Inside AI's $510 Billion Half and the Gulf's $49 Billion Bet

Six months. Five hundred and ten billion dollars. That is the headline number Crunchbase put on global startup funding for the first half of 2026, and it is not a rounding error or a one-off spike tied to a single mega-deal. It is the largest half-year of venture investment ever recorded, built quarter after quarter on the back of one theme: artificial intelligence. Investors did not spread that money evenly. A handful of labs, a handful of infrastructure operators, and a growing list of sovereign wealth funds from the Gulf absorbed a share of the total that would have looked implausible even eighteen months ago. This is a story about concentration. Capital is not just flowing toward AI as a category, it is pooling around specific companies and specific state-backed vehicles with the balance sheets to write nine and ten-figure checks. Understanding where that money actually went, and who is now positioned to shape the next phase of the industry, matters more than the topline record itself. Riyadh, Abu Dhabi and Doha are no longer bystanders watching Silicon Valley write the checks. They are writing some of the largest ones themselves, and doing it with distinct strategies that will shape who controls compute, models and data over the next decade. What follows is a walk through the numbers behind that half-year, quarter by quarter, deal by deal, and then a close look at three Gulf funds that have each chosen a different way to bet on the same industry. Read together, the figures explain both why the AI market has grown so top-heavy and why sovereign investors are treating that top-heaviness as an opening rather than a warning sign.

Start with the quarter that set the tone. Crunchbase's original Q1 2026 report counted $300 billion invested into roughly 6,000 startups worldwide, with $242 billion, or 80% of that total, going to AI companies alone. That AI share dwarfed the prior record of 55% set just a year earlier in Q1 2025. Analysts at Crunchbase called it an all-time high not approached by any other quarter on record, and the number alone exceeded full-year global venture totals from every year before 2018. Four of the five largest venture rounds ever closed inside those three months: OpenAI raised $122 billion, Anthropic raised $30 billion, xAI raised $20 billion and Waymo raised $16 billion. Combined, those four rounds alone equaled $188 billion, or 65% of everything invested globally that quarter. When Crunchbase later folded Q1 into its half-year accounting, the figure was revised slightly upward to $305 billion, a normal adjustment as later-reported deals get added to the dataset. Either number tells the same story. A market once known for spreading bets across hundreds of promising startups had narrowed, in the space of one quarter, to a contest among a handful of frontier labs and the investors able to back them at scale.

A slowdown after a record quarter would have been unremarkable. Instead, Q2 2026 closed at $205 billion, the second-largest quarter for global startup funding ever recorded, trailing only the quarter before it. Crunchbase's H1 report noted that AI-focused companies captured more than 70% of all global startup capital during the quarter, and that late-stage AI rounds alone rose 141% year over year compared with Q2 2025. Sixteen companies closed billion-dollar mega-rounds during the quarter, a combined $108.6 billion that made up 53% of everything invested in Q2. Put the two quarters together and the picture is a market running at a pace no prior year came close to matching, with AI investment functioning less like a sector rotation and more like a structural reallocation of where venture capital goes by default. For founders outside frontier AI labs, the practical effect has been brutal: capital that used to fund a wide spread of consumer, enterprise and fintech startups is now heavily weighted toward a short list of compute-intensive, foundation-model-adjacent companies.

Zoom out to the full half-year and the concentration gets starker. OpenAI and Anthropic together raised a combined $217 billion in H1 2026, according to Crunchbase, which is 43% of every dollar invested in startups globally across every sector, every geography and every stage. No other two companies in venture history have absorbed anywhere near that share of a six-month period. It helps explain why so many later-stage investors describe the current market less as a portfolio strategy and more as a queue: funds wait for allocation into a small set of names, and everything else competes for what is left. The concentration is not simply about model quality or product traction, though both companies have plenty of that. It reflects the sheer capital intensity of frontier AI, where training and serving the largest models requires data center capacity, chips and power contracts that only a handful of companies can plausibly commit to at scale. Investors who might once have backed twenty mid-sized AI startups are instead writing fewer, larger checks into the companies they believe can actually secure the compute to compete.

Geography tells a similarly narrow story. In Q1 2026, U.S.-based companies took in $250 billion, or 83% of global startup funding, while China accounted for $16.1 billion and the United Kingdom for $7.4 billion, with everyone else splitting a remainder of roughly $26 billion. By Q2, the U.S. share eased slightly to about two-thirds of global capital as some late-stage rounds closed elsewhere, but the broader pattern held: a small number of markets with deep capital pools and frontier AI labs captured most of the money, while founders in most other regions competed for a shrinking residual. That residual is exactly where Gulf capital has started to matter in a different way than most people assume. The Gulf is not primarily competing for a slice of that shrinking residual pool of startup equity. It is positioning itself on the other side of the table, as a source of the capital being deployed into the concentrated winners, and increasingly as an owner of the infrastructure those winners depend on.

The single deal that best captures the quarter closed on May 29, 2026, when Anthropic announced a $65 billion Series H round at a $965 billion post-money valuation, according to the company's own newsroom. Singapore's GIC co-led the round, with Temasek joining as a significant backer, marking one of the largest bets either fund has placed on a single private company. The round pushed Anthropic's valuation past OpenAI's for the first time, a symbolic shift in an industry that had treated OpenAI's lead as close to unquestionable for years. Coverage of the round noted that no major European public funds participated, underscoring how the capital behind frontier AI has increasingly clustered around U.S. tech investors, Gulf sovereign vehicles and a small set of Asian state funds rather than a broad international base. For Anthropic, the round is less about needing cash today and more about locking in the compute contracts, chip supply and data center capacity that will determine which lab can actually train its next generation of models on schedule. That is the same calculation driving Gulf sovereign funds toward AI, just from the other side of the check. It also raises a question boards at every major lab are quietly asking their bankers this year: once a private company is valued near a trillion dollars, does staying private still make sense, or does the exit-market boom described below start to look more attractive than another funding round.

Fresh capital going in is only half the story. Q2 2026 also delivered a record exit market, with 32 venture-backed companies going public above a $1 billion valuation, according to Crunchbase. The standout was SpaceX, which listed at a $1.77 trillion valuation and raised $75 billion in its IPO, instantly becoming one of the largest public listings in history. On the acquisition side, 24 companies were bought for $1 billion or more during the quarter, a combined $113 billion in deal value that Crunchbase flagged as a record quarter for M&A. The headline transaction there was SpaceX's own acquisition of Anysphere, the maker of the Cursor coding assistant, for $60 billion, a deal that folded one of the fastest-growing AI coding startups directly into a company that had just become a public markets giant. Together, the fundraising and exit numbers describe a market where capital is not just piling into private AI companies, it is also finding faster and larger paths back out, which in turn justifies the size of the checks investors are willing to write on the way in.

On July 1, 2026, Abu Dhabi's MGX closed its first fund at $49 billion, roughly $4 billion above its original target, in what CNBC and Bloomberg both described as one of the largest AI-dedicated funds ever raised. MGX is chaired by Sheikh Tahnoon bin Zayed, who also chairs Mubadala and G42, giving the fund direct lines into the UAE's broader technology and sovereign investment apparatus. The fund is targeting more than $100 billion in total assets over time and expects to deploy up to $10 billion a year. What sets MGX apart from a typical infrastructure fund is the breadth of its holdings. It has taken equity stakes in OpenAI, Anthropic and xAI simultaneously, giving it a position across three competing frontier labs rather than a bet on any single winner. It also owns a $40 billion data center operator, holds a stake in TikTok's American entity through the USDS joint venture with Oracle and Silver Lake, and made a $2 billion investment in Binance settled in the USD1 stablecoin. Analysts describe this as a full-stack aggregation strategy, one that spans labs, infrastructure and platform access rather than picking a single layer of the AI stack. It is a strategy only a handful of investors in the world could fund at this scale, and Abu Dhabi's balance sheet, with S&P putting its net asset position at 358% of GDP, is one of them.

Saudi Arabia's approach through HUMAIN, the AI vehicle backed by the Public Investment Fund, looks almost nothing like MGX's. Where MGX buys equity across labs and platforms, HUMAIN has signed roughly $23 billion in agreements with Nvidia, AMD, Amazon Web Services and Qualcomm, and struck a dedicated $10 billion joint venture with AMD for 500 megawatts of domestic compute capacity. The kingdom is targeting 1.9 gigawatts of AI-dedicated data center capacity by 2030, with plans to scale toward 6.6 gigawatts after that, all built inside Saudi Arabia rather than acquired as stakes in foreign companies. Aramco has taken an equity position tying part of the kingdom's oil balance sheet directly to its compute ambitions, a decision that links the country's traditional revenue base to its AI infrastructure buildout in a way few other national strategies attempt. HUMAIN has been explicit that it is not trying to acquire foundation model companies. Its bet is vertical integration at home: chips, power and data centers built domestically, positioning Saudi Arabia as a compute exporter to the region rather than a shareholder in labs it does not control. It is a slower, more capital-intensive path than MGX's, but one that keeps strategic assets inside Saudi borders.

The Qatar Investment Authority has taken a third path. In December 2025 it formed a $20 billion strategic partnership with Brookfield to build AI infrastructure inside Qatar and in select international markets. Like HUMAIN, Qai has said explicitly it will not build foundation models. Unlike HUMAIN, its infrastructure bet is not confined to domestic soil, and unlike MGX, it holds no direct equity in OpenAI, Anthropic or xAI. Qai's wager is that once model competition commoditizes, the infrastructure layer holding data centers, power and networking will retain durable value regardless of which lab wins the model race. Line up all three strategies side by side and a pattern emerges that says something about how each fund defines risk. MGX is betting on picking winners across the whole stack. HUMAIN is betting on building sovereign capacity that never depends on a foreign lab's decisions. Qai is betting on the layer beneath the competition, the one that gets used no matter who wins it. None of the three amounts are trivial next to the $217 billion OpenAI and Anthropic raised in H1 alone, but together they represent close to $92 billion in committed Gulf capital aimed squarely at AI, a figure that would have ranked among the largest single-country venture markets on earth just two years ago.

The record numbers have not silenced the skeptics. Morgan Stanley's Institute research has flagged the pace of AI infrastructure spending as a genuine risk factor for 2026, not just an opportunity, pointing to the gap between capital committed and revenue actually generated by many AI deployments. Fortune reported in late June that big technology companies are pouring trillions into infrastructure even as questions persist about whether current spending levels can be justified by near-term returns. Commentary elsewhere has used the word bubble directly, drawing comparisons to prior cycles where infrastructure buildout ran well ahead of proven demand. None of that skepticism has slowed the checks being written. What it has done is sharpen the case for why sovereign capital behaves differently than typical venture funds. A fund managing a country's oil wealth over decades can absorb a longer payback period and a higher tolerance for concentration risk than a ten-year venture fund answering to limited partners who expect returns on a fixed timeline. That patience is precisely why Gulf funds have been able to move so aggressively while others hesitate, and it is also why a correction in AI valuations, if one comes, would land very differently on a sovereign balance sheet than on a typical late-stage venture portfolio.

Three separate readings of the same half-year data lead to one conclusion for anyone tracking the Gulf's role in AI. Startup capital globally has concentrated around a small set of frontier labs at a pace that has no historical precedent. Exit markets have opened wide enough to justify the size of the checks going in, at least for now. And Gulf sovereign funds have stopped treating AI as one allocation among many, choosing instead to build entire strategies around it, each fund picking a different layer of the stack to own. For businesses and policymakers across the region, the practical question for the second half of 2026 is not whether AI investment will keep growing. It almost certainly will. The question is which of the three Gulf models, MGX's cross-lab equity approach, HUMAIN's domestic vertical build, or Qai's infrastructure-only bet, proves most resilient if the funding pace that produced $510 billion in six months cannot be sustained at the same intensity. Each fund is making a different wager on that question, and each has committed real capital to being right.

Crunchbase Data: Global Startup Investment Hit Record $510B In H1 2026

Q1 2026 Shatters Venture Funding Records As AI Boom Pushes Startup Investment To $300B

Anthropic raises $65B in Series H funding at $965B post-money valuation

GIC, Temasek anchor Singapore's investment in Anthropic's $65B Series H round

OpenAI, Anthropic backer MGX raises $49 billion AI fund

UAE: Abu Dhabi's MGX Raises $49 Billion for One of the Biggest AI Funds Ever

Abu Dhabi's $49 Billion AI Fund And Its Sovereign Rivals

AI Market Trends 2026: Global Investment, Risks, and Buildout

AI spending boom accelerates as Big Tech pours trillions into infrastructure