VCs look to AI datacenters, local LLMs, and domain models for growth

Venture capital firms are focusing on emerging AI opportunities that show promise for long-term growth as they look to emerge from the slump in deal activity and exit values. Pitchbooks’ latest Artificial Intelligence and Machine Learning report, released today, reflects the continued challenges facing VCs, including declining deal activity and exit values. According to Pitchbook analysis, three of the many growth catalysts venture capitalists need to sustain their company’s growth and profitability are AI data centers, local large-scale language models (LLM), and domain-specific foundations. It’s a model.

Further market disruption for VCs

AI and machine learning (ML) trading activity plummeted 19% in just one year, from 8,968 transactions in 2022 to 7,238 transactions in 2023. AI and ML deal value and volume also declined. Disclosed deal value in the fourth quarter of 2023 was $2.7 billion, the lowest quarter since the first quarter of 2019, according to Pitchbook. Merger and acquisition (M&A) activity continues to decline as major technology companies focus on partnering with LLM startups.

Pitchbook notes that exceptions to this trend include AMD’s acquisition of Nod.AI in machine learning operations (MLOps), IBM’s acquisition of Manta in database management, and ServiceNow’s acquisition of UltimateSuite in predictive analytics. The IPO of semiconductor startup Astera Labs is expected to reinvigorate trading volume in the first or second quarter of this year.

Amid sharp declines in trading activity and declining transaction values, there are also signs of long-term growth. Generative AI leaders raised $6 billion in 194 deals in the fourth quarter of 2023 alone. This is largely due to the support of Microsoft, Google, and other tech giants seeking access to the latest in his LLM technology. Pitchbook notes that horizontal platforms are also gaining momentum, setting a VC record by raising $33 billion in 2023. Investment in vertical applications has plummeted to levels not seen since 2020.

Where VCs say there are new opportunities

Creating an organizational structure and product strategy that can take advantage of Nvidia’s many innovations, including rapid advances in GPUs, is central to the new investment opportunity. Pitchbook analysis finds his three emerging areas – AI data centers, local LLM, and domain-specific foundational models – well positioned to benefit from his Nvidia momentum as a key driver of the AI ​​market. I found out something.

Nvidia reported fourth-quarter 2024 revenue of $22.1 billion, up 265% year-over-year and 22% sequentially. The data center segment grew 409% year over year and 27% sequentially to $18.4 billion. Jensen Huang, Founder and CEO of Nvidia, said: “Our data center platform is underpinned by an increasingly diverse set of drivers: large cloud service providers, GPU-focused providers, and even enterprise software for data processing, training, and inference. Demand. And consumer internet companies. Vertical industries centered around autos, financial services, and health care are now reaching multi-billion dollar levels.”

AI data centers show potential for exponential growth

These data centers are designed from the infrastructure layer to scale and support more AI-intensive workloads and are optimized to get the most value from high-performance servers, storage, networking, and dedicated accelerators. It has been. AI data centers must also be designed to optimize power consumption and heat dissipation for high-performance GPUs, with an emphasis on sustainability.

IDC estimates that $8 billion was invested in generative AI processors, storage, and networking, generating $2.1 billion in cloud revenue and $4.5 billion in application revenue. Pitchbook predicts that AI data centers will not reach SaaS (Software-as-a-Service) level margins until 2027. The startup is focused on providing cost-effective solutions and significant GPU time savings.

Pitchbook says, “According to hourly on-demand pricing, the startup offers 50% to 70% cost savings on advanced Nvidia A100 GPU hours and unique access to the latest H100 chips. ” states. The report notes that Lambda, a leading startup GPU cloud provider, has built the largest cluster of his H100 chips of any public cloud, surpassing Google and Oracle.

VCs will be building an ecosystem of colocation providers and evaluating opportunities to partner. Pitchbook reports that the specialist cloud provider has carved out its $4.6 billion market from the nearly $150 billion Internet-as-a-Service market, with more than 90% of it driven by U.S.-based hyperscalers and Chinese cloud giants. It is pointed out that it has been done. What makes specialty cloud providers unique is their ability to differentiate based on AI chip availability, local presence, multi-cloud support and support for multiple types of legacy hardware.