Weekly AI Digest: Nvidia’s $100B Bet on OpenAI, the effect of Distillation & OpenAI’s likely jump to Robotics
Week 39, 2025
This week’s big story is Nvidia’s potential $100B investment in OpenAI. Beyond the sheer number, it shows how intertwined hardware and models have become, with Nvidia having just so many advantages in this play. The deal with Nvidia isn’t purely a capital investment. The structure ties the “investment” to OpenAI’s commitment to buy Nvidia GPUs and infrastructure. In effect, Nvidia is locking in a large customer and substantial chip demand.
Analysts point out the circularity: OpenAI may receive funds from Nvidia, then use those funds (or commit to future payments) to purchase Nvidia’s hardware.
Other headlines: let’s look at model distillation in detail, and how OpenAI is doubling down on robotics.
And of course, the usual shenanigans — Elon dragging OpenAI to court and Zuckerberg swiping staff from everyone else.
Let’s dive in!
Nvidia To Invest Up To $100B in OpenAI
Nvidia is reportedly in talks to invest as much as $100B in OpenAI, according to multiple reports. This would give Nvidia closer integration with the company that drives GPU demand more than any other.
The deal, if finalized, could reshape the power map of AI, locking Nvidia deeper into the stack from chips to models to applications.
Read more on Crunchbase | NYT
Why Distillation Matters for the Next Wave of AI
Wired profiled how model distillation is becoming the core technique for shrinking models while preserving capability. Distilled models are cheaper, faster, and more deployable — key for real-world AI adoption.
Model distillation is a technique to make a large, complex AI model smaller and faster while keeping most of its performance.
As the industry pushes for efficiency at scale, distillation may become as central to AI in the future — this article on Wired explains why.
Read more on Wired
OpenAI Expands Robotics Effort
OpenAI is increasing focus on robotics as part of its AGI roadmap. The company has been hiring aggressively and setting up dedicated teams to accelerate research.
This is a shift back to its early roots in robotics, now powered by multimodal models that could handle both perception and action.
Read more on Wired
Meta Poaches OpenAI’s Yang Song
Meta has hired Yang Song, a high-profile researcher from OpenAI. He is known for foundational work in generative modeling.
The move is another sign of escalating talent wars between the major AI labs, with Meta continuing to be the main poacher.
Read more on Wired
Musk’s xAI Files Lawsuit Against Sam Altman and OpenAI
Elon Musk’s xAI has filed a new lawsuit against OpenAI and CEO Sam Altman, alleging misuse of technology and breaches of earlier agreements.
This legal clash adds more fuel to the already fractious relationships between Musk and Altman, something that seems to be a neverending feud.
Read more on The Guardian
OpenAI Introduces ChatGPT Pulse
Another week, another OpenAI product.
OpenAI launched ChatGPT Pulse, a tool that tracks and visualizes trends across user conversations. It provides aggregated insights into what people are asking, in real time.
Read more on OpenAI
Nvidia’s potential $100B bet signals consolidation between hardware and AI labs, while also pointing to some level of capital circularity withing the industry.
Distillation and robotics point to the next wave of technical focus.
And, of course, we had the full drama of lawsuits, talent moves, and new products showing how competitive and volatile the space remains.
As always, stay curious,
– Ivo
