--- url: 'https://qubit.capital/blog/ai-mega-rounds-funding-trends' title: How AI Mega-Round Funding Shapes Valuations and Founder Strategy author: name: Vaibhav Totuka url: 'https://qubit.capital/blog/author/vaibhav-totuka' date: '2026-05-11T04:28:00+05:30' modified: '2026-06-11T18:56:33+05:30' type: post categories: - Industry-Specific Insights image: 'https://qubit.capital/wp-content/uploads/2025/11/ai-mega-rounds-funding-trends.webp' published: true --- # How AI Mega-Round Funding Shapes Valuations and Founder Strategy Venture capital is concentrating. A handful of AI companies are pulling hundreds of millions, sometimes billions, in single rounds, and that money has to come from somewhere. The funds writing those checks are the same funds that would have led your Series A two years ago. When one deal absorbs what a fund would normally deploy across a dozen investments, the math for everyone else changes. That shift forces a real decision. You can wait and hope the cycle rebalances. You can target funds that are explicitly sitting out AI mega rounds. Or you can reframe your positioning so your company reads as infrastructure for the AI wave rather than a bet competing with it. Each path carries different timing risk, different investor conversations, and different windows, and founders who pick deliberately are closing faster right now than those still pitching the way they did in 2023. Table of Contents - [Capital is Concentrating into Fewer Mega Rounds](#capital-is-concentrating-into-fewer-mega-rounds) [The 73% of Value Going to $100M Deals](#the-73-of-value-going-to-$100m-deals) - [Why Only Category Leaders Land These Rounds](#why-only-category-leaders-land-these-rounds) - [Right-Sizing Your Raise to Real Traction](#right-sizing-your-raise-to-real-traction) - [AI Now Captures Most Venture Capital](#ai-now-captures-most-venture-capital) [From 30% to 81% of Funding](#from-30-to-81-of-funding) - [Raising While the AI Window is Open](#raising-while-the-ai-window-is-open) - [Tighter Terms Outside the AI Bucket](#tighter-terms-outside-the-ai-bucket) - [Infrastructure Builders Win the Billion Dollar Checks](#infrastructure-builders-win-the-billion-dollar-checks) [Why Infra Funding More Than Doubled](#why-infra-funding-more-than-doubled) - [Benchmarking Application Layer Raises Against Peers](#benchmarking-application-layer-raises-against-peers) - [Leading with Capital Efficiency over Compute Spend](#leading-with-capital-efficiency-over-compute-spend) - [Strategic Anchors Are Driving the Biggest Deals](#strategic-anchors-are-driving-the-biggest-deals) [Sovereign and Crossover Funds Leading Rounds](#sovereign-and-crossover-funds-leading-rounds) - [Lining Up a Credible Lead Investor](#lining-up-a-credible-lead-investor) - [Targeting Funds That Write Follow-On Checks](#targeting-funds-that-write-follow-on-checks) - [The Biggest AI Rounds Right Now](#the-biggest-ai-rounds-right-now) - [Your Next Move](#your-next-move) - [Key Takeaways](#key-takeaways) ## Capital is Concentrating into Fewer Mega Rounds The headline numbers in AI venture funding look broadly positive, but the distribution inside them tells a tighter story. Mega deals over USD 100 million made up about 73% of total AI investment value in 2025, per [OECD research](https://www.oecd.org/content/dam/oecd/en/publications/reports/2026/02/venture-capital-investments-in-artificial-intelligence-through-2025_3bcb227f/a13752f5-en.pdf), meaning the vast majority of every dollar invested in AI landed in a small number of very large deals. For most founders, that is not a market expanding in your direction; it is capital pulling away from the middle. ### The 73% of Value Going to $100M Deals The concentration is clearest when you look at what happened to repeat mega-rounds year over year. The count of US AI startups raising $100 million or more grew modestly, from 49 companies in 2024 to 55 in 2025, per [TechCrunch](https://techcrunch.com/2026/01/19/here-are-the-49-us-ai-startups-that-have-raised-100m-or-more-in-2025/). But the companies locking up capital repeatedly tell a sharper story: | Metric | 2024 | 2025 | | --- | --- | --- | | US AI startups raising $100M+ | 49 | 55 | | Companies raising multiple mega-rounds | 3 | 8 | The non-obvious read: while the entry bar for a single mega-round barely moved, the companies doing it repeatedly nearly tripled. Four startups raised rounds larger than $1 billion in 2025, with Anthropic closing two of them. That repeat-access pattern means a shrinking group is locking up an expanding share of available LP capital, compressing what is left for companies raising at that tier for the first time. ### Why Only Category Leaders Land These Rounds The starkest illustration of concentration sits at the top of the [Forbes AI 50](https://www.forbes.com/lists/ai50/). The 50 most-funded private AI companies have collectively raised $305.6 billion in venture funding. Two of them, OpenAI and Anthropic, hold $242.6 billion of that total, roughly 80 cents of every dollar raised by the entire cohort. That gap is not a revenue story; it is a narrative one. Investors writing mega-round checks are not pricing current multiples. They are paying for category ownership, the credible claim that this company defines what the category becomes. A company with strong traction but crowded or derivative positioning gets priced as a feature business rather than a platform, and platforms are the only asset class that justifies a billion-dollar check from a fund manager who also has to explain their own portfolio concentration to LPs. ### Right-Sizing Your Raise to Real Traction Most founders size their rounds by scanning recent headlines and anchoring upward. That logic works against you in this environment. Capital is not spreading wider; it is stacking deeper into a small group with proven category narratives. Targeting a round size that belongs to a different tier of company signals misread market position to an investor who has already seen the concentration data. The better frame: raise what your traction justifies, structured to reach the milestone that sharpens your category argument. The market is not short on capital overall. US AI firms captured roughly 75% of global AI VC deal value in 2025, representing about $194 billion, so the constraint is not broad scarcity. The constraint is that the top of the stack is locked by a small group of repeat raisers who keep returning for more. Positioning at the right tier for your current proof points, rather than chasing a round size that requires a category narrative you have not yet earned, keeps you competitive for the capital that is actually in play for you. [Startup funding stages](https://qubit.capital/blog/venture-capital-stages/) show how capital that once moved steadily from seed through Series A has increasingly been intercepted at the growth tier by AI companies raising at a scale that absorbs what a fund would previously have deployed across an entire vintage. ## AI Now Captures Most Venture Capital ![Infographic titled AI Now Captures Most Venture Capital showing: From 30% to 81% of, Raising While the AI Window, Tighter Terms Outside the AI.](https://qubit.capital/wp-content/uploads/2025/11/how-to-fundraise-when-ai-mega-rounds-are-eating-venture-capital-1-ai-now-capture.webp) The venture market has split into two tiers, and the divide is still growing. AI companies captured [81% of global venture funding in Q1 2026](https://tech-insider.org/q1-2026-venture-capital-297-billion-ai-startup-funding-record/), meaning four of every five dollars committed go to AI before non-AI founders step into a meeting. That is not a cycle you can wait out; it is the market you are raising into right now. ### From 30% to 81% of Funding AI’s share of global VC more than doubled, from 30% in 2022 to 61% in 2025. That 61% translated to $258.7 billion committed in a single year, and by Q1 2026 the share had climbed to 81%. Most founders read this as good news for AI and bad news for everyone else, but the structural signal is subtler: limited partners are explicitly reweighting toward AI-native funds, which compresses deployment capacity at non-AI GPs regardless of how much dry powder the broader market holds. | Year | AI Share of Global VC | What It Means for Your Raise | | --- | --- | --- | | 2022 | 30% | AI is competitive but not dominant | | 2024 | ~33% | $100-130 billion in AI; one-third of market | | 2025 | 61% | AI crosses majority; non-AI raises get harder | | Q1 2026 | 81% | Acceleration, not plateau; full tilt underway | ### Raising While the AI Window is Open The surge is demand-driven, not speculative, which is why it is holding longer than a typical hype cycle. Enterprise generative AI spend climbed from $11.5 billion in 2024 to [$37 billion in 2025](https://fundraiseinsider.com/blog/ai-startups/), and overall AI sector funding grew more than 75% year over year. Investors are following paying customers. If you are building in AI, the right move is to raise now and push your target size up, because the conditions that make that ask credible are unusually strong. The part most founders miss: within AI, infrastructure is capturing a disproportionate share. AI IT hosting and infrastructure funding surged from $47.4 billion in 2024 to $109.3 billion in 2025, a jump that tells you VCs are heavily backing picks-and-shovels alongside applications. If you are at the application layer, an AI label alone will not carry a round. Specific customer traction matters more now than it did a year ago. ### Tighter Terms Outside the AI Bucket When 81% of capital flows one direction, everything else competes for the remaining 19%, split across fintech, climate, SaaS, healthcare, and every other category. Non-AI founders are not in a frozen market, but the dynamics are materially different and most founders underestimate by how much. - **Extend runway before you start:** Raises in out-of-favor categories run longer than founders budget for. Your burn rate determines whether you can negotiate terms or just accept them. - **Target sector-specialist funds:** Generalist GPs under LP pressure to show AI exposure are a harder sell right now. Fintech, climate, and vertical SaaS funds still deploy actively and know how to underwrite your category without needing to justify it internally. - **Expect less leverage on valuation:** Thinner competition among investors in non-AI categories shifts negotiating power. If terms feel aggressive, the market is reflecting supply, not your fundamentals. - **Draw the AI connection honestly if it exists:** If your product uses AI as a specific tool, say exactly how. Vague positioning irritates investors who have seen a hundred decks claiming the label without substance behind it. The practical split: if you are inside the AI bucket, raise now, size up, and move before the window narrows. If you are outside it, plan for a longer cycle, go to investors who specialize in your space, and treat tight terms as a negotiating reality rather than a signal to abandon the raise. [Fundraising strategies](https://qubit.capital/blog/tech-startup-fundraising-strategies/) for companies outside the AI mega-round cohort have had to adapt quickly as the US market’s share of global VC has concentrated further, making fund selection and positioning more important than round timing alone. ## Infrastructure Builders Win the Billion Dollar Checks ![Infographic titled Infrastructure Builders Win the Billion Dollar Checks showing: Why Infra Funding More Than, Benchmarking Application Layer Raises Agains, Leading With Capital Ef](https://qubit.capital/wp-content/uploads/2025/11/how-to-fundraise-when-ai-mega-rounds-are-eating-venture-capital-2-infrastructure.webp) AI IT infrastructure and hosting firms absorbed $109.3 billion in VC in 2025, more than two-thirds as much as all other industries combined. That figure is not just large, it is structurally different from prior AI cycles. The cost of training and serving foundation models now tracks closer to semiconductor fab economics than software, and that is what drove the capital concentration toward the infrastructure layer rather than the applications built on top of it. ### Why Infra Funding More Than Doubled The shift was not about more companies entering the market. It was about the same handful needing exponentially more capital to stay in the race, with sovereign wealth funds, hyperscalers, and late-stage growth funds stepping in to supply it. | Segment / Company | Capital Raised | Year | | --- | --- | --- | | AI infra + hosting (global VC) | $47.4B | 2024 | | AI infra + hosting (global VC) | $109.3B | 2025 | | OpenAI + Anthropic combined | $242.6B | through 2026 | | xAI Series E | $20B | 2026 | OpenAI and Anthropic together have raised $242.6 billion in venture funding, a sum that represents roughly 80% of everything the entire Forbes AI 50 list has collectively raised. That concentration is the actual story. Two companies are absorbing the majority of available capital, which means every other AI company, including well-funded infrastructure plays outside the top tier, is competing for what remains. If you are not training a foundation model, those rounds are not your reference class, and treating them as such will distort every conversation you have with investors. ### Benchmarking Application Layer Raises Against Peers The reflex most founders have right now is to use the most recent headline round as their reference point, and those headlines are all infra. That is the wrong frame. Cursor hit a $29.3 billion valuation, a strong outcome by any measure for an application company, and it is still a fraction of what a single foundation-model raise looks like today. The right benchmark is your direct peer group at a comparable stage, not the market’s most capital-intensive outliers. Before your next investor conversation, get specific on these four data points from your actual peer set: - Median pre-money valuation for comparable-stage deals in your vertical over the last 12 months - ARR or usage thresholds your peers hit before closing their rounds - Gross margin benchmarks for application-layer businesses at your stage, not for foundation-model companies - Burn multiple comps from similar businesses, so you can frame your efficiency relative to the right cohort ### Leading with Capital Efficiency over Compute Spend Infrastructure companies justify large burn through compute CAPEX, and investors accept that framing because they understand the cost structure behind it. When Anthropic raised [$5 billion from Amazon](https://news.crunchbase.com/venture/biggest-funding-rounds-ai-autonomy-biotech-anthropic/), that capital was tied to specific infrastructure commitments, not general operating spend. If you are building on top of existing models rather than training them, borrowing the “we need more compute” narrative signals that you have not thought carefully about your own unit economics. Investors who have spent two years evaluating foundation-model deals will spot that mismatch quickly, and it will undermine your credibility rather than support your ask. The application-layer raise that closes today leads with gross margin trajectory, shows inference costs declining as a share of revenue as you scale, and makes a specific case for why the next tranche of capital produces a better return than the last one did. That is the story that positions you against your real peers, not against the outliers eating the billion-dollar checks. [Non-dilutive capital for AI startups](https://qubit.capital/blog/non-dilutive-funding-options-for-ai-startups/), including government grants and strategic accelerator programs, is increasingly relevant for infrastructure builders whose capital intensity leads them to pursue multiple funding tracks alongside their large VC rounds. ## Strategic Anchors Are Driving the Biggest Deals The biggest rounds in 2026 are not assembled from a long list of small checks. GIC anchored [Ramp’s $750 million financing at a $44 billion valuation](https://news.crunchbase.com/venture/biggest-funding-rounds-june-5-2026/), then led Supabase’s $500 million round at a $10.5 billion valuation in the same week, making it the primary validator on two of the market’s largest deals back to back. That kind of repeat anchor behavior is a structural signal about where conviction capital is concentrating, not a coincidence. ### Sovereign and Crossover Funds Leading Rounds The conventional wisdom is that mega-rounds get built by assembling many strong names. The reality in 2026 is a single dominant anchor that sets terms and draws the rest of the table in. Sovereign wealth funds and crossover investors have moved to the front because they can write large checks, follow on at scale, and hold positions through market cycles without the fund-life pressure that forces traditional VCs to push for exits before you are ready. | Company | Round Size | Lead Anchor(s) | Valuation Set | | --- | --- | --- | --- | | Ramp | $750M | GIC, Iconiq, Ontario Teachers’ | $44B | | Supabase | $500M | GIC | $10.5B | | Merge Labs | $250M seed | OpenAI | Undisclosed | ### Lining Up a Credible Lead Investor When OpenAI led Merge Labs’ $250 million seed, the signal was not primarily about scale. It was that the lead was also a natural distribution channel, which pre-answered the question every downstream investor asks: who is the credible buyer of this product. Most founders treat lead investor selection as the final step of a raise, something to close after building momentum with smaller checks. That is backwards. Your lead frames the narrative for everyone who follows, so identifying the fund with both the capital and the strategic reason to want your company to succeed should happen at the start of your process, not as the last piece to slot in. ### Targeting Funds That Write Follow-On Checks The trap most founders miss is closing a round with investors who cannot follow on. A fund that stretches to anchor your seed will not be able to defend its position at Series B, and a cap table where no one maintains conviction tells later investors the story they are most afraid of. What to filter for when building your target list: - **Sovereign and pension funds:** GIC’s lead role on both Ramp and Supabase in the same week confirms these institutions can deploy across multiple large positions without hitting concentration limits. They are structurally built to follow on, which means they stay at the table when you need them most. - **Strategics with acquisition history:** [Google’s $32 billion acquisition of Wiz](https://www.forbes.com/lists/midas/), the largest software takeover on record, and its $2.4 billion deal to bring Windsurf’s founders and IP in-house both started with strategic proximity. The investor who eventually acquires you rarely appears from nowhere. - **Crossover funds bridging private and public markets:** Iconiq’s anchor role in Ramp’s round shows these funds are willing to set terms, not just fill allocations, and hold through the public offering window. That gives founders continuity instead of pressure to sell on someone else’s timeline. Line up a credible anchor first, build the round around that conviction, and treat follow-on capacity as a qualifying criterion rather than a nice-to-have. Spreading a round thin to maximize optionality made sense in a different market. Today, the founders getting the largest deals done are the ones who can point to one investor with the balance sheet and the incentive to stay. [Private equity funding for startups](https://qubit.capital/blog/attract-private-equity-funding-startups/) follows a different return and governance logic than early-stage venture, and recognizing that distinction helps founders understand what strategic anchors are optimizing for when they commit to leading a billion-dollar round. ## The Biggest AI Rounds Right Now Sovereign wealth funds and hyperscalers are anchoring the largest checks on this leaderboard, not traditional venture funds acting alone. The week of June 5, 2026, logged more than a dozen rounds in the multiple hundreds of millions among U.S. startups source, with GIC appearing as lead investor in two separate nine-figure rounds that same week. The pattern across each deal is consistent: the companies raising at this scale have either category-defining infrastructure or a strategic partner whose check size exceeds what the institutional VC market can write on its own. | Company | Raised | Lead Investor | Why It Matters | | --- | --- | --- | --- | | Anthropic | $5B source | Amazon | Amazon had already invested $8B in Anthropic before this round, and committed up to $20B more in future investment beyond the current $5B. This is hyperscaler infrastructure spending at a scale no venture fund can replicate. | | Ramp | $750M | Iconiq, GIC, Ontario Teachers’ | Sets a $44 billion valuation. Sovereign wealth (GIC) and pension capital (Ontario Teachers’) as co-leads signals that the biggest checks in the market are no longer coming from venture funds. | | Supabase | $500M | GIC | $10.5 billion valuation for open-source developer infrastructure. GIC led this round and the Ramp round in the same week, writing two separate nine-figure checks inside seven days. | | Impulse Space | $500M | Not disclosed | Series D brings total investment past $1 billion. Physical-world infrastructure joining the same mega-round tier as software. | | Reliable Robotics | $160M | Nimble Partners | Autonomous aircraft systems funded by a specialist lead. Physical AI attracting institutional capital in the same cycle as the software rounds above it. | | Ray Therapeutics | $125M | Not disclosed | Series B for vision restoration therapies. Biotech clearing nine figures in the same market moment as AI infrastructure, which tells you capital is not flowing into one vertical only. | The range from $125 million to $5 billion on this list reflects three distinct check-writing cohorts operating at the same time: hyperscalers funding AI infrastructure at a scale no VC can match, sovereign wealth funds anchoring growth-stage rounds in the hundreds of millions, and specialist investors backing physical systems and biotech in the $100 million to $200 million band. Most founders will not compete in the top tier. The more useful signal from this table is that the market has not collapsed to one winner-take-all stratum. Companies with a clear infrastructure thesis and a credible anchor investor are still raising meaningful rounds, they just need to know which cohort to target and what that cohort needs to see before writing the check. [Venture capital IRR benchmarks](https://qubit.capital/blog/vc-return-expectations/) explain part of why the largest AI rounds keep attracting capital even at valuations that would have seemed implausible three years ago, since top-quartile returns in this cycle depend heavily on concentrated positions in companies with credible category ownership claims. ## Your Next Move Stop waiting for the market to normalize. The concentration of capital in AI mega rounds is not a blip, and generalist funds that used to write $2M seed checks are now reserving that capacity for follow-ons into their AI bets. Your raise lives or dies on how fast you get specific: which investors are still actively deploying in your stage and sector right now, what your non-AI traction signals look like on paper, and whether your round size fits the check sizes that are actually moving. Founders who are closing deals today built a short, targeted list and moved with urgency. If you want help mapping the right investors and sharpening your pitch before you go out, start with [fundraising assistance](https://qubit.capital/startup-services/fundraising-assistance). ## Key Takeaways - Mega-rounds are concentrating capital into a shrinking pool of category leaders, so right-size your raise to your actual traction rather than anchoring to headline numbers you cannot credibly defend. - If you are building in AI, the fundraising window is unusually open right now and you should move before sentiment shifts or the mega-round cycle cools. - If you are building outside AI, price in a longer raise, tighter terms, and fewer competing term sheets, then plan your runway accordingly before you go out. - Billion-dollar checks are reserved for foundation-model and infrastructure builders, so if you are at the application layer, benchmark your round size and valuation against application-layer peers, not those outliers. - Efficient capital use is your differentiation against infrastructure players burning on compute, so lead with unit economics and capital efficiency in every pitch. - The largest rounds are anchored by strategic, sovereign, or crossover investors who can write follow-on checks, so identify those names early and build relationships before you are actively fundraising. - Assembling a round from many small participants leaves you without a credible lead and signals to the market that no one is convicted, so prioritize landing one strong anchor over collecting small checks. - Category leadership is the threshold that unlocks mega-round access, so if you cannot make a clear case for that position, structure a smaller, clean round that you can close fast rather than running a drawn-out process chasing a size that will not come.