Cognition's valuation nearly doubles to $47B in a new $1B round

The agent economy repriced itself again overnight. Cognition, the studio behind the Devin coding agent, is raising around $1 billion at a valuation near $47 billion — nearly double the $26 billion it commanded just four months ago, per Bloomberg.

Cognition is raising roughly $1 billion at a valuation near $47 billion, nearly double its May mark, as demand from investors ran to almost $10 billion. Almost exactly a year after the company raised at $4 billion and months after it closed a billion-dollar round at $26 billion, the new figure puts the autonomous-coding shop above most of the labs whose models it runs on. Bloomberg reported the figure from people familiar with the matter; the company has not confirmed a signed round. The surge in interest — nearly $10 billion chasing a $1 billion allocation — is the loudest signal yet that investors now see the application layer, not the model layer, as where this cycle's value is captured. We flagged the trajectory when talks were at $40 billion — Devin maker Cognition eyes $40B valuation in new round — and tracked the underlying numbers when the business hit roughly $900 million in annualized revenue, double its pace at the left of the year — Cognition tops $900M annualized revenue.

The valuation math has outrun the revenue. Cognition was doing about $492 million in annualized revenue at its May round and roughly $900 million by late August, with executives projecting more than $1.5 billion for 2026. A $47 billion figure on that trajectory is a roughly 30x forward multiple — a bet that autonomous coding agents keep compounding well past the current wave of enterprise pilots. It is also a pointed answer to SpaceX's reported acquisition interest from earlier this year: founder Scott Wu publicly ruled out a sale, and this round prices the refusal. The open question isn't whether Devin-style agents are the future of software work — it's whether the multiples now attached to them hold when inference costs keep falling and larger labs ship their own agents with far better margins.

A House panel advanced a bipartisan ban on Chinese-linked robots in federal use. The House Energy and Commerce subcommittee voted to advance the American Security Robotics Act, which would bar federal agencies and federally funded contractors from procuring or operating ground robots, patrol tech, and humanoid robots tied to Chinese entities. The bill defines "covered unmanned ground vehicle systems" to include anything from remote surveillance vehicles to humanoids that move on their own, and it hits procurement and operation plus federal funds flowing through grants and contracts. There are carve-outs for national-security research, counter-terrorism, and systems modified to cut off data links to foreign entities — a wide enough aperture that the intelligence and defense agencies keep their toys. The deeper story is that a humanoid robot is now a national-security question, not a factory-floor curiosity; a supply-chain rule that could reshape who builds and sells robotics in the US market.

What to watch: whether Cognition's round closes at the reported number and how much appetite remains for agent startups at these multiples as inference costs compress. The robotics bill still needs full committee and floor votes.

Do $47 billion valuations for coding agents make sense, or is the market pricing autonomy before it's proven? Tell us in the comments.

AI 101 — How do AI video generators work?

An AI video generator is a model that starts with a grid of random noise and repeatedly refines it into a moving picture that matches your prompt — the same trick image generators use, stretched across time. Nothing is filmed, and nothing is stitched together from clips the model has seen. The pixels in the finished video were invented, frame by frame, all at once. Why it matters right now Kling 4.0, Kuaishou's flagship video model, went fully live on October 15 with up to 30 seconds of vid

Houmo moves compute inside memory — and A20 enters production

Beijing chipmaker Houmo used its partner conference this week to lay out two compute-in-memory routes at once — and to say the first chip built on the newer one is already in production. Houmo's A20, its first chip built on a 3D compute-in-memory architecture, has entered production. The part supports up to 16 chips fully interconnected, claims a 50% improvement in compute utilization over the previous generation, and covers models from 3B to 30B parameters, according to co-founder and product

Kling 4.0 goes live with 30-second native video at 10-bit HDR

Kuaishou's Kling AI has spent two years becoming the video model people actually finish projects with. Its new flagship, Kling 4.0, went fully live on October 15 — and this round is less about prettier frames than about the two things that make a generated shot usable in a real edit: length and control. Kling 4.0 generates up to 30 seconds in a single native pass, outputs 10-bit HDR at 4K, 1080p and 21:9, and takes up to 15 reference assets in one task. That budget splits across ten images, fi

Brain organoids classified Braille at 83% — in a trio, not alone

Two stories today: human brain tissue on a chip decoded a tactile signal, and the data-centre company behind some of the largest AI contracts in Europe turns out to be six months behind on its flagship site. Human forebrain organoids grown on a microelectrode array classified Braille letters with 61 percent accuracy on their own and 83 percent when the responses of three organoids were combined, in a peer-reviewed study from the University of Bristol. The team took event-based tactile readings

Read more

AI 101 — How do AI video generators work?

Houmo moves compute inside memory — and A20 enters production

Kling 4.0 goes live with 30-second native video at 10-bit HDR

Brain organoids classified Braille at 83% — in a trio, not alone