructure is the bottleneck everyone is racing to solve, and the cost of participation keeps climbing. OpenAI’s $852 billion valuation and Nvidia’s Marvell deal show that the biggest players are locking in strategic positions for a multi-year buildout. Oracle’s mass layoffs underscore a harsh reality—companies are choosing silicon over headcount. Meanwhile, Huawei’s 950PR chip reveals that U.S. export controls are accelerating, not slowing, China’s push for domestic AI self-sufficiency. On the startup side, Tenex’s meteoric rise to unicorn status proves that AI security is no longer a niche—it’s a necessity. And Google’s Flash-Lite pricing offensive is a reminder that the race to commoditize inference is well underway. The message is clear: AI in 2026 is an infrastructure story, and the winners will be whoever can build, connect, and power the compute fastest.

Sources

• OpenAI Raises Record $122B at $852B Valuation – Tech Startups
• OpenAI Closes Funding Round at $852B Valuation – CNBC
• Nvidia Invests $2 Billion in Marvell – American Bazaar
• The $2 Billion Nvidia Deal With Marvell – The Next Platform
• Oracle Cutting Thousands in Latest Layoff Round – CNBC
• Oracle Lays Off Workers as AI Investment Accelerates – Allwork.Space
• Huawei’s New AI Chip Finds Favor with ByteDance, Alibaba – CNBC
• Huawei’s 950PR: The Chip That Learned to Speak CUDA – Shashi
• Tenex Raises $250M at $1B+ Valuation – Tech Funding News
• Cybersecurity Startup Tenex Raises $250M – SiliconANGLE
• Gemini 3.1 Flash Lite – Google Blog
• Google Releases Gemini 3.1 Flash-Lite at 1/8th the Cost of Pro – VentureBeat