Prediction: This Artificial Intelligence (AI) Stock Will Be Worth 2 Times as Much by the End of 2026

Applied Digital says its contracted backlog has expanded to $36 billion, boosted by a new 15-year $5.2 billion hyperscaler lease covering 210 MW of AI data-center capacity. The company already has 100 MW operational at Polaris Forge 1, with another 150 MW expected in 2026 and a second campus ramping in 2026-2027. Despite a $100.9 million net loss and $2.7 billion of debt, the stock is framed as potentially doubling by end-2026 if investors re-rate it as a long-term AI infrastructure platform.

Analysis

APLD is transitioning from a development story to a contracted-capacity story, which matters because the equity multiple should migrate from near-term revenue skepticism toward backlog visibility and build-rate execution. The non-obvious second-order effect is that the market is likely to re-rate the whole liquid-cooled, high-density infrastructure niche before it fully believes APLD’s own free cash flow — meaning suppliers, power equipment vendors, and adjacent data-center developers can get sympathy bids even if their own bookings lag. CRWV is the clearest read-through beneficiary because it validates hyperscaler demand for specialized AI capacity, while NVDA benefits indirectly if accelerated deployment tightens the bottleneck on GPU time-to-training rather than chip availability.

The biggest risk is not demand; it’s financing and schedule slippage. A $36 billion backlog sounds de-risked, but the balance sheet still forces continuous capital-market access, so any wobble in debt spreads, equity issuance, or construction timelines can compress the multiple quickly. Over the next 3-9 months, the key catalyst sequence is: capacity coming online on time, leasing conversion into cash collections, and evidence that the contracted revenue can outgrow interest expense and capex drag; if either of the first two steps slips, the market will revert to valuing APLD as a levered developer rather than a platform.

Consensus appears underestimating how much optionality is embedded in the renewal stack and how quickly a large, verified backlog can attract passive and factor flows into a previously illiquid name. That said, the stock is already pricing in a meaningful portion of the success case; the asymmetry is better on pullbacks or through structured exposure than via an outright chase after a sharp run. The trade is less about whether AI demand exists and more about whether APLD can convert scarcity of powered, liquid-cooled capacity into durable per-share value before dilution or execution risk catches up.

Key Decisions for Investors

• Long APLD on a 2-6 week pullback rather than strength; prefer entry only if the stock digests recent gains and implied execution risk resets. Upside is another 20-35% on clean delivery, but downside is a fast 15-25% derating if any delay hits the 2026 build schedule.
• Pair trade: long APLD / short a basket of lower-quality AI infrastructure developers with weaker backlog conversion or more balance-sheet leverage. The setup works best over 3-6 months if the market keeps paying up for contracted megawatts while penalizing funding risk.
• Add CRWV on any post-print weakness as the most direct customer-validation beneficiary. Risk/reward favors a medium-term move higher if the market starts capitalizing hyperscaler demand for specialized AI capacity at a higher duration multiple.
• For high-conviction exposure, use APLD call spreads 6-9 months out instead of common stock. This captures rerating potential while limiting dilution and financing risk that could cap equity upside.
• Avoid chasing the move in NVDA off this headline alone; treat it as a secondary beneficiary, not a primary catalyst. Best use is as a hedge against an APLD-specific execution miss rather than as a direct directional expression.