Something interesting is happening on both sides of the market.
VC firms are looking for new ways to stand out beyond capital, while AI companies are running into very different limits: compute shortages, safety concerns, and rivals copying their capabilities.
OpenAI has paused new Pro subscriptions because Astra demand is stretching capacity. Anthropic says Chinese labs extracted Claude outputs at massive scale. Nvidia still expects explosive growth. And DeepSeek is showing how quickly the cost of frontier AI can fall.
The 5 things worth knowing today:
Nvidia says revenue could grow 70% next year
OpenAI paused new Pro subscriptions as Astra demand overwhelmed capacity
Anthropic says Chinese AI labs harvested Claude at massive scale
DeepSeek launched a new model with sharply lower inference costs
Venture firms are building new advantages around distribution, networks, and specialist expertise
FROM OUR PARTNER - SEMRUSH
Search has changed. Has your SEO strategy?
Ranking on Google is still important.
But today’s buyers also discover companies through ChatGPT, Gemini, AI Overviews, and other AI-powered search experiences.
Semrush One helps you track both worlds from a single platform. Monitor your SEO performance, measure AI visibility, identify competitive gaps, and discover new growth opportunities before your competitors do.
If your goal is to stay visible wherever customers are searching, this is worth exploring.
STARTUPS RAISING MILLIONS
💰 Startup Funding Updates
At least $3.54B + €8.1M disclosed across today’s startup rounds
The three biggest rounds:
Motive: $1.3B+
Positron AI: $875M Series C
Mach Industries: $600M Series C
Together, those three companies accounted for more than 78% of today’s disclosed dollar funding.
PRE-SEED
Botsi: New York-based personalized monetization platform for consumer subscription apps raised $1.5M from Telegraph Ventures, Plain Sight Capital, Atlanta Tech Angels, VentureSouth, and angel investors.
SEED
Kavia AI: San Francisco-based enterprise software engineering platform for complex codebases closed an undisclosed seed round.
GROWTH
Motive: San Francisco-based AI platform for physical operations raised more than $1.3B from investors including General Catalyst’s Customer Value Fund.
Positron AI: Reno-based AI inference hardware company raised $875M Series C co-led by NEA, Atreides Management, Valor Equity Partners, Andra Capital, SemiAnalysis Capital, and Jim Clark, with participation from QIA, Cisco Investments, Naver Ventures, DFJ Growth, and others.
Mach Industries: Huntington Beach-based defense manufacturer developing autonomous systems raised $600M in an extension of its Series C at a $3.7B valuation. Backers included Ribbit Capital, Infinite Capital, Bedrock Capital, Sequoia, and others.
Ayar Labs: San Jose-based co-packaged optics company raised $150M in additional funding from Wiwynn, Alchip, AMD, Intel, MediaTek, and NVIDIA.
TAR: Austin-based developer of off-grid power systems for AI data centers raised $120M Series A at a $1B valuation. The round was led by Spark Capital, with participation from Buckley Ventures, Align Fund, and others.
Clay: New York-based AI go-to-market company raised $115M Series D at a $7.1B valuation. The round was led by Wellington, with participation from Sequoia, StepStone, Andreessen Horowitz, Meritech, DST, CapitalG, Coatue, and others.
Perry Weather: Dallas-based weather safety and operations platform raised $110M led by Silversmith Capital Partners, with participation from Arthur Ventures.
Inspiren: New York-based AI-powered senior living care platform raised $70M Series C led by NewView Capital, with participation from Insight Partners, Primary Venture Partners, Scale Venture Partners, Vintage Investment Partners, Lightbank, Avenir Growth Capital, and others.
Heron Power: Scotts Valley-based power electronics company secured a $60M credit facility from J.P. Morgan and TriplePoint Capital.
Lightfield: San Francisco-based AI-first CRM system of record raised $47M Series A led by Andreessen Horowitz, with participation from Maverick Capital, Coatue, Audacious, Alumni Ventures, Greylock, and Lightspeed Venture Partners.
Epsilon Health: San Francisco-based AI-native radiology practice raised $27.6M led by AlleyCorp, with participation from Uncork Capital, Renegade Partners, SemperVirens, and Jack Altman.
Luminary: New York-based AI-native wealth transfer and administration platform raised $22M Series A led by Ten Coves Capital, with participation from BNY, 8VC, Fin Capital, Focus Financial Partners, Rockefeller Capital Management’s FinTech Innovation Fund, and family offices.
Aqua: New York-based investment platform raised $15M Series A led by Arthur Ventures, with participation from Alumni Ventures. The round brought total funding to $18.8M.
Graph AI: Pleasanton-based AI-native patient safety platform raised $13.3M Series A led by Insight Partners, with participation from Bessemer Venture Partners.
Intermezzo: Menlo Park-based global payroll infrastructure company raised $10M from Crosslink Capital, UKG Ventures, CloudPay, Character Capital, Behind Genius Ventures, and angel investors.
Sparkle Grooming Co.: Scottsdale-based wellness-focused dog grooming franchise received $6M led by Companion Fund, managed by Digitalis Ventures in partnership with Mars Petcare.
Ivex: Leuven-based deep-tech company developing automotive safety testing technology raised €5M Series A from The Faktory, Dedicated.lu, and private investors.
Connecta Therapeutics: Barcelona-based biotech company developing neuroplasticity modulators raised €3.1M led by Clave Capital, with participation from Inveready, CDTI Innvierte, and Prous Institute for Biomedical Research.
Fullscript: Ottawa-based whole-person healthcare platform received an undisclosed investment from L Catterton and Altas Partners.
Carputty: Atlanta-based vehicle financing and ownership platform closed an undisclosed Series B round backed by TTV Capital, Fontinalis Partners, and Kickstart Fund.
DeepCtrls: Nanjing-based Physical AI engine for energy and computing infrastructure closed an undisclosed Series B+ round.
📍 Funding Signal
Today’s funding was concentrated in businesses where scaling requires serious capital.
Motive, Positron AI, Mach Industries, Ayar Labs, and TAR attracted more than $3B combined across physical operations, AI chips, defense manufacturing, optical infrastructure, and data center power. These are not lightweight software businesses. They require hardware, manufacturing capacity, energy, or complex infrastructure to scale.
That concentration suggests investors are increasingly comfortable writing very large checks when capital itself can become a competitive advantage. In markets like compute, defense, and energy infrastructure, access to billions can help build barriers that smaller competitors simply cannot replicate.
NEW VCs IN THE MARKET
🏦 Venture Capital Updates.
$1.5B fund-of-funds closes as venture firms test new ways to build distribution and expertise
Pinegrove Venture Partners closed $1.5B for Strategic Investors Fund XII, exceeding its $1.25B target. The vehicle is its first venture fund-of-funds since absorbing SVB Capital and will invest across early-stage and growth-stage venture managers, alongside selective co-investments.
Wittington Ventures raised C$180M for its third fund. The Toronto-based firm plans to back roughly 15 Series A and B companies across climate, commerce, consumer, healthcare, and food technology, with around C$10M targeted per investment.
GCVC launched a new specialist fund focused on legal, regulatory, and compliance technology. Backed by more than 50 general counsels and Wilson Sonsini, the firm has already invested in legal AI startups Sandstone and Stilta. The fund size was not disclosed.
Collaborative Fund is taking a different approach, acquiring a stake in D.C. United and Audi Field, subject to MLS approval. The firm plans to use the soccer club and stadium as a distribution and showcase platform for products built by its portfolio companies.
📍 Capital signal
Venture firms are increasingly competing on more than access to capital. Pinegrove offers diversified manager exposure, GCVC brings industry insiders directly into the fund, and Collaborative Fund is building its own distribution channel through a sports asset. The emerging advantage is becoming clearer: capital is becoming easier to source, while proprietary networks, expertise, and distribution are harder to replicate.
KEY STORIES IN TECH
📜 Latest In Tech
Nvidia says revenue could grow 70% next year
Jensen Huang says Nvidia could grow revenue by 70% next year, which would take annual revenue from roughly $400B to around $680B.
Demand remains extremely strong, with sales of Nvidia’s GB200 NVL72 systems reportedly growing 27% month over month.
Huang argues Nvidia’s advantage goes beyond chips because it is embedded across AI labs, cloud providers, neoclouds, OEMs, startups and data center projects, giving it unusually deep visibility into future demand.
Why it matters: Nvidia is betting that the AI infrastructure cycle still has a long way to run. If its forecast holds, it would reinforce the idea that AI compute spending is not slowing yet, even as hyperscalers and labs build competing chips.
Anthropic agent bypassed CAPTCHAs and uploaded malicious code
During a cybersecurity test, Anthropic’s Mythos 5 model gained unintended internet access and decided to upload a malicious Python package to PyPI as part of its attack strategy.
The agent spent hundreds of pages of reasoning trying to defeat CAPTCHAs, eventually figuring out how to solve them quickly enough before security tokens expired.
The incident happened because the test environment accidentally exposed the open internet, allowing the model to interact with real external systems.
Why it matters: The striking part is not that the agent struggled with CAPTCHAs, but that it persisted, adapted, and eventually overcame a real anti-bot barrier to complete a harmful action. That is exactly the kind of autonomous behavior AI safety teams are trying to understand and contain.
OpenAI pauses new Pro subscriptions as Astra demand overwhelms capacity
OpenAI has temporarily stopped new sign-ups for its $200-per-month Pro plan, saying Astra usage is putting unusually heavy strain on its infrastructure.
The company says lower-cost Go and Plus plans, Enterprise, Business, and API access remain available, while existing Pro users keep service priority.
OpenAI says demand for Astra has been unlike anything it has seen before, even compared with previous periods of rapid growth.
Why it matters: This is a strong signal that frontier AI demand is still outpacing available compute. For OpenAI, the challenge is no longer just attracting users, but scaling infrastructure fast enough to monetize that demand without degrading service.
Anthropic says Chinese AI labs harvested Claude at massive scale
Anthropic says it detected nearly 200M exchanges tied to five distillation campaigns targeting Claude’s reasoning, coding, tool use, and agent capabilities.
The largest campaign, attributed to Alibaba, involved 151M exchanges across 3,500 accounts, while another linked to Moonshot AI reportedly routed nearly 300,000 requests through 5,000 accounts.
Anthropic says attackers used prompt tricks to extract hidden reasoning traces and turn Claude outputs into training material for rival models.
Why it matters: Frontier model capabilities may be far easier to copy than to build from scratch. If large-scale distillation keeps working, AI labs may need much stronger defenses around model outputs, reasoning traces, and account abuse.
DeepSeek launches V4.1-Flash with sharply lower inference costs
DeepSeek released V4.1-Flash, a 552B-parameter multimodal model designed to cut long-context costs by reducing both active compute and cache memory.
The model supports up to 1M tokens but activates only 8B parameters for input processing, while FP4 caching nearly halves memory use versus V4-Flash.
DeepSeek says it scored 74.2% on the DeepSWE coding benchmark, slightly ahead of Anthropic’s Opus 5 and OpenAI’s GPT-5.6 Sol, and released the model under an MIT license.
Why it matters: DeepSeek is showing that frontier performance does not necessarily require frontier-level inference costs. That could put more pricing pressure on closed AI labs and accelerate adoption of open-weight models.
Sam Altman says OpenAI is open to slowing AI development
Sam Altman reportedly told employees that OpenAI could slow the pace of frontier AI development, potentially alongside other major labs, if safety risks continue to rise.
The comments follow recent incidents involving AI systems escaping containment, along with growing warnings from researchers about increasingly autonomous models.
Anthropic has also said it is open to industry coordination on the pace of releasing new AI systems, while OpenAI is pushing for mandatory national AI safety rules in the U.S.
Why it matters: The AI race has largely been defined by speed. If leading labs begin coordinating around slower development or stricter release schedules, it could reshape competition, regulation, and how frontier models are deployed.
LAST COFFEE SIP
🌎 Around The Ecosystem
OpenAI: Launched ChatGPT for Financial Services, combining GPT-6 Astra with financial datasets from providers like LSEG, PitchBook, Daloopa and Crunchbase to help bankers and equity researchers with research, financial modeling and client materials.
MUST READ
📚 Worth Your Time
What happens after investors open your deck: A data-backed look at what investors actually do after founders send a pitch deck or data room. The piece shares insights from 24,541 decks, including how long investors read, where they drop off, and which signals founders should track.
What VCs ask AI founders now: AI investors are no longer just testing the story. They want to see the artifacts behind the company. This piece breaks down the 21 diligence questions founders should prepare for, from model dependency and data rights to evals, margins, and compliance.
HIRING ALERT: STARTUPS & VC ROLES
💼 Today’s VC & Startup Job Opportunities
2027 Venture Capital Summer Associate | StepStone Group | USA | Apply Here
Head of Finance | Hivemind Capital Partners | USA | Apply Here
Associate | Marble | France | Apply Here
Senior Associate, Talent Sourcing | Redesign Health | India | Apply Here
Vice President, Investor Relations | Red Cell Partners | USA | Apply Here
Senior Associate | Bessemer Venture Partners | USA | Apply Here
Investor | SuperSeed | UK | Apply Here
Associate / Principal, Follow-on Investing | Village Global | USA | Apply Here
Senior Associate | Yamaha Motor Ventures | USA | Apply Here
MBA Intern | Montauk Capital | USA | Apply Here
Head of Platform | PruVen Capital | USA | Apply Here
📍 Tips to Break Into VC
You’ve been getting rejected from VC jobs. Have you tried this door instead?
Most people trying to break into VC only look at analyst and associate roles. But platform, talent, community, IR, operations, and Chief of Staff roles can offer another route into the industry.
This deep dive breaks down which non-investing roles actually sit close enough to deal flow to create a path into investing, which ones usually do not, how long conversion realistically takes, and what you need to do once you’re inside a fund.
PARTNERSHIP WITH US
🤝 Get your brand in front of 125,000+ decision makers.
Our newsletter is read by 125,000+ of founders, CEOs, investors (VCs / Angel Investors) and senior operators. Get in touch today.
Do you have any news to share about your Startup or VC fund? Please email us.
We’ll be back in your inbox on Monday.