Venture Bytes #118: AI-Native FinTech Startups Redefining Value Creation

Over the past 25 years, fintech has evolved with each major tech cycle. The Internet era brought financial services online, the cloud era modularized the stack and made fintech ubiquitous, and the mobile era turned finance into a tap-and-go habit. With artificial intelligence (AI) now in the mix, fintech is gradually expanding its focus — one centered on intelligence, automation, and operational leverage.

‍
The next great arbitrage in fintech isn’t between capital and risk, it’s between talent and compute. As AI takes center stage in this shift, it is reshaping fintech across multiple fronts, creating investible opportunities in both established and emerging segments. While areas such as stablecoins, cross-border payments, and embedded finance remain strong, AI is also unlocking new potential in areas like regulatory technology, wealth management, banking, and corporate finance tools.

‍
Median pre-money valuations for all venture-backed fintech companies saw a remarkable 94.4% increase year-over-year in 2024, per Pitchbook. This momentum is closely tied to the transformative potential of AI in fintech, as investors are increasingly betting on AI- driven innovations to lead the next wave of growth in the sector.

Compliance stands out as a prime target for disruption. Over the past 50 years, the US Code of Federal Regulations has more than doubled in complexity, growing from 400,000 to over 1 million restrictive words, per Arlington-based research institute Mercatus Center. This  regulatory expansion  has  driven  compliance  costs for global banks to $270 billion annually, according to Deloitte, resulting  in  inefficiencies  that  hinder  progress.  Startups  such as  Norm  AI  are  addressing  this  challenge  with  regulatory  AI, enabling compliance and business teams to navigate rules swiftly and accurately. This positions AI-powered regulatory technology (regtech) start-ups as not just a cost mitigator, but also a high- growth investment area. The company, a leader in the emerging AI-driven compliance automation market, recently raised a $48 million Series B round, bringing its total raised over the last 20 months to $87 million.

‍
The wealth management sector is also experiencing a paradigm shift with AI as it shifts towards personalization and efficiency. Morgan Stanley’s internal deployment of AI illustrates this shift. By using AI@Morgan Stanley Debrief, a tool developed in collaboration with OpenAI, Morgan Stanley’s 16,000+ financial advisors have access to real-time, context-aware research. Compared to traditional wealth management, AI-powered solutions significantly reduce costs by automating routine tasks and portfolio rebalancing, leading to efficiency gains of 20-30%.
The upside is massive. McKinsey estimates AI could create $1 trillion in value across global banking, with wealth management capturing a meaningful portion. Startups are capitalizing on the trend. Range, based in Virginia, offers AI-powered financial planning at a fraction of the cost of traditional advisors—reporting up to 90% fee reductions.
Importantly, this transformation is not limited to high-net-worth individuals. Deloitte expects AI-based platforms to become the primary source of advice for this segment by 2027, with usage hitting 80% by 20     ritical inflection point for mass- market adoption.

‍

Similarly, the accounting and invoicing sector is experiencing a surge in investor interest, particularly for startups leveraging AI to automate complex workflows. With over 3 million accountants in the US and a 33% drop in CPA exam candidates from 2016 to 2021, the industry is ripe for disruption.

‍
The AI in accounting market is expected to expand from $6.7 billion in 2025 to $37.6 billion globally by 2030, at a CAGR of 41.27%, per Mordor Intelligence. Startups are moving quickly to seize this opportunity. Among early movers, New York-based Basis stands out, having raised $34 million in Series A funding from Khosla Ventures and BoxGroup to build AI-native financial operations from the ground up. The platform aims to automate workflows, reduce repetitive tasks, and enable accountants to focus on higher-value advisory roles.

‍
Finally, AI is fundamentally transforming financial services infrastructure, with agentic AI reshaping the core of the banking tech stack. This shift is driving renewed VC interest as banks urgently need scalable, real-time systems to support AI’s data and compute demands. Infrastructure providers are now mission- critical, enabling seamless data integration, robust security, and compliance for AI-driven operations.

‍
Startups such as Bud exemplify this trend. Its agentic AI, launched in  September  2024,  empowers  banks  to  deploy AI  agents  that autonomously  optimize  customer  finances,  reduce  overdraft fees, and boost profitability. Bud’s platform also extends agentic AI to  back-office  functions  like  fraud  detection  and         it  risk, highlighting the broad impact on banking operations.

Data Center Cooling Takes Center Stage

Thirty years ago, cooling a data center was as simple as scaling up an office HVAC system. But with today’s extreme compute density and rising energy loads, cooling infrastructure has evolved into a highly engineered, mission-critical system. Hyperscale facilities now generate 50 times more heat per square foot than office spaces and often draw more than 30 megawatts of power—nearly all of which is converted into heat. This shift has forced a leap from generic HVAC to precision cooling—a shift with massive investment implications.

The AI boom is fueling one of the most aggressive infrastructure buildouts in recent history. Goldman Sachs estimates that US companies will pour over $1 trillion into AI data centers in the coming years. Recently, the US Department of Energy (DOE) has also detailed 16 possible sites for the development of data centers on federal land. This spending is creating opportunities for growth across the entire AI data center value chain—but one area is emerging as especially critical: data center cooling.

‍

The shift to AI-optimized GPUs is fundamentally reshaping thermal profiles. A conventional CPU draws between 75 and 200 watts—comparable to a household lightbulb. In contrast, NVIDIA’s flagship GPUs for AI workloads draw 5–10 times as much power and generate proportionally more heat, according to Mitsubishi Heavy Industry. This change is already driving hyperscale data centers from the 30–40 megawatt range to well over 100 megawatts in the next few years.

‍
This strain is being compounded by climate volatility. In July 2022, a record heatwave in London sent temperatures above 40°C (104°F), knocking Google and Oracle data centers offline due to cooling failures. Similar events followed in California. The combination of denser compute and harsher weather patterns is making cooling a top priority.

‍
Operators are already feeling the pressure. AFCOM’s 2024 State of the Data Center Industry report reveals that only 46% of facilities have cooling systems that meet their current needs, while 35% report routinely running out of cooling capacity. Rack density is rising fast—jumping from 8.5 kW per rack in 2023 to 12 kW in 2024—and more than half of data center operators pect that figure to climb further in the next two to three years.

Data center cooling has become one of the most critical variables in hyperscale architecture. It’s the second-largest capital expense after electrical systems and a primary factor in operating expenses. Data from SemiAnalysis suggests that in a typical facility, 60 to 80% of non-IT electricity spend goes toward cooling, primarily through  chillers,  with  the  rest  split  between  power  conversion inefficiencies and marginal loads like lighting.

‍
Of all the core systems in a data center, cooling is arguably evolving the fastest, driven by rising rack densities and surging thermal loads. That’s where a new wave of startups is placing its bets. Texas-based LiquidStack, which develops advanced liquid cooling solutions for data centers and high-performance computing environments, is a standout. Fueled by the AI infrastructure boom, the company is seeing 4–5x demand for its coolant distribution units. Notably, LiquidStack has partnered with Standard Power to deliver the first large-scale US colocation data center using two-phase immersion cooling, marking a key milestone in industry-wide adoption.

Spain-based  Submer  is  taking  an  ecosystem-wide  approach. While it began as a liquid immersion cooling startup, it now offers datacenter design and construction services, as well as datacenter management  and  AI  as  a  Service  (AIaaS).  The  company  has secured $122.2 million in total funding, including a $55.5 million Series C in late 2024 that valued it at over $520 million. Its recent launch  of  a  56MW  liquid-cooled  facility  in  Barcelona  offers  a powerful proof point for the tech’s viability at scale.

‍
Accelsius is also at the forefront of cooling technology for next- generation AI computing. The company has set new industry benchmarks, achieving an exceptional 250kW cooling capacity per AI server rack while delivering a 25% reduction in cooling-related energy consumption. Strategic partnerships with Nordik Data Centers and Computacenter have accelerated its expansion across North America and Europe, strengthening its global footprint.

‍
The market opportunity for these start-ups is massive. According to  Research  &  Markets,  the  global  data  center  liquid  cooling market  is  expected  to  grow  from  $5.7  billion  in  2024  to  $48.4 billion  by  2034—a  staggering  24%  CAGR.  In  the  US,  growth  is expected at 17.1% annually through 2033. As capital continues to pour into GPUs, cloud infrastructure, and AI software, the physical infrastructure  enab                            oling—is  primed  for  its  own breakout moment.

‍