BST: Gain Pre-IPO Anthropic Exposure And A 6% Yield

BlackRock Science and Technology Trust (BST) has 7.1% of its portfolio invested in Anthropic, its second-largest holding after Nvidia, giving investors indirect pre-IPO exposure to the AI company. Anthropic is described as targeting a Q4 IPO and potentially a $1 trillion-plus post-listing valuation, which could create a meaningful revaluation catalyst for BST. The trust's portfolio repositioning increases its exposure to high-growth generative-AI businesses.

Analysis

The investable issue is not Anthropic’s prospective headline valuation but BST’s transmission mechanism: a private-mark revaluation affects only the marked value of a roughly 7% position, while the closed-end-fund discount/premium to NAV determines whether shareholders capture it. Even a 100% uplift in that holding adds only about 7% to NAV before fees, taxes, valuation-lag effects, and any widening of the fund discount. A trade therefore requires verification of BST’s current discount, the date/method of its private-company mark, and whether a public listing would impose lockups that defer realizable NAV.

NVDA is a weak direct proxy. A successful frontier-model listing would validate aggregate AI infrastructure spending, but Anthropic’s compute mix and cloud relationships make the marginal hardware read-through less clean than for firms more tightly tied to its hosting stack, including AMZN and GOOGL. The more important second-order effect is competitive: a liquid public valuation for a major model developer could reset expectations for AI monetization and pressure Microsoft/OpenAI and Google to sustain elevated model and capex spending, supporting the broader AI supply chain but potentially delaying cloud-margin recovery.

Consensus may be overestimating the scarcity value of indirect access. Public-market investors can price an anticipated listing well before it becomes a cash or NAV event, while CEF discounts commonly persist through portfolio wins. Over the next 1-3 months, independently verifiable filing, pricing, lockup, and BST NAV-mark disclosures are the relevant catalysts; over 6-18 months, inference revenue growth and model-training economics—not a listing itself—determine whether a premium valuation is durable. The thesis is falsified if BST’s discount fails to narrow after a confirmed valuation event, or if disclosed marks already reflect the implied public valuation.

Key Decisions for Investors

• Do not treat NVDA as a primary expression of this event; maintain AI-semiconductor exposure only on independent demand evidence. A confirmed Anthropic listing is a sentiment catalyst, not sufficient evidence of incremental NVDA revenue.
• Place BST on an event-driven watchlist rather than initiate blindly: consider a long BST / short QQQ hedge only if its discount to NAV is materially wider than its 12-month median and the underlying private holding is still marked below an independently supported transaction valuation. Target discount normalization, not the full headline valuation; exit if the discount widens by 5 percentage points or no filing/mark catalyst emerges within 90 days.
• For a cleaner second-order basket, monitor AMZN and GOOGL versus cloud peers following verified disclosures on training and inference commitments. Favor the provider showing accelerating AI revenue without a disproportionate step-up in capex; avoid extrapolating valuation benefits before quarterly capex and cloud-margin data confirm it.
• Require confirmation of a public filing, offering terms, lockup duration, and BST’s post-event NAV methodology before sizing any catalyst trade. Absence of these data leaves the timing and realizable value of the private stake indeterminate.