As artificial intelligence reshapes the global economy, one of Wall Street’s most influential voices is sounding a clear alarm — and delivering a roadmap for where investors should place their capital next.
Ulrike Hoffmann-Burchardi, Global Head of Equities and Chief Investment Officer for the Americas at UBS Wealth Management, says the AI boom poses an existential threat to many software-backed businesses. Her prescription? Investors should pivot from “coders” (the bits-and-bytes world of pure software and services) to “builders” — the companies operating in the physical world that supply the atoms, infrastructure, and energy required to make AI possible.
Speaking on the sidelines of a financial conference in Miami Beach, Hoffmann-Burchardi told Bloomberg that UBS is actively “morphing portfolios from targeting ‘bits to atoms.’” The firm has downgraded the information technology and communication services sectors within the S&P 500 and is instead overweighting “the physical parts” of the benchmark — equipment makers, power generators, resources miners, industrial firms, and power producers.“This is about companies that build the physical infrastructure necessary to drive the modern economy,” she emphasized.
Why Energy Builders Are the Real AI Winners
The math behind the shift is straightforward and energy-intensive. Hyperscalers — Microsoft, Amazon, Alphabet, Meta, and others — are projected to pour roughly $700 billion into capital expenditures in 2026 alone, the vast majority earmarked for AI data centers, chips, and supporting infrastructure. That spending is exploding electricity demand.U.S. data centers already consumed about 4.4% of national electricity in 2023–2024. Forecasts from Lawrence Berkeley National Laboratory, S&P Global, and the IEA project that share could climb to 6.7–12% by 2028 and potentially higher by 2030, with some models showing data-center demand nearly tripling. In high-growth regions like Virginia and Texas, utilities are scrambling to interconnect hundreds of gigawatts of new load.
Every query, every training run, every inference at scale requires reliable, dispatchable power — much of it still anchored by natural gas, nuclear, and the upstream energy complex that keeps the grid humming.
That’s why traditional energy producers and service companies are perfectly positioned as the ultimate “builders” in the AI supply chain.
Proven Shareholder Returns: Energy’s Track Record vs. Tech’s Uncertainty
Unlike many high-flying AI software names trading at stretched valuations and facing disruption risk, the energy complex has a long history of returning cash directly to shareholders through dividends and buybacks — even in volatile commodity cycles.
ExxonMobil (XOM): 43 consecutive years of dividend increases. In 2025 alone, the company returned $37.2 billion to shareholders ($17.2 billion in dividends + $20 billion in buybacks) while generating industry-leading free cash flow. At current levels, the yield sits around 2.9–3.5%, backed by low-cost assets and structural cost savings targeting $20 billion cumulatively by 2030.
Chevron (CVX): 39-year dividend growth streak. Yield hovers near 3.9%, with management projecting double-digit free-cash-flow growth through the end of the decade at $70 oil. The recent Hess acquisition and major project start-ups further bolster the payout outlook.
Liberty Energy (LBRT): A nimble oilfield services leader now expanding aggressively into distributed power generation for data centers. The company recently declared its quarterly dividend of $0.09 per share and is capitalizing on the “speed-to-power” gap that hyperscalers face. Oil-service peers entering the modular power market (via gas turbines and reciprocating engines) have seen sharp outperformance precisely because they can deliver reliable electricity to AI projects faster than traditional utilities.
These firms aren’t just riding commodity prices — they are structurally benefiting from the AI-driven surge in power demand while maintaining disciplined capital allocation and generous shareholder returns.
NVIDIA and Hyperscalers: Part of the Build, But Not the Whole StoryNVIDIA (NVDA) remains the indispensable enabler of AI training and inference — its GPUs are the picks and shovels of the gold rush.
Hyperscalers’ massive capex directly flows through to NVIDIA’s data-center revenue, which already accounts for ~90% of the company’s top line.
Yet UBS’s framework places hardware and infrastructure in the “atoms” camp while flagging pure software and service models as more vulnerable to AI’s own disruptive force (e.g., coding agents that can replace or compress traditional software labor).
Investors can therefore own the full stack:
NVIDIA for the silicon edge,
Hyperscalers for the data-center buildout (and their own massive power consumption),
and the energy complex for the indispensable fuel that keeps the lights on.
Bottom Line for Energy Investors
The AI trade is no longer just about who writes the best code or trains the largest model. It’s about who can physically deliver the compute, the power, and the infrastructure at scale.UBS’s call to move from bits to atoms is a loud endorsement of the energy sector’s multi-year tailwind.
Companies like Exxon, Chevron, and Liberty Energy have already proven they can generate robust free cash flow and return it to shareholders — traits that become even more valuable in an environment where software margins face compression and capex-heavy tech names burn cash at record rates.
For investors seeking real, tangible returns backed by the physical realities of the AI economy, the message is clear: bet on the builders. The energy names that power the AI revolution are poised to deliver both growth and income for years to come.
Energy News Beat will continue tracking AI-driven power demand, hyperscaler capex updates, and energy-sector capital returns. Stay tuned.
Get your CEO on the #1 Energy Podcast in the United States: https://sandstoneassetmgmt.com/media/
Is oil and gas right for your portfolio? https://sandstoneassetmgmt.com/invest-in-oil-and-gas/
Be the first to comment