From January to June 2025, global investment in AI startups far exceeded that in the first half of 2024. In just the first quarter of 2025, approximately 60 to 73 billion dollars were attracted, exceeding half of the total amount for all of 2024, with a year-on-year growth of over 100%. In the first quarter, AI companies received about 58% of the total venture capital, compared to about 28% a year earlier. This clearly indicates the investors' 'AI FOMO' mentality.
This means: capital is concentrating in the AI field on an unprecedented scale, with major institutions doubling down on those believed to be the winners in AI, which could reshape the funding allocation landscape in the second half of the year.
During this period, the super-sized late-stage financing rounds led by leading companies are particularly notable. In March, OpenAI raised 40 billion dollars (the largest private financing round in history), achieving a valuation of 300 billion dollars, while Anthropic's 3.5 billion dollars in E round financing raised its valuation to 61.5 billion dollars. Other deals, such as Safe Superintelligence's 2 billion dollar financing and Neuralink's 650 million dollar E round financing, further increased the total.
This means: a 'winner-takes-all' situation is concentrating most of the funds into very few enterprises, squeezing out funds that could have flowed to earlier stages or smaller companies.
In addition to those highly publicized mega-financings, medium-sized transactions surged, while seed funding activity remained selective. The median seed round financing in the AI sector reached about 15 million dollars (average about 41 million dollars), and the median A round financing was about 75 to 80 million dollars, both well above historical averages (the median A round financing across all industries globally in 2022 was about 10 million dollars). The median financing for growth-stage rounds C and D concentrated between 250 million and 300 million dollars, with the average being pushed up by extreme cases like OpenAI.
This means: the expansion of transaction volume reflects fierce competition among industry leaders. Investors unable to write nine-figure checks may turn to niche areas or earlier-stage investments, while any startup claiming to have an AI narrative can secure larger-scale financing and higher valuations.
Participants in the generative AI and core model/infrastructure space attracted over 45 billion dollars in funding in the first half of the year, accounting for more than 95% of the total disclosed funding. Application-oriented AI verticals are relatively underfunded (about 700 million dollars in healthcare/biotech; about 2 to 3 billion dollars in fintech/enterprise). Geographically, the U.S. (especially Silicon Valley) dominates: more than 99% of the global AI funding in the first half of the year flowed to companies based in the U.S. Asia and Europe lagged behind, with China's largest deal (Zhizhu AI) raising 247 million dollars; while Europe only saw some medium-sized financing rounds (for example, UK's Latent Labs raised 50 million dollars).
This means: this wave is centered in the U.S., led by a few large companies; governments and investors outside the U.S. are expected to respond in the second half of the year by establishing national AI funds, providing incentives, or making cross-border investments to avoid falling behind.
Despite record capital investments, investors' caution is returning. Many financing rounds in the first half of the year focused on strategic or corporate investors (cloud service providers, chip manufacturers, defense companies), indicating that investors prefer projects with practical application scenarios and strategic synergies. As we enter the second half of the year, investors will closely monitor the performance of startups that have received large amounts of funding in product delivery, revenue, and regulatory responses, especially in the context of intensifying competition.
This means: in the second half, capital may favor those companies that demonstrate efficiency and real market attractiveness — especially 'tools and shovels' suppliers (tools, chips, enterprise software), which will raise the entry barrier for new entrants, consolidate existing companies' advantages, while posing challenges to newcomers.
The first half of 2025 is a critical moment for AI investment. The substantial influx of capital into the AI sector (and its tilt towards a few participants and regions) will shape the innovation landscape and competitive dynamics for years to come. For investors, understanding the flow of funds and the reasons behind it will be crucial for navigating the second half of 2025. Will winners be able to justify their valuations, or will there be a pullback and refocusing? The data from the first half of the year provides early clues for portfolio strategy, policy considerations (such as antitrust and national security issues), and the financing prospects for founders in the coming half year.
The most noteworthy financing in the AI sector over the past month.
1. Financing momentum: unprecedented surge year-on-year.
In the first half of 2025, venture capital investment in AI startups far exceeded the levels of the same period in 2024. Reliable data shows that approximately 70 billion dollars flowed into AI companies in just the first quarter, exceeding half of the total financing amount for the entire AI sector in 2024. This means that the financing amount in the first half of 2025 is more than double that of the first half of 2024 (in dollars).
In the first quarter of 2025, AI's share of global venture capital surged to about 53% to 58%, compared to about 25% to 30% a year ago. This means that currently more than half of global venture capital is directed towards the AI sector.
Drivers: a few mega-financings; without them, global venture capital funding year-on-year is roughly flat.
Impact on the second half of 2025: overall venture capital indicators may depend on the transaction flow in the AI field; any cooling of enthusiasm in the AI sector could pull down overall financing levels.
2. Financing stages: late-stage financing has surged, while early-stage financing is uneven.
Data shows that the transaction volume in the AI sector has a barbell distribution.
Late stage (C+ rounds) dominates: in the first quarter of 2025, the total amount of late-stage financing across all industries reached 81 billion dollars, a year-on-year increase of about 147%, with AI as the main driver.
Early stage: the number of deals has decreased (global early-stage deals dropped about 19% year-on-year), but the financing scale has significantly increased.
Key point: investors will put money into fewer, higher-stakes projects — confident in specific AI themes while remaining cautious in other areas. This polarization is expected to continue in the second half of the year.
3. Industry configuration: core models and infrastructure development.
About 95% of AI funding is chasing generative AI model developers and their infrastructure (cloud computing, chips, development platforms). Just OpenAI and Anthropic attracted about 60% of the AI sector's funding in the first half of the year.
In contrast, vertical application areas are negligible.
Investor logic: control the 'AI stack'; vertical applications may become commoditized (note: unique values such as branding that commodities originally possess may disappear due to sufficient market competition) or face longer GTM cycles.
4. Regional distribution: concentrated in the U.S., with the Bay Area accounting for half of the financing amount.
In the first quarter, 71% to 73% of global venture capital flowed into North America; by value, the concentration of funds in the AI sector is about 99% in the U.S. The San Francisco Bay Area (including OpenAI) alone accounts for nearly half of global venture capital.
Europe, the Middle East, and Africa: only a few medium-sized AI transactions (Latent Labs raised 50 million dollars, Speedata raised 44 million dollars).
Asia-Pacific: In the first quarter of 2025, only 1.8 billion dollars were raised for AI (a year-on-year decline of 50%); China's largest round was Zhizhu AI's 247 million dollars.
In conclusion: the U.S. holds an advantage in funding in this 'AI arms race'.
5. Investor landscape:
Sovereign wealth funds and cross-border funds (Saudi's Prosperity7, Malaysia's Khazanah, Thrive Capital) led multiple rounds of financing.
Corporate venture capital departments of large tech companies (Microsoft, Salesforce, Google) are very active.
Net effect: capital is flooding in from all sides.
Regulatory milestones.
Governments are still exploring how to respond to AI. In the EU, the AI Act is expected to be finalized by the end of 2025. In the second half of the year, startups are expected to engage in lobbying battles, and early compliance signals may emerge. In the U.S., any movement from the executive order on AI and Congress — hearings, proposed legislation — will be crucial. New regulations around data usage, model transparency, or chip export controls may reshape the economic conditions for startups and investor confidence.
Additionally, attention should be paid to AI procurement by the U.S. government — rumors of a multi-billion dollar plan could provide important demand signals for AI companies focused on enterprises.
IPO channels and exit routes.
Despite a surge in private financing in 2025, no groundbreaking AI IPOs have been seen yet. This situation may change in the second half of the year. Companies like Databricks, Stripe (AI-related), or even OpenAI could be potential IPO candidates.
Meanwhile, M&A activity may escalate. Large tech companies may take action: Google, Microsoft, or Nvidia may acquire smaller AI teams or core infrastructure providers. A significant AI acquisition could reshape the competitive landscape and deliver returns for venture capital firms.
Technological breakthroughs and product releases.
Expecting major news disclosure: possibly OpenAI's next-generation model, or hardware launched in collaboration by Sam Altman and Jony Ive.
Any significant breakthroughs in capabilities (e.g., models capable of reasoning or models that reduce costs tenfold) could validate overvaluation and trigger a new wave of capital.
Attention should also be paid to enterprise-level attractiveness — API sales, SaaS adoption, and revenue. But risks exist; if a security incident or public misuse occurs, it may provoke strong regulatory backlash, dampening market sentiment.
In summary, the technical and business performance in the second half of the year will determine whether the optimism of the first half can continue.
Regulatory and ethical resistance.
If the government or public feel that AI has gone out of control, rapid intervention measures are expected: such as implementing licensing systems, imposing fines based on the General Data Protection Regulation (GDPR), or imposing strict restrictions on certain models.
Ethical resistance: scandals, mass layoffs caused by automation, or AI-generated misinformation may quickly change market sentiment, making it harder for funds to be deployed.
Computing and talent constraints.
The lifeblood of AI — graphic processing units (GPUs) and elite engineers — remains scarce.
GPU bottlenecks may force cash-strapped teams to exit, while well-capitalized companies will hoard computing resources.
The talent war is intensifying, with OpenAI and Google recruiting top talent.
The burn rate is skyrocketing: some startups are spending over 100 million dollars a year on cloud services without being able to quickly launch products. If the gap between costs and products continues to widen, financing discounts and a brutal market reset are expected.
Model commoditization.
Ironically, the large language model (LLM) competition is driving rapid commoditization. Open-source releases (such as Meta's LLaMA, Mistral, etc.) blur the differences.
The moat is shifting towards data quality, distribution channels, or vertical integration.
If OpenAI begins to lose to streamlined open-source participants or models developed in-house by enterprises, venture capitalists may reassess the true meaning of 'defensibility'.
The second half of the year may sound the alarm: not every well-packaged wrapper deserves a billion-dollar valuation.
Financing scale slows down but remains high.
After the surge in the first half of the year, the pace of transactions will slow down. It is expected that there will not be another financing round of 40 billion dollars, but the quarterly AI financing amount will still be double that of 2024. The boom continues, just more robust.
A major liquidity event is coming.
At least one exit of over ten billion dollars is expected: IPO (e.g., Databricks) or acquisition by a traditional company trying to maintain influence.
This will affect investor sentiment and reset pricing expectations.
A clear stratification of the startup ecosystem, with differentiation becoming apparent by the fourth quarter.
Top 5-10 AI companies (with strong funding and development momentum) will gradually exit and may recruit talent through acquisitions.
For those startups in the mid-stage or overhyped but yet to achieve product-market fit? Many will undergo transformation, experience valuation downgrades, or gradually disappear.
Investors will reward execution that can generate revenue, not just investments in research programs or GPUs.
The next six months will put the AI narrative to the test. Will 2025 be the beginning of continuous change, or will it be a bubble that needs correction?
Some bubbles will burst, but the core argument still holds. AI remains the most attractive frontier in the venture capital space, but the flow of funds will be more cautious.