Investors continue to increase their bets on artificial intelligence, but the focus is shifting from applications to "hard" infrastructure: energy, specialized chips, data centers, and cybersecurity. At the same time, the volatility of recent IPOs is prompting funds to reevaluate their exit strategies in favor of M&A and secondary transactions. Below are the key events and trends shaping the agenda of the venture market as of Wednesday, August 5, 2026.
Main Deal: Nvidia Invests $5 Billion in Safe Superintelligence
A central event of recent days is the strategic partnership between Nvidia and Safe Superintelligence (SSI) — the lab founded by Ilya Sutskever, co-founder of OpenAI. According to sources familiar with the terms of the deal, the chipmaker's investment amounts to around $5 billion — one of Nvidia's largest bets during the entire AI boom.
The deal's details are impressive even within the context of a heated market:
- SSI will receive priority access to the Vera Rubin computing platform — Nvidia's latest architecture;
- the startup's computing capacity is expected to increase tenfold over the next 12 months;
- total funding for SSI has reached approximately $7 billion with a valuation of around $32 billion;
- the company still has no commercial products and publicly states that it does not plan to release intermediate models until achieving its primary goal.
The deal underscores a new market logic: the largest tech corporations are willing to pay billions not for revenue, but for access to cutting-edge research and talent. For the venture funds that had previously invested in SSI — among them Andreessen Horowitz, Sequoia, Lightspeed, and Greenoaks — the partnership with Nvidia serves as a powerful validation of their positions.
Record-Half Year: $510 Billion and Unprecedented Capital Concentration
Statistics for the first half of 2026 shattered all historical records. The global volume of venture investments reached $510 billion — approximately 36% above the previous record set in the second half of 2021. The first quarter brought in $305 billion, making it the largest quarter in the industry’s history; the second added another $205 billion, distributed among over five thousand startups.
However, behind these impressive figures lies a market structure that is alarming for allocators:
- about 43% of all capital for the half-year went to just two companies — OpenAI and Anthropic;
- U.S. startups accounted for nearly 80% of global funding from seed to late stages — a sharp contrast to the pre-AI era, when the U.S. share did not exceed half;
- in the artificial intelligence segment, the concentration is even higher: about 88% of AI investments, or roughly $319 billion, went to companies based in the U.S.;
- the five largest managers collected over 73% of all venture commitments, while the top 15 firms accounted for nearly 89%.
Analysts caution that the venture asset class increasingly resembles public indices, where returns are defined by a narrow group of mega-capitalizations. For institutional investors, this poses a risk of hidden exposure duplication when investing in multiple large funds simultaneously.
IPO Market: A Record Year with a Bitter Aftertaste
The IPO market in 2026 is formally experiencing a renaissance: 44 venture company IPOs have already occurred in the U.S. — compared to 50 for the entire previous year. The culmination was the June debut of SpaceX, which achieved a historical valuation of around $1.77 trillion, followed by listings for Cerebras, Quantinuum, X-Energy, and HawkEye 360.
However, the post-debut dynamics have tempered enthusiasm. SpaceX's shares fell approximately 30% below the offering price within six weeks, while Cerebras lost up to 35%. Consequences soon followed:
- OpenAI postponed its plans for a public offering to 2027;
- Databricks completely excluded itself from the listing queue — the CEO called 2026 a "terrible year for IPOs" due to the overloaded calendar of mega-listings;
- late-stage investors are increasingly using secondary transactions and structured liquidity instead of waiting for IPOs.
A curious counter-trend is being set by Robinhood: the broker is bringing a second venture fund to market with a volume of up to $200 million, offering retail investors access to early-stage private companies through a publicly-listed structure. The IPO is slated for mid-August — a signal that the democratization of the venture asset class continues, independent of sentiment in the traditional IPO segment.
Where the Money Goes: AI Infrastructure Instead of Applications
Recent funding rounds demonstrate a clear shift of capital towards the physical infrastructure of the AI economy. Investors are financing the "bottlenecks" of the boom — energy, computing, and security:
- Valar Atomics secured $1 billion in a Series B round at a valuation of $6 billion for the serial production of modular nuclear reactors for data centers;
- Commonwealth Fusion Systems received $1 billion for the construction of an industrial-scale fusion power plant, bringing total funding to $4 billion;
- Antora Energy closed a $550 million C round for thermal energy storage systems for data centers;
- K2 Space raised $500 million for the production of powerful satellites;
- British developer of photonic chips for AI inference OLIX raised around $312 million at a valuation of $3.3 billion;
- Horizon3.ai received $250 million for autonomous cybersecurity testing tools.
The logic of investors is clear: while the outcome of competition among AI applications remains uncertain, suppliers of "shovels and picks" — energy, computing, and security — win under any scenario.
Consolidation and M&A: Strategists Reshape the Landscape
Against the backdrop of a narrowing IPO window, mergers and acquisitions are becoming the main channel for liquidity. The first half of the year has already seen landmark deals: Qualcomm acquired AI chip developer Modular for approximately $4 billion, Salesforce absorbed client AI solutions provider Fin, and the acquisition of Cursor has gone down in history as the largest acquisition of a venture company.
Corporate venture arms are also changing tactics: instead of a broad portfolio of small bets, they are focusing on a smaller number of substantial investments in AI startups, viewing them as a way to gain priority access to computing capabilities and technologies. For early-stage funds, this widens the map of potential buyers for portfolio companies.
Discipline Amid Abundance: How Funds Manage Their "Dry Powder"
Despite record levels of available capital, there is no talk of easy money. Managers describe the current market as selective: subsequent rounds are going to teams with clean metrics, clear unit economics, and a coherent path to exit. Valuations are rising rapidly only for category leaders — primarily in AI and late stages — while the rest of the market is undergoing a rigorous test of resilience.
It is telling that the record exit environment does not help small and new venture firms: institutional money continues to flow toward the largest brands in the industry, complicating fundraising for first and second-time fund managers.
Russia and the CIS: Cautious Recovery from a Low Base
The Russian venture market is moving in its own logic. By the end of 2025, its volume was approximately $159 million from 102 deals; however, the average check grew by two-thirds to $1.7 million. Projections for 2026 expect growth of 10–15%, gradually recovering to around 17 billion rubles.
Key drivers are private and state funds, while business angel activity is hampered by high key rates and competition from bonds. Among notable initiatives is the launch of the first specialized fund for AI agent-based projects in the country, as well as a busy calendar of industry events: by mid-August, Moscow is set to host the anniversary forum "Venture Landscape," which will gather key players from the local ecosystem.
Looking Ahead: What It Means for Investors
The venture market enters the second half of 2026 with a unique combination of factors: unlimited private capital, record concentration, a cooling public window, and an increasing role of M&A. For funds and allocators, there are three practical conclusions. First, diversification beyond consensus mega-deals is becoming a source of alpha — competition for quality assets is significantly lower in less efficient market segments. Second, liquidity strategies require reevaluation: the secondary market and sales to strategics are displacing IPOs as the baseline exit scenario. Third, a bet on AI infrastructure — energy, chips, cybersecurity — appears most resilient to potential corrections in application segment valuations. The market remains generous but rewards discipline rather than a mere appetite for risk.
