Cognition hits $48B, Forus triples in four months, and the week AI funding went vertical. Cognition AI raises $2B at a $48B valuation, Forus jumps to $3B in four months, and September 8 brought the kind of capital velocity that makes 2024 rounds look quaint. Published September 9, 2026 September 8 was a day for checking valuations twice, because the numbers moved faster than most term sheets can close. Cognition doubles in four months. Cognition AI raised $2 billion in a Series E at a $48 billion valuation, led by Andreessen Horowitz and Accel with participation from Founders Fund, General Catalyst, and Avenir. That puts the AI coding assistant maker - best known for its Devin product - at nearly double the $26 billion valuation it carried in May when it raised $1 billion. The speed matters. Four months is not enough time to fundamentally rewrite a product roadmap or capture a new market segment. What changed is investor conviction that AI coding is far from winner-take-all. GitHub Copilot, Cursor, and half a dozen smaller players are all shipping code-completion tools that developers actually use, and the market is still expanding. Cognition's valuation is a bet that category leaders in AI tooling will coexist rather than consolidate - at least for now. The round also signals that late-stage AI funding has not cooled off. If anything, the bar for "mega-round" keeps moving up. A $2 billion Series E would have been headline news in 2024. In September 2026, it shares the day with three other nine-figure deals. Forus triples its valuation in four months. Healthcare automation platform Forus raised $150 million in a Series C at a $3 billion valuation, tripling the company's worth since May. Bain Capital Ventures led the round, with participation from Thrive Capital, General Catalyst, Accel, Redpoint, BoxGroup, and Pear VC. Total funding now exceeds $300 million after a $160 million raise four months ago. The jump is notable because Forus is not building a chatbot or a coding assistant. It automates prior authorizations, claims processing, and other administrative workflows that healthcare providers handle manually. The market is huge - U.S. healthcare admin costs run into the hundreds of billions annually - but the technology is less flashy than frontier models. What matters to investors is unit economics: Forus sells into hospitals and payers that measure ROI in saved labor hours and faster reimbursement cycles. That kind of defensibility justifies a fast step-up in valuation even when the broader AI narrative shifts. Forus now sits at a $3 billion valuation roughly 18 months after its founding. For context, that is faster than most SaaS unicorns from the 2010s reached their first billion-dollar mark. The difference is that AI automation companies can target enterprise buyers with immediate cost savings rather than abstract productivity gains. Infrastructure and fintech stay expensive. Networking chip maker Celero Communications raised $275 million at a valuation above $3 billion, co-led by Atreides Management, Valor Equity Partners, and Alphabet's CapitalG. Celero builds hardware to improve data center efficiency for AI workloads - think interconnects and network fabric rather than GPUs. The round reflects continued investor appetite for picks-and-shovels infrastructure plays, especially when hyperscalers like Google are active buyers. Meanwhile, Ramp is in early talks to raise $1 billion at a $60 billion valuation, up from $44 billion three months ago. The corporate card and spend management platform is not an AI company in the traditional sense, but its trajectory shows that fintech valuations are still climbing for category leaders. If the talks close, Ramp will be one of the most valuable private startups in the U.S., and it will have gotten there by selling expense automation to finance teams rather than consumer-facing AI features. Seed rounds still happen. Not every deal today cleared nine figures. Healthcare claims intelligence startup ClaimQI closed a seed round just under $500,000, backed by four private investors and healthcare firm CQuence, with a pending $150,000 Nebraska state innovation grant. The company plans to scale its claims platform across Medicare Advantage, Medicaid, Medigap, and self-funded ERISA markets. AI startup Sapien raised a new round at a $180 million valuation, led by Neo's Ali Partovi. Sapien started in financial planning software but is now building a broader platform for analyzing how operational decisions drive profit - think margin analysis powered by LLMs rather than spreadsheets. The company raised an $8.7 million seed from General Catalyst in 2024. What this week means. September 8 brought five funding announcements above $100 million, two above $1 billion, and one that tripled a valuation in four months. The common thread is that AI infrastructure and automation continue to pull capital at a pace that makes 2024 rounds look conservative. Cognition's $48 billion valuation is the headline, but Forus's trajectory - $1 billion to $3 billion in 120 days - is the more interesting data point. It suggests that applied AI in healthcare and enterprise workflows can command the same valuation velocity as frontier model labs, as long as the unit economics are legible. The other takeaway is investor concentration. Andreessen, Accel, General Catalyst, and Bain Capital Ventures all appeared in multiple deals today. The same names keep funding the same category of startup, which means the market is less about picking winners and more about maintaining exposure to a sector that is still expanding faster than any single firm can track.

Alphabet's CapitalG joins $275m round for AI chip startup Celero. US-based networking chip startup Celero Communications said on September 8 it raised US$275 million at a valuation of more than US$3 billion for AI data center networking. The round was co-led by Atreides Management, Valor Equity Partners, and CapitalG, Alphabet's venture capital arm. Atreides managing partner and CIO Gavin Baker is joining Celero's board. Celero makes coherent digital signal processors, optical chips designed to speed data transmission over fiber links between racks and nearby data centers. The startup said the technology was needed as AI systems moved more data over longer distances. Its competitors include Marvell, Broadcom, and Cisco Systems. Celero said it had validated its chip using Taiwan Semiconductor Manufacturing Co.'s 2-nanometer process technology, which it described as the most advanced commercially available. The company plans to finalize the design and begin production, a transition CEO Nariman Yousefi said he aims to start next year. Stay updated on the go with our mobile app. Get latest insights with smoother, more personalized experience through TIA mobile app. How would you feel if you could no longer use Tech in Asia? Share, tag us, and land on our Wall of!

Celero reels in $275M for its coherent digital signal processors. Celero Communications Inc., a startup that develops data center networking chips, today announced that it has raised $275 million in funding. The Series C round was led by Alphabet Inc.'s CapitalG fund, Atreides Management and Valor Equity Partners. Returning backers Sutter Hill Ventures and Maverick Silicon chipped in as well. Celero is now valued at $3 billion. Data center operators link together their facilities using fiber-optic cables that can span more than 100 miles. Those fiber-optic cables are managed by specialized chips called coherent digital signal processors, or DSPs. Irvine, California-based Celero is developing a coherent DSP specifically optimized for artificial intelligence workloads. The company says its technology provides lower latency than current chips and uses less power. Coherent DSPs are responsible for managing a data center's transceivers. Those are optical modules designed to turn digital data, which is stored as electricity, into light that can travel over fiber-optic cables. Once the light arrives at its destination, transceivers turn it back into electricity. Many workloads require the ability to determine when inbound network requests were generated. The simplest way to obtain that information is to create a physical network link through which servers can exchange timing data. However, that approach comes with tradeoffs: it adds complexity to network designs and raises costs. Coherent DSPs make it possible to obtain timing information without extra hardware. Such chips also play an important role in maintaining network reliability. Light disperses and dims inside fiber-optic cables, which causes some of the data that it contains to be lost. Coherent DSPs address that issue. They do so by extending data streams with extra bits that enable the recipient to fix certain transmission errors. Celero says its coherent DSP can power not only long-haul links between data centers but also their internal networks. In particular, the chip supports scale-up and scale-out networking. Scale-up describes the hardware that connects chips in the same rack. Scale-out networks link together different racks. Under the hood, Celero's coherent DSP combines standard digital transistors with analog circuitry. Digital transistors process discrete electrical signals that each represent a one or a zero. By contrast, analog circuitry processes a continuous current that fluctuates over time. Celero's chip uses the analog circuitry to connect to the optical hardware that it manages. The company's product engineering efforts reached a major milestone ahead of today's funding announcement. It validated a version of its chip that can be made using a two-nanometer process, the most advanced semiconductor production technology on the market. Taiwan Semiconductor Manufacturing Co.'s two-nanometer node provides 15% better performance than its predecessor. Celera says its two-nanometer design can support 1.6T and 3.2T interconnects. Such technologies transmit up to 3.2 terabits of data per second per port. Today's AI networks are significantly slower. "AI infrastructure is rapidly outgrowing the capabilities of traditional optical interconnect architectures," said Celero co-founder and Chief Executive Officer Nariman Yousefi. "Our two-nanometer silicon validation demonstrates that coherent technology can deliver the bandwidth density, power efficiency and scalability required for the next generation of AI networks." Celero will use the funding round to speed up its product development and manufacturing roadmap. Image: Celero. A message from John Furrier, co-founder of SiliconANGLE: Support its mission to keep content open and free by engaging with theCUBE community. Join theCUBE's Alumni Trust Network, where technology leaders connect, share intelligence and create opportunities. * 15M+ viewers of theCUBE videos, powering conversations across AI, cloud, cybersecurity and more * 11.4k+ theCUBE alumni - Connect with more than 11,400 tech and business leaders shaping the future through a unique trusted-based network Are you an AWS customer? Support SiliconANGLE financially by buying your AWS services from its Marketplace portal page and links: https://siliconangle.com/aws-marketplace/. About SiliconANGLE Media. SiliconANGLE Media is a recognized leader in digital media innovation, uniting breakthrough technology, strategic insights and real-time audience engagement. As the parent company of SiliconANGLE, theCUBE Network, theCUBE Research, CUBE365, theCUBE AI and theCUBE SuperStudios - with flagship locations in Silicon Valley and the New York Stock Exchange - SiliconANGLE Media operates at the intersection of media, technology and AI. Founded by tech visionaries John Furrier and Dave Vellante, SiliconANGLE Media has built a dynamic ecosystem of industry-leading digital media brands that reach 15+ million elite tech professionals. Its new proprietary theCUBE AI Video Cloud is breaking ground in audience interaction, leveraging theCUBEai.com neural network to help technology companies make data-driven decisions and stay at the forefront of industry conversations.