What are the fundraising trends in the AI infrastructure market?
Download our beautiful pitch about the AI infrastructure market
In our AI infrastructure market deck, you will find everything you need to understand the market
Between 2022 and 2025, pure-play AI infrastructure startups raised over $24 billion across 134 equity rounds, turning what was once a niche hardware category into the fastest-growing segment of the tech funding landscape.
The AI infrastructure market went from $1.3 billion in annual funding in 2022 to nearly $12.8 billion in 2025, a tenfold increase driven by the explosion in demand for GPU compute, custom AI chips, and optical interconnects.
Nvidia alone participated in over 30 funding rounds during this period, effectively acting as both the market's dominant supplier and its most prolific investor.
And if you want to better understand this new industry, you can download our pitch covering the AI infrastructure market.
Insights
- AI infrastructure funding grew roughly 10x in just three years, from $1.3 billion in 2022 to $12.8 billion in 2025, faster than almost any other enterprise tech category in recent memory.
- Nvidia invested in 30 AI infrastructure deals worth a combined $10 billion or more, making it the single most influential investor-customer in the AI infrastructure market.
- GPU cloud and AI data center startups captured around $10.3 billion over 2022 to 2025, nearly half of all AI infrastructure funding, reflecting the industry's desperate need for compute capacity.
- The average AI infrastructure deal size jumped from $72 million in 2022 to $242 million in 2025, showing this market is rapidly leaving traditional venture territory.
- Photonic and optical interconnect startups raised close to $2 billion across 19 rounds between 2022 and 2025, making light-based data movement the quiet breakout category in AI infrastructure investing.
- Five of the ten largest AI infrastructure rounds ever happened in September 2025 alone, including Lambda, Crusoe, Nscale, Cerebras, and Groq.
- South Korea's AI chip ecosystem punched well above its weight, with Rebellions, FuriosaAI, DeepX, and HyperAccel raising over $700 million combined despite operating outside the US.
- No AI infrastructure startup below $2 million was tracked across all 134 deals in four years, confirming that this is not a market where small seed checks can compete.
- Four top-funded AI infrastructure companies (Groq, Celestial AI, and others) were acquired before the end of 2025, signaling that M&A is now a primary exit path in AI infrastructure.
- The share of funding captured by the top 10 deals actually dropped from 85% in 2023 to 62% in 2025, suggesting the AI infrastructure market is broadening even as total capital grows.
First, how do we define the AI infrastructure market?
We define the AI infrastructure market as the technologies and services required to run AI training and inference reliably at scale.
We include AI compute (accelerators and servers), AI-optimized cloud and cluster platforms, and the networking and storage required to move and serve model data and outputs.
We exclude end-user AI applications, foundation-model API services as a "model product," and general-purpose data, analytics, and MLOps tools that are not necessary to operate the compute-and-cluster layer.
This is also the definition we use in our pitch about the AI infrastructure market.
In our AI infrastructure market deck, we have designed useful charts to give you full market clarity
How has funding activity in the AI infrastructure market changed over time?
2025 was by far the most active year for AI infrastructure funding, with $12.8 billion raised across 53 deals, driven by multiple billion-dollar rounds from Lambda, Crusoe, Nscale, and Cerebras rather than a single outlier.
2022 was the quietest year, with just $1.3 billion across 18 deals, because the market was still in its pre-ChatGPT phase and AI infrastructure remained a specialized bet.
Compared to 2024, AI infrastructure funding in 2025 grew by 53%; compared to 2023, AI infrastructure funding in 2025 was nearly 7 times larger; and compared to 2022, AI infrastructure funding in 2025 was roughly 10 times higher.
Even if you strip out the top two deals each year to remove mega-round distortion, the underlying trend still holds: AI infrastructure funding went from about $970 million in 2022 to $1.2 billion in 2023, then jumped to $6.6 billion in 2024 and $10 billion in 2025.
If you're interested in this industry, please note that we regularly keep in touch and share funding updates for this market on this page, which we keep continuously updated.
We also make quarterly analyses of the funding activity in the AI infrastructure market here.
| Year | Number of deals | Total raised ($) | Comment |
|---|---|---|---|
| 2022 | 18 | $1.3B | A quiet, pre-ChatGPT year where AI infrastructure funding was still a niche bet dominated by chip IP and storage players. SiFive's $175M Series F was the largest round, and optical interconnects quietly emerged as an active subcategory. |
| 2023 | 19 | $1.6B | The post-ChatGPT wave arrived, and GPU cloud providers like CoreWeave captured over a quarter of all funding with $421M. Photonic interconnect companies Lightmatter and Celestial AI raised a combined $409M as investors recognized the data movement bottleneck. |
| 2024 | 44 | $8.4B | AI infrastructure funding exploded 5x versus 2023, with CoreWeave's $1.1B Series C setting the pace. Asian semiconductor players emerged as a force, capturing over $2B, and the average deal size hit $191M. |
| 2025 | 53 | $12.8B | Record-breaking AI infrastructure investment featured four billion-dollar rounds in a single quarter. Lambda raised $2B across two rounds, while Nscale's $1.5B showed Europe emerging as a sovereign AI compute player. |
In our AI infrastructure market deck, we help you understand how the market is structured
Which startups in the AI infrastructure market raised the largest rounds over the last few years?
These startups raised the most over the last years in the AI infrastructure market:
- Lambda raised $1.5 billion in November 2025 to build gigawatt-scale AI factories, on top of an earlier $480M round the same year, making Lambda the single largest AI infrastructure fundraiser of the period.
- Crusoe raised $1.375 billion in October 2025 to scale its clean-energy AI data centers in Texas and Wyoming, backed by Valor Equity, Mubadala, and Nvidia.
- Nscale raised $1.1 billion in September 2025, making it the largest Series B ever in the UK, to build AI factories for OpenAI's Stargate project across Europe.
- Cerebras Systems raised $1.1 billion in September 2025 to expand wafer-scale AI processor manufacturing at an $8.1B valuation, later climbing to $23.1B.
- CoreWeave raised $1.1 billion in May 2024 to expand its specialized GPU cloud infrastructure across the US and Europe at a $19B valuation.
- Firmus raised $830 million across two 2025 rounds to build 1.8GW of AI data center capacity in Australia, backed by Nvidia and valued at $3.9B.
- Groq raised $750 million in September 2025 for its Language Processing Unit inference chips, before being acquired by Nvidia in December 2025 for around $20B.
- SJ Semiconductor raised $700 million in December 2024 to expand its advanced 2.5D/3D packaging foundry, critical for assembling multi-die AI chiplets.
- Horizon Robotics raised $696 million through Hong Kong's biggest IPO of 2024, with Alibaba and Baidu as cornerstone investors, to fund autonomous driving AI chip R&D.
- Tenstorrent raised $693 million in December 2024 backed by Jeff Bezos, Samsung, and Fidelity to scale its RISC-V AI chip architecture as an open alternative to Nvidia.
And, yes, we do cover most of them in our our beautiful pitch about the AI infrastructure market.
In our AI infrastructure market deck, we answer all the common questions from investors and entrepreneurs
Is the AI infrastructure market shifting toward smaller or bigger deals?
According to our own data, the average deal size across all 134 AI infrastructure funding rounds from 2022 to 2025 was approximately $180 million, which is far above what you would normally expect in a startup market.
Breaking it down by year, the average AI infrastructure deal was $72 million in 2022, $86 million in 2023, $191 million in 2024, and $242 million in 2025. The jump between 2023 and 2024 was the most dramatic, reflecting the moment when billion-dollar GPU cloud and AI chip rounds became normal in the AI infrastructure market.
Even when excluding the top two outlier deals per year, the average AI infrastructure deal size still rose steadily from about $61 million in 2022 to $195 million in 2025, confirming this is a genuine structural shift rather than a few mega-rounds pulling the average up.
| Year | Number of deals | Avg. deal size ($) | Deals below $2M | Deals above $50M |
|---|---|---|---|---|
| 2022 | 18 | $72M | 0 | 9 |
| 2023 | 19 | $86M | 0 | 11 |
| 2024 | 44 | $191M | 0 | 29 |
| 2025 | 53 | $242M | 0 | 35 |
| 2022-2025 | 134 | $180M | 0 | 84 |
In our AI infrastructure market deck, we identify repeatable patterns you can use if you’re building in this market
How concentrated was funding activity in the AI infrastructure market?
AI infrastructure funding has always been top-heavy, but the concentration has actually decreased as the market matured. In 2023, the top 10 AI infrastructure deals captured 85% of all capital, while in 2025 that figure dropped to 62%, even though total funding was eight times larger.
The share captured by the single largest AI infrastructure deal stayed remarkably stable at around 13-15% each year, meaning no single company ever dominated the market completely. What changed is that the pool of well-funded AI infrastructure startups got much wider, with 35 deals above $50 million in 2025 compared to just 9 in 2022.
| Year | Number of deals | % by Top 1 | % by Top 3 | % by Top 10 |
|---|---|---|---|---|
| 2022 | 18 | 13.5% | 31.6% | 76.1% |
| 2023 | 19 | 13.5% | 35.3% | 84.9% |
| 2024 | 44 | 13.1% | 29.7% | 71.9% |
| 2025 | 53 | 15.4% | 34.8% | 62.4% |
| 2022-2025 | 134 | 9.9% | 25.0% | 55.1% |
In our AI infrastructure market deck, we track adoption trends and shifts in consumer behavior
Which categories in the AI infrastructure market received the most funding?
GPU cloud and AI data center companies captured around $10.3 billion across roughly 32 deals from 2022 to 2025, representing about 43% of all AI infrastructure funding. This dominance reflects the brutal reality that access to GPU compute became the single biggest bottleneck for AI development, pushing startups like Lambda, CoreWeave, Crusoe, and Nscale into multi-billion-dollar fundraising territory.
AI chips and accelerators came in second with roughly $8.5 billion across about 49 deals, making it the most active AI infrastructure category by deal count. The category spans everything from Cerebras' wafer-scale processors to Groq's inference LPUs to Unconventional AI's neuromorphic approach, showing that investors are backing a wide range of architectural bets to dethrone Nvidia.
Photonic and optical interconnect companies raised close to $2 billion across around 19 deals, making light-based data movement the third-largest AI infrastructure funding category. Celestial AI, Lightmatter, and Ayar Labs led the charge, with investors recognizing that moving data fast enough between processors will determine how large AI models can get.
| Category | Number of deals | Total raised ($) | Startups and amount |
|---|---|---|---|
| GPU Cloud & AI Data Centers | ~32 | ~$10.3B | Lambda ($2.4B), Crusoe ($2.0B), Nscale ($1.7B), CoreWeave ($1.5B), Firmus ($1.0B), Together AI ($534M), FluidStack ($200M), Ori ($175M), Applied Digital ($160M), TensorWave ($100M), Anyscale ($99M), Neysa ($80M), Infinigence AI ($70M), NextGen Cloud ($45M), Modal Labs ($23M), RunPod ($20M), GMI Cloud ($15M), Lepton AI ($11M) |
| AI Chips & Accelerators | ~49 | ~$8.5B | Groq ($1.4B), Cerebras ($1.1B), Tenstorrent ($793M), Horizon Robotics ($696M), SiFive ($175M), Etched ($625M), D-Matrix ($429M), Rebellions ($447M), Hailo ($240M), SiMa.ai ($115M), Axelera AI ($118M), Tsing Micro ($283M), EnCharge AI ($100M), Mythic ($125M), DeepX ($81M), FuriosaAI ($125M), SiPearl ($135M), Blaize ($106M), Recogni ($102M), and others |
| Photonic & Optical Interconnects | ~19 | ~$2.0B | Lightmatter ($709M), Celestial AI ($642M), Ayar Labs ($310M), nEYE Systems ($58M), Scintil Photonics ($58M), Q.ANT ($72M), Xscape Photonics ($44M), OpenLight Photonics ($34M), Arago ($26M), iPronics ($23M), Lumai ($10M) |
In our AI infrastructure market deck, we have collected signals proving this market is hot right now
Who are the biggest investors in the AI infrastructure market?
Nvidia is the single largest investor in the AI infrastructure market by a wide margin, participating in roughly 30 deals worth a combined $10 billion or more between 2022 and 2025. Nvidia invested across every layer of the stack, from GPU cloud providers like Lambda and CoreWeave to optical interconnect companies like Ayar Labs and storage platforms like WEKA, effectively funding the ecosystem that drives demand for its own chips.
Fidelity Management & Research appeared in about 12 AI infrastructure deals, including billion-dollar rounds for Cerebras, Crusoe, CoreWeave, and Nscale. Fidelity's heavy presence signals that large institutional asset managers increasingly view AI infrastructure as a core allocation, not a speculative venture bet.
Samsung Catalyst Fund also participated in around 12 AI infrastructure rounds, including investments in Groq, Tenstorrent, Enfabrica, Axelera AI, and Rebellions. Samsung's strategy reflects its position as both a memory supplier and strategic investor seeking to ensure its chips are designed into next-generation AI systems.
Atreides Management backed about 8 AI infrastructure deals, including Cerebras, Enfabrica, WEKA, and Ayar Labs. Atreides consistently picked companies solving data movement and compute bottlenecks, making it one of the most thesis-driven infrastructure investors.
Temasek and its Xora Innovation fund participated in roughly 7 AI infrastructure rounds, with a particular focus on Celestial AI (backing the company across multiple rounds) and d-Matrix. Temasek's persistence with photonic and inference chip bets shows a long-term conviction in alternative AI compute architectures.
Lux Capital invested in about 7 AI infrastructure deals, including Together AI (three rounds), Unconventional AI, and Modal Labs. Lux Capital has been the most consistent backer of alternative AI compute platforms, following Together AI from seed through its $305M Series B.
Intel Capital participated in about 6 AI infrastructure deals, mostly concentrated in 2022 with SiFive, Anyscale, Pliops, and Ayar Labs. Intel Capital's early infrastructure bets were well-timed, though the firm was notably less active in the later boom years.
Valor Equity Partners backed around 6 AI infrastructure deals, including Crusoe (twice), Cerebras, Enfabrica, and WEKA. Valor showed a strong preference for companies with physical infrastructure moats, particularly energy-optimized data centers.
Prosperity7 Ventures (Saudi Aramco's venture arm) invested in about 6 AI infrastructure rounds, including Together AI (multiple rounds), SiFive, Rain Neuromorphics, and TensorWave. Prosperity7's involvement reflects the broader trend of Middle Eastern sovereign capital flowing into AI infrastructure.
IAG Capital Partners participated in roughly 6 AI infrastructure deals focused primarily on optical interconnects and networking, including Celestial AI, Ayar Labs, Enfabrica, and Xscape Photonics.
Disclaimer: this investor list may be incomplete; we focus on publicly disclosed lead and prominent recurring investors, so some frequent minority participants may be underrepresented. "Total funded" does not represent the amount personally invested by an individual investor. Instead, it refers to the aggregate amount raised across all fundraising rounds in which the investor participated.
| Investor | Number of deals | Total funded ($) | Startups |
|---|---|---|---|
| Nvidia | ~30 | ~$10.4B | Lambda (x3), CoreWeave, Crusoe (x2), Nscale (x2), Together AI, WEKA (x3), Ayar Labs (x2), Pliops, Firmus (x2), Lightmatter, Enfabrica, VAST Data, Applied Digital, Fireworks AI, Xscape Photonics, Scintil Photonics, TensorWave, nEYE Systems, Phaidra, MetAI |
| Fidelity Management & Research | ~12 | ~$7.1B | Cerebras, Crusoe, CoreWeave, Nscale (x2), Lightmatter (x2), Lambda, Tenstorrent, VAST Data, Celestial AI, SiMa.ai |
| Samsung Catalyst Fund | ~12 | ~$2.7B | Groq, Tenstorrent (x2), Celestial AI (x2), Enfabrica, Axelera AI, Axiado, Rebellions, SiFive, EnCharge AI |
| Atreides Management | ~8 | ~$2.8B | Cerebras, Enfabrica (x2), WEKA (x2), Ayar Labs, Axiado, Mythic |
| Temasek / Xora Innovation | ~7 | ~$1.2B | Celestial AI (x4), D-Matrix (x2), Blaize, Rebellions (2022) |
| Lux Capital | ~7 | ~$1.0B | Together AI (x3), Unconventional AI, Modal Labs, 1001 AI |
| Intel Capital | ~6 | ~$0.6B | SiFive, Anyscale, Ayar Labs (x2), Pliops, RunPod |
| Valor Equity Partners | ~6 | ~$3.7B | Crusoe (x2), Cerebras, Enfabrica (x2), WEKA |
| Prosperity7 Ventures | ~6 | ~$0.8B | Together AI (x3), SiFive, Rain Neuromorphics, TensorWave |
| IAG Capital Partners | ~6 | ~$0.6B | Celestial AI (x2), Ayar Labs (x2), Enfabrica, Xscape Photonics |
In our AI infrastructure market deck, we cover the latest tech updates shaping the market
What are the 2025 narratives around fundraising in the AI infrastructure market?
These are the dominant narratives shaping fundraising in the AI infrastructure market in 2025:
- GPU cloud companies are racing to lock up power and land, with AI data center operators like Crusoe, Lambda, and Firmus raising billions because securing gigawatt-scale energy is now a competitive moat.
- Inference is overtaking training as the biggest AI infrastructure spending category, with chip companies like Groq, Etched, D-Matrix, and Cerebras all raising massive rounds specifically for inference workloads.
- Nvidia's acquisitions of Groq and Celestial AI in late 2025 signal that the era of independent AI chip startups may be ending, as vertical integration accelerates across the AI infrastructure supply chain.
- European sovereign AI ambitions are turning into real money, with Nscale raising $1.5 billion to build AI factories for OpenAI's Stargate project across Norway and the UK.
- Photonic computing is graduating from R&D to production, with seven photonic AI infrastructure companies raising a combined $280 million in 2025 alone, including industrial-scale players like Scintil Photonics and Q.ANT.
- Neuromorphic and brain-inspired chips attracted their first mega-round when Unconventional AI raised $475 million in a seed round, the clearest signal yet that investors believe GPUs are not the permanent answer to AI compute.
- South Korea is building a national AI chip ecosystem, with Rebellions, FuriosaAI, and Dnotitia raising over $450 million backed by coordinated state and corporate capital from Korea Development Bank, Samsung, and Arm.
- Clean energy is becoming a deal requirement for AI infrastructure investors, with Crusoe, Nscale, and Firmus all emphasizing renewable or energy-efficient power as a core differentiator in their fundraising pitches.
- The AI infrastructure market saw its first $400 million-plus seed rounds in 2025 (Unconventional AI at $475M), breaking all previous records and raising the floor for what counts as a "serious" entry into AI hardware.
- Middle Eastern and sovereign wealth fund capital is flowing heavily into AI infrastructure, with Mubadala, Qatar Investment Authority, Wa'ed Ventures, and DAMAC all appearing as investors in major 2025 rounds.
In our AI infrastructure market deck, we will give you useful market maps and grids
Related blog posts
- What is the latest update in the AI infrastructure market?
- What are the latest news in the AI infrastructure market?
- What is the real market size of the AI infrastructure market?
- What are the latest funding news in the AI infrastructure market?
- Evolution of the funding activity in the AI infrastructure market
