AI Bottleneck Investment Strategy: 14 Targets Covering Everything from Power to Lithography

Original Author: George Kikvadze

Original Translation: TechFlow

Preface: Bitfury Group Vice Chairman George Kikvadze proposes a contrarian approach: the most profitable opportunities in the AI sector lie not in the model layer, but in infrastructure bottlenecks like power, cooling, memory, and networking. He outlines 7 critical chokepoints in AI systems and reveals his portfolio of 14 picks, which has yielded approximately 60% returns so far. This "bottleneck investment" framework is worth a close look for anyone focused on AI investing.

To understand where the money is in AI, don't look at the headlines. Look at where the system is under strain.

The simplest analogy: today's AI is like a factory with unlimited orders, but the power, cables, and cooling can't keep up.

This mismatch itself is an opportunity.

After conducting detailed due diligence, we placed our bets on the following "AI Bottleneck" portfolio:

$CEG $GEV $VST $WMB $PWR $ETN $VRT $MU $ANET $ALAB $ASML $LRCX $CIFR $IREN

The Real Question to Ask

Most investors ask: "Who will win AI?" That's the wrong question.

The right question is: Where will the system break? Who gets paid to fix it?

In markets, dependencies are leverage.

AI's dependencies are anything but abstract; they are all physical:

• Megawatt-scale power
• Transformer lead times
• Cooling capacity per rack
• Memory bandwidth

The economic center of gravity is shifting towards these areas.

The Only Analytical Framework You Need

AI Expansion → Infrastructure Strain → Forced Investment → Bottlenecks → Pricing Power → Earnings Upgrades

When demand is inelastic and supply is constrained: prices move first, earnings follow, and stock re-ratings come last.

Why Now

A few numbers tell the whole story:

Nearly 50% of all data center projects in the US are currently delayed, not due to a lack of demand or funding, but because they can't secure power. Transformer lead times have stretched from 24 months before 2020 to over 5 years now. The data center construction cycle is 18 months. This math doesn't add up.

Hyperscalers' AI infrastructure spending alone will reach $700 billion by 2026, nearly 6 times the level in 2022. Amazon is spending $200B, Google $175-185B, Meta $115-135B. None are slowing down.

Semiconductors now account for 42% of the total market cap of the S&P 500 Information Technology sector, more than double the bottom of the 2022 bear market and over four times their weight in 2013. Semiconductors also contribute 47% of the sector's forward EPS, nearly triple the 2023 level.

The market is converging on the compute layer at an unprecedented density.

But compute is no longer the bottleneck.

Capital is flooding into chips, while the real constraints have shifted elsewhere.

This gap is the trading opportunity.

Bottleneck Map: Where the Pressure Really Is

• Power: The Foundation

AI cannot scale without power. Period.

The US needs to add the equivalent of the entire current data center power base every two years just to keep up with AI demand forecasts through 2030. Nuclear is the only source of baseload power that can provide the scale and reliability hyperscalers require, but even the fastest reactor restarts take years.

Picks: $CEG $GEV $VST $WMB

These aren't utility stocks; they are AI capacity providers. The market hasn't finished this reclassification yet. This mispricing is the opportunity.

Constellation Energy ($CEG) operates the largest fleet of nuclear plants in the US and is one of the few suppliers capable of providing large-scale, reliable, zero-carbon baseload power. Hyperscalers are rapidly signing long-term Power Purchase Agreements (PPAs) with nuclear operators, placing Constellation directly in the path of this demand.

GE Vernova ($GEV) is building the generation backbone for the next energy cycle, spanning gas turbines, renewable energy, and grid solutions. When AI demand accelerates, the ability to deploy power quickly and at scale becomes critical, and GE Vernova's gas turbine and electrification capabilities are central to this.

Vistra Corp ($VST) owns a diversified generation portfolio including nuclear, natural gas, and retail power, capable of handling both baseload and peak demand. This flexibility becomes particularly valuable as AI workloads create highly volatile power demand patterns.

Williams Companies ($WMB) operates one of the largest natural gas pipeline networks in the US, providing fuel to bridge the gap between current demand and future nuclear scale. For AI infrastructure expansion, natural gas offers the fastest pathway to incremental online power. Williams is essentially the fuel supplier for AI growth.

The Grid and Electrification: The Constraint Behind the Power

Generating power is one thing; delivering it is harder.

The US grid interconnection queue is now booked beyond 2030. Meeting just the existing commitments over the next decade requires over $50 billion in transmission investment, and this doesn't account for a single new AI data center coming online.

Picks: $PWR $ETN

This is where timelines slip and margins expand. Companies solving the "last mile" delivery problem possess durable, long-cycle pricing power.

Quanta Services ($PWR) is a leading contractor for building and upgrading the transmission infrastructure that connects power generation to consumption. As grid congestion becomes a primary bottleneck for AI expansion, Quanta sits directly in the path of multi-year, non-discretionary capital expenditure. Its backlog is a leading indicator of grid strain.

Eaton Corporation ($ETN) provides power distribution systems, switchgear, and power management technologies that enable safe and efficient power delivery at scale. As data centers push towards higher power densities and more complex energy flows, Eaton's components transition from standardized hardware to critical infrastructure.

Cooling: The Silent Ceiling

Heat kills performance. There is no software patch for thermodynamics.

Next-generation AI facilities target 250 kW per rack, compared to the 10-15 kW for standard enterprise data centers a decade ago. Liquid cooling is no longer optional; it's essential infrastructure. Every GPU sold requires a corresponding cooling capacity, and this ratio is fixed.

Pick: $VRT

Vertiv holds a near-monopoly position in hyperscale data center cooling. This is one of the most undervalued parts of the entire AI stack because no one pays attention to cooling until a cluster goes down.

Vertiv Holdings ($VRT) designs and deploys thermal management systems that keep high-density AI clusters operational under extreme power loads. As racks shift from air cooling to liquid cooling, Vertiv is at the center of this structural upgrade cycle, expanding in lockstep with AI compute deployment. This is not optional spending; it's a prerequisite for uptime.

Memory: The Next Bottleneck

AI is transitioning from being compute-constrained to being memory-constrained.

As models grow larger and inference volumes explode, the bandwidth and capacity of memory become the binding constraints, not raw processing power. HBM (High Bandwidth Memory) supply is already tight. The world's top three AI memory suppliers control over 90% of global HBM output. Micron is the primary Western beneficiary.

Core Pick: $MU

This is the next wave of earnings upgrades. Most portfolios are not positioned for it yet. When the market catches on, they will be.

Micron Technology ($MU) is one of the few global players capable of mass-producing advanced HBM, a critical component for AI training and inference workloads. As memory becomes the binding constraint on system performance, Micron transitions from a historically cyclical supplier to a structural beneficiary of AI demand. This shift is not yet fully reflected in its valuation, leaving room for continued earnings upgrades and multiple expansion.

Networking: The Throughput Layer

An AI cluster's speed is determined by its slowest connection.

A single network bottleneck can stall an entire cluster of thousands of GPUs, wasting hundreds of millions of dollars of capital investment per facility. As cluster scales expand towards 100,000 GPU configurations, the interconnect problem amplifies exponentially. One choke point, and the entire line shuts down.

Picks: $ANET $ALAB

Quiet, critical, under-owned. No one talks about networking until it breaks.

Arista Networks ($ANET) builds high-performance network infrastructure that enables data to flow seamlessly across massive AI clusters. When workloads demand ultra-low latency and high throughput, Arista's software-defined networking becomes crucial for maintaining cluster efficiency. The cost of downtime or inefficiency is enormous, and Arista captures value by ensuring systems run at full speed.

Astera Labs ($ALAB) operates inside the data path, ensuring high-speed connections between GPUs, CPUs, and memory within AI systems. As cluster density increases, bottlenecks shift from the network edge to chip-to-chip communication, which is precisely Astera's domain. In high-performance AI environments, if components can't talk to each other fast enough, the entire system slows down.

Manufacturing: The Long-Cycle Constraint

AI cannot scale without chip fabrication capacity. And advanced chips cannot be made without manufacturing tools.

ASML's EUV lithography machines have a production cycle of over a year, cost over $200 million each, and have no credible substitutes. Every advanced chip on earth, from NVIDIA's H100 to Apple's M-series, requires their equipment. Lam Research's etching and deposition tools are embedded in the production lines of every major wafer fab globally.

Picks: $ASML $LRCX

Long-cycle constraints. Structurally harder to disrupt than any software moat. Much less discussed than they should be.

ASML Holding ($ASML) is the sole supplier of EUV lithography systems, the most advanced chipmaking tools in existence and a prerequisite for producing cutting-edge semiconductors. With a multi-year order backlog and no viable competition, ASML controls the critical chokepoint in the global chip supply chain.

Lam Research ($LRCX) supplies the etching and deposition equipment that forms the backbone of semiconductor manufacturing. Its tools are deeply embedded in all major wafer fabs, making it a recurring and indispensable partner in chip capacity expansion. As AI demand drives continuous capacity expansion, Lam captures long-cycle revenue directly tied to global semiconductor manufacturing growth.

Misclassification: The Source of Alpha

This is the part most investors overlook, and the most asymmetric opportunity on the entire map.

There is a class of companies the market prices as A, but whose operational and financial reality is already B.

Take $CIFR (Cipher Digital) and $IREN (IREN Limited), for instance.

What the market still sees are Bitcoin miners.

What they are becoming is far more valuable: AI power infrastructure and HPC data center platforms.

These companies locked in low-cost power and built infrastructure before anyone was paying attention, before the demand was there. Today, hyperscalers are scrambling for exactly these two things.

Cipher Digital is already executing on this transition, signing a 15-year lease with an investment-grade hyperscale tenant (its third AI/HPC campus) and securing a $200 million revolving credit facility from top-tier global banks. These aren't speculative moves; they are long-cycle revenue commitments.

IREN is executing the same playbook across multiple sites, combining energy access with scalable data center construction. Its advantage is speed: it already controls the land, power, and infrastructure needed to pivot towards AI workloads.

The market sees miners. The balance sheets already look like infrastructure companies.

This gap will converge. And when it does, it won't be gradual.

Portfolio at a Glance

This isn't a collection of stocks; it's a system.

Each position corresponds to a specific constraint within the AI stack. Each constraint must be resolved for the system to function. That is the discipline.

• Power: $CEG $GEV $VST $WMB
• Grid: $PWR $ETN
• Cooling: $VRT
• Memory: $MU
• Networking: $ANET $ALAB
• Manufacturing: $ASML $LRCX
• Misclassification: $CIFR $IREN

The Cognitive Shift Most Investors Haven't Made Yet

We are moving from compute scarcity to infrastructure scarcity.

This means:

• GPUs are no longer the only narrative
• Power, grid, memory, and cooling become the dominant earnings drivers
• Returns follow constraints, not hype

Most portfolios are still positioned in the old world.

Risks: Discipline Matters Too

This framework breaks down under certain conditions. They deserve honest acknowledgment.

Hyperscaler CapEx Slowdown. If Amazon, Google, and Meta reduce infrastructure spending due to margin pressure or disappointing demand, the assumption of inelastic demand weakens. This is the primary risk to monitor, focusing on quarterly CapEx guidance as a leading indicator.

Bottleneck Resolution Faster Than Expected. Government intervention in transformer manufacturing, accelerated nuclear permitting, or grid interconnection queue reforms could compress the premium on constrained infrastructure. These changes are slow but real.

Regulatory Friction