Solar to dominate energy by 2035, but AI data centers will keep fossil fuels in business

BloombergNEF expects solar to become the largest source of power within the next decade, with solar generation projected to outcompete coal and natural gas on cost and reach more than twice natural gas output by 2050. The report also ties incremental electricity demand to AI and data centers, forecasting an extra 1 terawatt of utility-scale solar, 370 gigawatts of natural gas, and 110 gigawatts of coal for data-center load growth. The outlook is broadly bullish for solar, batteries, and hybrid renewable projects, while underscoring continued demand for fossil backup generation and the strategic impact of energy imports and geopolitics.

Analysis

The key market implication is not simply that solar takes share, but that power procurement becomes an operating-cost arms race for AI infrastructure. Data center developers will increasingly optimize around the cheapest marginal kilowatt-hour, which means the real winners are not just module makers but firms that can bundle generation, storage, interconnection, and software into a dispatchable product. That shifts value upstream and downstream of pure PV hardware toward grid equipment, inverters, transformers, and control systems, while commoditizing standalone generation over time.

A second-order effect is margin compression in daytime power markets that forces capital toward hybridization. As solar penetration rises, the economic value of batteries should improve faster than installation volumes suggest, because the spread between midday and evening power prices becomes the relevant monetization metric. That creates a favorable setup for storage enablers and for utilities or IPPs with existing interconnection queues, since the bottleneck becomes permitting and grid access rather than panel supply.

The contrarian read is that the market may be underestimating how much of the incremental demand will be met by gas on reliability grounds, even in a low-cost solar world. AI load is highly concentrated, latency-sensitive, and politically visible, so buyers will pay for firm capacity premiums; this supports gas midstream, peakers, and turbocharged LNG supply chains longer than the clean-energy narrative implies. The biggest risk to the renewable thesis is not cost, but integration: if transmission delays and curtailment rise faster than storage deployment, the winners will be the firms that own dispatchability, not the cheapest electrons.

For equities, the article is modestly bullish for GOOGL insofar as large hyperscalers with balance-sheet strength can arbitrage energy procurement and lock in long-duration PPAs; for Ford, the storage angle is more optionality than core upside. The bigger trade is to own the picks-and-shovels tied to grid buildout and to avoid pure-play solar manufacturers if pricing remains under pressure. The time horizon is 12-36 months for the storage/reliability trade to inflect, while the solar commoditization trend is already well advanced.

Key Decisions for Investors

• Go long GOOGL on a 6-12 month horizon via common or 12-18 month calls; thesis is that AI-driven power scarcity strengthens hyperscaler pricing power and justifies pre-purchasing clean, dispatchable capacity. Risk/reward: moderate upside with lower fundamental risk than most energy plays.
• Initiate a relative-value pair: long grid equipment/infrastructure names versus short a basket of pure-play solar manufacturers over 6-9 months. The setup favors firms exposed to interconnection, transformers, inverters, and storage integration over module pricing, which should stay competitive and volatile.
• Buy downside-hedged exposure to gas flexibility assets through midstream/LNG or utility peakers for 12-24 months; the market is likely underpricing the persistence of firm capacity demand from data centers. Use this as a hedge against over-owned renewables.
• Add F as a long-dated optionality position only, not a core fundamental long; treat the storage business as a call option on scale rather than near-term earnings contribution. Prefer staged entries after market pullbacks.
• Avoid chasing standalone solar equities after the initial rerating; if desired, express the theme through hybrid developers or storage-integrated platforms where the moat is grid access and dispatchability, not panel ASPs.