AI Buildout Meets AI/Energy Deflation. Wen Hyperscalers?
The “AI Buildout Trade”
Over at LaDucTrading, we have an active Live trading room which I run and then a lively Slack live huddle which my moderators/bench of pro contributors run: Luke, Archna, Mike & Alex.
Obviously, we have been on this MAG7 short basket a few weeks now:
But today, Luke made a solid point,
“Given the epic performance of the mag7 and hyperscalers today (insert snarky emoji) - it looks like the market is once again rewarding the buildout names and punishing the ones who have to pay for it!
But the AI Buildout trade appears alive and well”
From his channel:
NEW 52 WEEK HIGHS TODAY:
Micron +17% $MU
Applied Materials +11% $AMAT
Corning +9% $GLW
Teradyne +9% $TER
Caterpillar +6% CAT 0.00%↑
United Rentals +5% $URI
Rockwell Automation +4% $ROK
Trane Technologies +4% $TT
Wabtec +4% WAB 0.00%↑
The list of tickers hitting new all-time highs points strongly to the “AI data center infrastructure buildout” phase of the AI trade being in the spotlight right now (as of late June 2026).
This is the “picks and shovels” / physical scaling theme — companies supplying the hardware, materials, equipment, and systems needed to actually construct and operate massive AI compute facilities — rather than pure AI software/models, hyperscaler end-users, or just GPU designers like NVIDIA.
Breakdown by Theme
Here’s how the tickers map to the dominant sub-themes:1. AI Semiconductor Supply Chain & Memory (Core enablers of AI accelerators)
MU (Micron): Leading the move (+17%). Strong beneficiary of HBM (High Bandwidth Memory) demand for AI GPUs. Recent earnings showed explosive growth from AI-driven memory pricing and volumes, with 2026 HBM supply reportedly sold out.
AMAT (Applied Materials): Semiconductor capital equipment — tools for manufacturing advanced chips.
TER (Teradyne): Automated test equipment critical for validating complex AI chips and advanced packaging (2.5D/3D). New solutions specifically highlighted for AI/data center devices.
2. Data Center Physical Construction & Equipment
CAT (Caterpillar): Heavy machinery + power generation equipment (e.g., natural gas generators) for hyperscale data center projects and backup/on-site power.
URI (United Rentals): Equipment rental for data center construction and maintenance (construction gear, power, climate control, etc.).
3. Cooling / Thermal Management (Critical bottleneck for high-density AI racks)
TT (Trane Technologies): HVAC and advanced cooling systems. AI servers generate enormous heat, making cooling infrastructure a major growth area.
4. High-Speed Optical Networking & Connectivity (Interconnects inside/between AI clusters)
GLW (Corning): Optical fiber, cables, connectors, and co-packaged optics (CPO) solutions. Major deals with Meta (up to $6B), Amazon, and collaborations tied to AI data center scaling for higher bandwidth and lower power.
Peripheral / Broader Industrials
ROK (Rockwell Automation): Industrial automation and control systems (benefits from smart manufacturing/fabs or data center operations).
WAB (Wabtec): More tangential — rail/transport equipment; benefits from general industrial activity or logistics tied to large infrastructure projects.
Key Conclusions About the Current AI Trade
The market is emphasizing real-world scaling constraints: Power, cooling, construction, memory supply, chip production capacity, and high-speed interconnects are the current focus. This reflects hyperscalers pouring hundreds of billions into data center capex, creating bottlenecks that benefit these suppliers.
Broadening beyond “pure AI” tech: Strength is spilling into traditional industrials, materials, and equipment providers. This suggests the AI trade is maturing from hype around models/chips into tangible infrastructure investment.
Supply constraints are driving gains: Examples include sold-out HBM capacity and surging demand for optics/fiber. This supports pricing power and visibility for these companies.
Rotation or diversification within AI: While some big AI names (or certain segments) may have pulled back, money is flowing into the infrastructure layer that enables further AI expansion. This can be a more “defensive” or sustained part of the trade compared to valuation-sensitive software or direct AI plays.
Overall signal: Continued bullishness on the durability and scale of AI investment. The physical buildout is still in early-to-mid innings, with power, cooling, and connectivity often cited as key long-term challenges.
In short, the AI infrastructure buildout theme — especially memory/semiconductor tools + data center construction/power/cooling/optics — is the clear winner in this list. It shows investors are betting on the hard work of actually building out the AI future, not just talking about it. This phase tends to favor companies with real capacity constraints and direct exposure to hyperscaler spending.
Note: I actively review Swing & Trend long trades in all of the above but especially CAT, GLW, URI, TER as these I recommended early & often. And I continue to say, “they haven’t done anything wrong.”
But timing the next turn for Hyperscalers is my focus right now, as is the play on dollar longs AND yields short. So I want to share my response with the members in response to Luke’s post above to shed some light on the timing of this very important ‘sector’ rotation:
