Article 144: Funds Flow into AI Startups... "Unique Data and Computing Infrastructure" Become the Key to Victory

In 2025, the artificial intelligence(AI) venture capital ecosystem is experiencing intense fluctuations driven by unprecedented investment scales and technological innovation. AI startups have raised approximately $100 billion(about 144 trillion Korean won) in just the first half of the year, accounting for nearly half of the total global startup investment. The capital is centered around massive transactions, focusing on AI technology, infrastructure, and application layers, with related companies rapidly emerging.

Against this trend, Crunchbase interviewed six leading AI venture investors and compiled their views on market structure, startup investment strategies, and next-generation growth drivers.

Accel partner Philippe Botteri pointed out that the six major “super companies” in the US—NVIDIA(NVDA), Microsoft(MSFT), Apple(AAPL), Alphabet(GOOGL), Amazon(AMZN), and Meta(META)—are pouring billions of dollars in cash flow into AI infrastructure. However, he emphasized that despite these huge capital inflows, startups based on core AI technologies still have opportunities to explore new fields. In fact, Accel has proactively invested in AI companies covering model and application layers, such as Anthropic, Perplexity, Synesthesia, and Sierra.

Foundation Capital partner Steve Baser analyzed that the bottleneck in the AI industry is gradually shifting toward physical infrastructure, with hardware ecosystems based on chips, power, and data centers becoming the core opportunities of the future. Leveraging experience from early-stage investment in AI chip manufacturer Cerebras, Foundation Capital has invested in over 100 AI startups, recently funding medical automation company Tenor and code error prediction tool provider PlayerZero.

Dell Technologies(Dell Technologies)’s venture capital division DTC is actively investing in AI chips and application software while selling GPU servers. DTC Managing Director Daniel Docter stated, “The investment speed in AI startups is faster than ever,” and revealed that some deals were signed within two days after meetings. Dell has also begun acquiring AI chip startup Rivos, which is currently awaiting regulatory review.

Sierra Ventures CEO Tim Guleri pointed out, “If computing power is the bottleneck, then data is the differentiating factor.” Sierra divides AI investments into five levels, from infrastructure to vertical applications, with a particular focus on the application layer that relies on unique data and distribution strategies. Guleri noted that most of the global $110 trillion GDP remains concentrated in service industries that have yet to digitize with AI, and he predicts explosive growth in AI-based value creation over the coming decades.

AI funds and Google(GOOGL)'s brain co-founder Andrew Ng have adopted a strategy of exclusive data access through a venture studio model. His fund has established partnerships with large companies such as AES, HP, and Mitsui, designing AI companies based on data from physical industries, and directly hiring CEOs to start ventures. Ng emphasized, “AI is not a single technology but a network creating customized opportunities across various industries.”

Finally, Alphabet’s investment arm GV boldly invests in AI startups that may conflict with existing Google businesses. GV provides funding from early stages—covering chips, compilers, to user applications—and, based on rapid revenue growth and market opportunities, also accepts high valuation premiums.

From Accel to Sierra Ventures and GV, the key words shared by major venture capital institutions are “computing resources,” “unique data,” and “application innovation.” The AI boom is not merely a trend but a technological and industrial revolution, likely to become a decisive variable in spawning new unicorns across various industries in the coming years.