
OHUB @ohub
🚨 OHUBNext | Sarvam Raises $234M at a $1.5B Valuation as PhysicsX Hits $2.4B and the AI Compute Stack Concentrates
🚨 OHUBNext | Sarvam Raises $234M at a $1.5B Valuation as PhysicsX Hits $2.4B and the AI Compute Stack Concentrates
📍 Sarvam crossed into unicorn territory yesterday with a $234 million first close of a $300 million Series B at a $1.5 billion post-money valuation, days after PhysicsX raised $300 million at a $2.4 billion valuation to scale AI that replaces physics simulation. The capital defining this cycle is concentrating into companies that own a layer of the AI stack, and the founders who position themselves where that money clusters will own the decade.
─────
Hey Builders!
The story of June 2026 is concentration. Sarvam, India's full-stack sovereign AI company, raised $234 million at a $1.5 billion valuation. PhysicsX, an AI-native engineering company, raised $300 million at a $2.4 billion valuation. Crusoe's contracted infrastructure capacity now approaches 5 gigawatts. In every case the capital is clustering around companies that own something structural, the model, the engineering layer, the compute, rather than something incremental. The boom is real, but it is not evenly distributed, and understanding where it pools is the difference between riding the wave and watching it.
The clearest signal came from Sarvam's cap table. HCLTech wrote a $150 million strategic check, taking a position north of 10% in a company that trained its frontier models from scratch and now handles over 2 million conversations a day. This is not a model-layer bet in the abstract. It is an enterprise services giant buying its way into owning the infrastructure that underpins its own future delivery. When a publicly traded IT firm decides the smartest use of its capital is equity in a frontier lab rather than a vendor contract, the signal to founders is unmistakable.
PhysicsX tells the same story from the industrial side. The London company builds AI models that predict physical behavior in seconds rather than the hours or days that conventional simulation takes, and it doubled year-over-year revenue while doubling headcount to more than 300 people, per the company. That growth pulled in a $300 million Series C at a $2.4 billion valuation, led by Temasek with NVIDIA, Siemens, and Applied Materials on the cap table. Capital is rewarding the companies that turn AI into measurable engineering output, not the ones that merely talk about it.
For the builders OHUB serves, the Crunchbase data on Black founders is the necessary counterweight. Black founders raised $643 million in Q1 2026, the highest quarterly figure since 2022, yet that sum was driven by a handful of large rounds and still represented a fraction of the roughly $252 billion that flowed to U.S. startups in the same period. Records and scarcity coexist. The capital is real, but it clusters around a narrow set of companies and a narrower set of relationships.
The infrastructure layer underneath all of it kept compounding. Crusoe announced its contracted AI infrastructure capacity now approaches 5 gigawatts, and Hydra Host raised $100 million at a valuation near $800 million to monetize underused GPU capacity across independent data centers. The pattern across every story today is ownership. Who owns the model, who owns the rails, who owns the compute. That is the question this brief returns to.
Read on for the five stories defining the moment and what each one means for the company you are building.
─────
1️⃣ Sarvam Raises $234M at a $1.5B Valuation to Build India's Sovereign AI Stack
Sarvam, India's full-stack sovereign AI company building training and inference infrastructure, frontier models, and a go-to-market motion across enterprises, developers, and government, announced on June 15 that it raised $234 million in the first close of a $300 million Series B at a $1.5 billion post-money valuation (sarvam.ai/...). HCLTech and Bessemer Venture Partners led the round, with continued support from existing backers Khosla Ventures and Peak XV Partners. The raise made Sarvam India's fifth billion-dollar startup of 2026.
HCLTech invested $150 million as the lead strategic investor, an unusual structure in which a publicly traded IT services company takes a double-digit equity stake in a frontier AI lab rather than simply contracting for its services. The capital funds Sarvam's research on its next frontier model for agentic, coding, and cybersecurity use cases, alongside compute access at scale.
Sarvam has shipped at pace. The company trained Sarvam 105B and Sarvam 30B from scratch in India, digitized over 35 million pages of insurance and land records through Sarvam Vision, and now processes 10 million API calls daily. Its conversational platform handles over 2 million interactions a day, with usage doubling in the prior two months.
💡 For Founders
The strategic investor who takes equity instead of buying your service is the most durable capital you can raise, because their incentive shifts from extracting margin to growing your enterprise value. When you pitch a large incumbent, do not lead with what you will sell them. Lead with the position they cannot afford to let a competitor own.
─────
2️⃣ PhysicsX Raises $300M at a $2.4B Valuation to Replace Physics Simulation With AI
PhysicsX, the physics AI company for industrials, announced on June 8 an oversubscribed $300 million Series C at a valuation of approximately $2.4 billion, per the company's newsroom (physicsx.ai/...). Temasek led the round, with new investors M&G Investments and Intrepid Growth Partners joining existing backers including NVIDIA, Siemens, Applied Materials, Atomico, and General Catalyst.
The company builds an AI-native engineering platform whose models predict physical behavior in seconds rather than the hours or days that conventional simulation requires. The platform is deployed across aerospace and defense, semiconductors, automotive, energy, and materials manufacturing, letting engineering teams evaluate orders of magnitude more design variants and carry physics insight from early design through to real-time digital twins in operation.
The raise came amid rapid growth. PhysicsX doubled year-over-year recognized revenue, tripled booked revenue, and more than doubled its customer count over the past year, while its team grew past 300 people. The capital funds global expansion and frontier research into larger pre-trained models the company calls Large Physics Models.
💡 For Founders
PhysicsX is collapsing a workflow that used to take days into seconds, and capital rewarded the company that owns that speed advantage. Find the slow, expensive bottleneck in your own industry that everyone tolerates as a cost of doing business, and build the layer that removes it. The deepest moats come from compressing time, not adding features.
─────
3️⃣ Black Founders Raise $643M in Q1 2026, the Highest Quarterly Total Since 2022
Black founders in the United States raised $643 million in the first quarter of 2026, the highest quarterly figure since 2022, according to Crunchbase data reported by TechCrunch. The total represented nearly 70% of the entire prior year's fundraising by Black founders compressed into a single quarter.
The surge was driven by a small number of large transactions, including SambaNova's $350 million Series E, Noviq's $75 million Series B, and Harper's $47 million round. Strip out the top deals and the broader picture is unchanged. The $643 million represented a small fraction of the roughly $252 billion that flowed to U.S. startups in the same period.
Gené Teare of Crunchbase pointed to persistent structural barriers, including limited access to networks and early introductions and a venture market increasingly concentrated in AI. The record and the scarcity are not in tension. They are the predictable output of a market where capital clusters around a narrow set of companies and relationships.
💡 For Founders
A record quarter built on a few mega-rounds is a signal about deal concentration, not access. If you are raising outside the AI core, your strategy is relationship infrastructure built months before the pitch. Map the specific partners writing checks in your category now, and get a warm introduction into the room before you need the money.
─────
4️⃣ Crusoe's Contracted AI Infrastructure Capacity Approaches 5 Gigawatts
Crusoe announced on June 9 that its contracted AI infrastructure capacity across data centers and cloud now approaches 5 gigawatts, per the company's newsroom. The milestone underscores how quickly the physical layer of the AI economy is being locked up under long-term contracts.
The company's vertically integrated model spans energy sourcing, AI-optimized data center construction, and its Crusoe Cloud platform. Its flagship 1.2 gigawatt Abilene, Texas campus went live roughly a year after construction began, and Crusoe has announced additional gigawatt-scale campuses including a 1.8 gigawatt site in Wyoming and a 900 megawatt campus in Abilene supporting Microsoft.
Crusoe closed a $1.375 billion Series E at a valuation above $10 billion in late October 2025, co-led by Valor Equity Partners and Mubadala Capital with participation from NVIDIA, Founders Fund, and Fidelity. The contracted capacity figure shows that capital converting directly into committed physical infrastructure, the scarcest input in the entire AI stack.
💡 For Founders
The bottleneck in AI is no longer talent or models. It is power and compute under contract. If your company depends on inference at scale, lock your capacity commitments early and treat them as a strategic moat, because the firms that contracted gigawatts in 2025 are the ones who will have margin in 2027.
─────
5️⃣ Hydra Host Raises $100M at a Valuation Near $800M to Monetize Idle GPU Capacity
Hydra Host, a Miami-based AI infrastructure company, raised $100 million at a valuation close to $800 million, per the June 15 Tech Startups funding roundup. Kindred Ventures led the round, with participation from NVIDIA, ARK Invest, Magnetar, Founders Fund, and Flume Ventures.
The company operates a GPU marketplace that helps independent data centers monetize underused AI server capacity, effectively building a liquidity layer on top of compute that would otherwise sit idle. It is an arbitrage on the same scarcity that pushed Crusoe's contracted capacity toward 5 gigawatts.
The investor list is the story within the story. NVIDIA and Founders Fund appear in both the Crusoe and Hydra Host cap tables, a pattern that shows the smartest capital is hedging across both sides of the compute market, the firms building new capacity and the firms extracting value from existing capacity.
💡 For Founders
The biggest opportunities in any boom are not the scarce assets themselves but the infrastructure that makes them liquid and efficient. Look at where supply is stranded in your own market, the inventory, capacity, or expertise sitting idle, and build the marketplace that connects it to demand.
─────
🔧 Three moves to make this week
1️⃣ Attack the bottleneck everyone tolerates
PhysicsX raised $300 million by collapsing a workflow that took days into seconds. Identify the slow, expensive step your industry accepts as unavoidable, and build the layer that removes it. Compressing time is a deeper moat than adding features.
2️⃣ Build the strategic-investor thesis
Sarvam raised $150 million from a single strategic investor who took equity rather than buying a service. Identify the one incumbent in your sector whose position you could threaten or strengthen, and build a one-page thesis on why owning a stake in you beats competing with you. That document is worth more than any pitch deck.
3️⃣ Lock your compute and capacity early
Crusoe's path to 5 gigawatts shows that the scarce input is contracted infrastructure, not ambition. If your roadmap depends on inference at scale, secure your compute commitments now while terms are still negotiable, and treat capacity as a balance-sheet asset rather than an operating expense.
─────
💬 Quote of the Day
"We are clear that research-led innovation to create AI that works at India's scale is a very large opportunity. That means models that understand our voices, read our documents, and serve intelligence at a cost every enterprise and government can afford." — Pratyush Kumar, Co-Founder, Sarvam
─────
🏁 Stay Ahead With OHUBNext
This brief is part of what OHUBNext members get every day.
For $5.99/month — or $59/year — you get the full daily brief, access to OHUB's Library of Opportunity, self-paced certificates in High-Growth Company Building and Tech Ecosystem Investing, career accelerator tools, and live monthly labs with the OHUB team. The annual plan includes exclusive masterclasses, early course access, and wealth tools built for builders who are serious about owning what comes next.
Institutional-grade. Fraction of the cost.
🚀 Join at opportunityhub.co/next
─────
🎬 Closing Thought
The defining feature of this moment is concentration. Capital is concentrating into a narrow band of companies that own a structural layer of the AI economy, the frontier model, the engineering platform, the contracted compute. Sarvam's $1.5 billion valuation, PhysicsX's $2.4 billion, and Crusoe's 5 gigawatts are three expressions of the same logic. Institutional money has decided that owning the layer beats renting access to it, and it is paying accordingly.
For the builders OHUB exists to serve, this is the most important strategic environment in a generation, and not because the doors are open. The Black founder data makes that plain. A record $643 million in a quarter still amounted to a rounding error against $252 billion in total venture flow. The opportunity is not that capital has become fair. It is that the rules of who gets funded are being rewritten in real time, and the companies that position themselves to own a layer rather than service one will be where the next decade of value accrues.
The move is not to wait for the concentration to ease. It will not. The move is to build the kind of company that strategic capital cannot afford to ignore, to own the scarce inputs rather than rent them, and to prove your value in numbers a customer or investor cannot argue with. Sarvam did not ask permission to build India's sovereign AI stack. It built models from scratch, deployed them at population scale, and let a $150 million strategic check find it. PhysicsX did not pitch a vision. It compressed days of simulation into seconds until a $2.4 billion valuation became the obvious price.
That is the playbook. Own what comes next, and make the value impossible to dispute. The builders reading this brief have the same tools, the same models, and the same compute markets available to them as the founders raising nine figures this week. What separates them is positioning, conviction, and the willingness to move before the room tells them they are allowed to.
Move now.
─────
⚡️ OHUBNext Daily Brief - investments, edge tech, and moves that matter.
For 12+ years, OHUB has been building pathways and on-ramps to multi-generational wealth without reliance on pre-existing wealth. Through exposure, skills, entrepreneurship, capital markets, and inclusive ecosystems, we've helped people create new jobs, new companies, and new wealth.
Sarvam raises $300M Series B
Sarvam, India's full-stack sovereign AI company, announces a $300M Series B led by HCLTech.
www.sarvam.ai
