Lambda raises up to $4B from Blackstone ahead of its IPO
The neocloud money is consolidating fast, and today's inbox shows both ends of the market: a heavyweight pre-IPO round on one side, and a Google open model you can run on a phone on the other.
Lambda is raising up to $4 billion led by Blackstone and Coatue at a $14.5 billion pre-money valuation — its last private round before a planned IPO. The Wall Street Journal reported the scoop from a letter to limited partners, and Reuters independently confirmed the headline terms: the round is led by two of the most active AI-infrastructure investors of the cycle, and the Nvidia-backed neocloud is targeting a 2027 listing subject to market conditions. The same LP letter puts Lambda's unfilled-order backlog at $50 billion in September, up from $15 billion in June — a 3.3x jump in a single quarter that says GPU rental demand is outrunning what the company can currently deliver. The round is not closed yet, and Lambda declined to comment, so treat the $4 billion as a ceiling rather than a done deal. Our take: this is the biggest test yet of whether the neocloud tier — companies renting GPUs that aren't hyperscalers — can go public as an asset class rather than a footnote, and Blackstone writing the check is the strongest signal so far that Wall Street believes it can.
Google DeepMind released EmbeddingGemma 2, a 740-million-parameter open model that maps code, images, video and audio into one shared embedding space. It ships under an Apache 2.0 license, with weights on Hugging Face today and a footprint small enough for on-device use — about 191 megabytes quantized for text-only, roughly 567 megabytes for the full multimodal model. Google claims a big jump on code search specifically, from 68.76 to 78.68 on MTEB Code, and the 8,192-token context window is four times the first version's. Embeddings are the least glamorous layer of the stack and the one every retrieval pipeline sits on — if the vendor numbers hold up independently, this is the boring open model more builders will actually ship with. We explained the layer itself in AI 101 — What is an embedding?.
Microsoft's AI-funded journal published Nobel laureate Daron Acemoglu's bear case: AI adds about 1.5% to GDP over a decade and replaces at most 5% of jobs. The forecast itself is not new — it restates Acemoglu's 2024 peer-reviewed estimate — but Microsoft platforming a skeptic in The Humanist Review of AI is the story, and the essay concedes that "advances in models have been more rapid than many, including me, predicted." Acemoglu's bottleneck argument is organizational rather than technical: companies need to reassign tasks and retrain workers before productivity shows up, a process that could drag longer than electrification did. That thesis happens to suit Microsoft's strategy of bolting AI onto products it already sells rather than betting on wholesale automation — and it sits in stark contrast to the spend-side picture in The AI buildout is 3.6% of GDP a year.
What to watch: whether Lambda's round closes at the full $4 billion, and whether any outlet independently reproduces EmbeddingGemma 2's benchmark jumps.
Is the neocloud tier a real asset class headed for public markets, or is Blackstone buying a story? Tell us in the comments.
