Startup and Venture Capital News — Wednesday, August 26, 2026: Anthropic Prepares for IPO, Nvidia Acquires Complete AI Stack, and Physical AI Emerges as New Megatrend

Current updates in startups and venture capital as of August 26, 2026: Anthropic's preparation for a record-breaking IPO, Nvidia's strategic expansion, mega-rounds in defense technology and robotics, as well as key trends in the venture market for funds and institutional investors.

Venture capital is approaching the end of August 2026 in a state that is difficult to describe with a single word. On one hand, there are historical records: the global amount of venture investments in the first half of the year reached $510 billion, surpassing the figure for the entire year of 2025 ($440 billion) and exceeding the previous peak of $375 billion set in the second half of 2021. On the other hand, there is an unprecedented concentration of capital: OpenAI and Anthropic accounted for $217 billion, or 43% of all venture capital invested in startups worldwide.

For venture investors and funds, this means that traditional benchmarks are no longer effective. The average round size is skewed by deals that most LPs will never receive allocations in, and “normalcy” in the market has to be measured outside the top ten mega-rounds. Below are the key events and trends shaping the venture market agenda for Wednesday, August 26, 2026.

Headline of the Day: Anthropic on the Brink of the Largest IPO in Tech Sector History

The central event of the week is Anthropic’s preparation for public disclosure of IPO documents. The company, which confidentially submitted a registration statement to the SEC back in June, is ready to publish its prospectus by the end of August. The target for the offering is at or above the record IPO of SpaceX, which in June raised approximately $75 billion ($85.7 billion including underwriter options) at a valuation of $1.77 trillion.

What is important for venture investors in this deal:

• Valuation. The last private round, Series H at $65 billion, established a post-money valuation of about $965 billion. Market expectations for the IPO range from $1 trillion to $2 trillion.
• First disclosure of frontier lab economics. The prospectus will show the market for the first time the revenue structure, growth rates, business segmentation, and critically, the actual cost of inference.
• Risk factors. Sources indicate that key risks will include growing public discontent regarding AI and data center construction, as well as concerns about AI's impact on employment.
• Governance structure. The status of a public benefit corporation and Long-Term Benefit Trust with the right to appoint an increasing share of the board of directors is a topic that will spark discussions among institutional buyers.
• Margin. Gross margin forecasts have been revised down from about 50% to 40% amid unexpectedly high computing costs.

An additional backdrop is provided by Nvidia's quarterly results, which are expected to be published on August 26. For the entire AI startup ecosystem, this is a major macro-indicator of the resilience of the infrastructure cycle.

Nvidia Builds a Vertical: From Chips to Models, Applications, and Talent

In recent days, Nvidia has demonstrated how the largest beneficiary of the AI boom converts cash flow into control over the entire stack. The company is discussing an investment in Perplexity as part of a round that could value the AI search startup at more than $30 billion—compared to approximately $20 billion a year earlier. Perplexity's annual revenue reportedly grew from less than $250 million at the beginning of 2026 to more than $750 million.

At the same time, Nvidia has entered into an agreement with Poolside worth around $6 billion, including about $1 billion in direct investment, access to the startup's technology, and the transfer of more than 100 engineers to the Nemotron project. The goal is to create a competitive American alternative to Chinese open-weight models.

For venture funds, this presents a new structural risk: strategic investors with a balance of such scale simultaneously serve as suppliers, shareholders, and potential competitors to portfolio companies. Classic licensing-investment-hiring structures are increasingly being replaced by full acquisitions, which directly impacts exit scenarios.

Physical AI and Robotics: A New Category of Mega-Rounds

The robotics division of Chinese automaker XPeng raised over $900 million in its first external round at a valuation of over $6.3 billion. The round was led by IDG Capital and Gaorong Ventures, with strategic investors including Tencent and Alibaba. The funds will be used for the development of humanoid robots, mass production, and physical AI models.

A notable context: at the World Humanoid Robot Games in Beijing, two Chinese machines completed the 100-meter dash faster than Usain Bolt's record—9.39 and 9.47 seconds compared to 9.58 seconds. A year earlier, the same platform recorded a time of 21.5 seconds.

Takeaways for investment committees:

1. Physical AI has transitioned from the category of demonstrations to that of capital-intensive industrial bets.
2. Automakers gain a structural advantage over pure robotics startups through the reuse of chips, perceptual systems, and manufacturing capabilities.
3. Chinese tech giants are aggressively positioning themselves in embodied AI as the next computing platform.

Biggest Rounds of the Week: Defense, Inference, and Infrastructure

The list of the largest American deals for the week confirms the capital shift towards “hard” sectors:

• Castelion — $800 million (plus $250 million in debt financing), defense technologies, hypersonic strike missile. The round was led by JPMorgan Chase, Andreessen Horowitz, and Carlyle, with a valuation of $13 billion.
• Etched — $700 million, semiconductors for accelerating inference, valuation $21 billion, lead investor Jane Street.
• Higgsfield — $400 million, AI video generation platform, valuation $5.4 billion, led by DST Global.
• Groq — $350 million, data centers, valuation $3.5 billion, with Nvidia’s participation.
• Wispr Flow — $280 million, voice AI interfaces, valuation $2 billion, led by Menlo Ventures.
• Muon Space — $250 million, satellite constellations, led by Eclipse.

Finishing off the top ten are Also ($150 million, micromobility), Velaura AI ($110 million, ultra-low-power computing), Rillet ($100 million, agentic finance, valuation $1 billion), and Happy Health ($75 million, sleep apnea diagnostics).

Europe: Steady Deal Flow with No Mega-Rounds

Over the past week in Europe, more than 45 deals were registered with a total volume exceeding €684 million. The leading sectors were fintech (€239.2 million), robotics (€178 million), and artificial intelligence (€99 million). By country, Switzerland topped the list (€172.5 million), followed by France (€150 million) and the United Kingdom (€122.9 million).

The largest deals included SoftBank's investment of $200 million in Swiss Gravis Robotics, €150 million raised by French Ingenico, Rillet's $100 million round at a valuation of $1 billion, and a seed round for British Callosum at $100 million—a rare example of nine-digit seed funding.

In the context of the first half of the year: European tech companies raised €44.1 billion across 1,740 deals, with the UK accounting for €18.7 billion, while AI startups attracted €5.92 billion. There were 252 exits recorded.

M&A Market and Liquidity: Power Infrastructure as a New Asset

The exit channel remains open but is increasingly shifting towards infrastructure assets. nVent has acquired Maverick Power for $1.75 billion with a potential earnout of up to $550 million upon reaching performance targets in 2027-2028. Infineon has absorbed Indian C2i Semiconductors, specializing in power management in AI data centers.

A separate story is Hugging Face, which is exploring the possibility of a sale at a valuation of $13 billion or more, having hired a bank to assess buyer interest. The company's last major valuation in 2023 was around $4.5 billion.

In the second quarter of 2026, the exit market set records: 32 companies went public with valuations above $1 billion and 24 were acquired for at least $1 billion, totaling $113 billion. For LPs, this means distributions are finally returning, fueling a new fundraising cycle for venture funds.

Market Structure: Record Without Breadth

A key analytical takeaway for investors: record absolute numbers mask the bifurcation of the market. Excluding the four largest deals—OpenAI, Anthropic, xAI, and Waymo—activity in the rest of the market is close to levels seen in 2024-2025.

Additional structural observations:

• Over 70% of the capital in the second quarter went to AI companies compared to less than 50% a year earlier.
• 16 companies raised rounds exceeding $1 billion for a total of $108.6 billion—53% of the quarterly volume.
• Late-stage funding grew by 141% year on year: capital is concentrating in already proven winners.
• In the first half of the year, 195 companies joined the unicorn list—the highest since the second half of 2022.
• The share of the U.S. in the global volume dropped from 83% in the first quarter to 67% in the second.

Local Context: Russia and Markets with Limited Access to Capital

Against the backdrop of a global boom, the Russian venture market is moving in the opposite direction. Industry research estimates that the market volume in the first half of 2026 shrank by nearly 48% year on year, to 4.6 billion rubles. The share of foreign investment has virtually vanished, with Moscow accounting for about 64% of the volume and 63% of the number of deals.

The structure of the market has also changed: seed rounds make up 62% of deals but only 8% of the volume, whereas late-stage rounds account for 8% of deals and 43% of all investments. The most significant decrease was seen among private investors—a decline of 59% in the number of deals. For global funds, this illustrates how quickly local ecosystems lose touch with the international flow of capital in the absence of exit channels.

What This Means for Venture Funds and Investors

The agenda for August 26, 2026, shapes several practical takeaways for capital managers:

1. The Anthropic IPO will be the primary test for valuations in the AI sector. The public response to the prospectus will set a benchmark for the entire private AI universe—from frontier labs to applied startups.
2. The infrastructure layer continues to absorb capital. Energy, power distribution, cooling, and inference chips are segments with the most predictable unit economics in the current cycle.
3. Strategic investors are changing the game. The presence of Nvidia, Alibaba, Tencent, and hyperscalers in cap tables necessitates a reevaluation of approaches to protecting minority positions.
4. Defense technologies and physical AI are resilient categories of mega-rounds. Geopolitics have transformed them from niche topics into the mainstream of the venture portfolio.
5. The exit window is open, but selectively. Record IPOs and M&A are concentrated in the upper segment; the median portfolio company still requires proven revenue.

The market has entered a phase where record venture investment volumes coexist with stringent selectivity. Capital is available—but primarily to those who control technically, legally, or physically replicable layers of the AI economy.