Seligman’s Paul Wick: Selective AI Plays & "Old Tech" Bargains
Veteran tech investor Paul Wick (Seligman Investments) argues that the AI trade is maturing and shifting away from pure chip hype toward undervalued software, infrastructure, and "power-adjacent" plays. Wick emphasizes public market valuations over "overheated" late-stage private ventures.
Top Investment Convictions
- Broadcom ($AVGO): Highlighted as a top "Trillion-Dollar" pick following a recent sell-off. Wick anticipates $79 billion in AI-related revenue over the next four quarters.
- Alphabet ($GOOGL): Favored for its superior free cash flow profile among the "Magificent Seven," aggressive stock buybacks, and leadership in autonomous driving via Waymo.
-
Bloom Energy ($BE): A high-conviction energy play. Wick views Bloom’s fuel cells as the fastest solution to AI's "power constraint," noting they can be deployed in six months—far faster than utility grid upgrades.
- Note: Wick projects 2026 earnings of $2–$3/share, significantly higher than the $1 consensus.
- Marvell Technology ($MRVL): Considered a "glass-half-empty" stock that has grown into its valuation.2 Wick cites a massive ASIC (custom chip) backlog with growth expected to accelerate through 2027.
- Wix.com ($WIX): A contrarian play in software. Wick argues fears that AI "vibe coding" will kill website builders are overblown, noting Wix’s strong free cash flow ($500M+) and its own strategic AI acquisitions.
Sector Strategy & Market Themes
| Theme | Wick's Outlook |
| AI Hype Cycle | Not a "dot-com" bubble. While the path will "zigzag," public markets remain more disciplined than private venture capital. |
| Enterprise Software | Currently viewed as "AI losers," but fears are overstated. Stability in corporate workflows will lead to a revaluation of laggards like Wix. |
| Power & Nuclear | Power is the #1 constraint for AI. While bullish on energy, Wick is skeptical of SMRs (Small Modular Reactors) like Oklo, citing long regulatory timelines. |
| "Neocloud" Risks | High-risk warning on commodity data-center firms (e.g., CoreWeave, Nebius) that lack proprietary intellectual property. |
The "Avoid" List
- Quantum Computing: Wick remains highly skeptical, labeling it "perpetually 5-10 years away."3 He views the commercial applications as trivial and the total addressable market as too small for significant returns.
- Late-Stage Private Tech: Cites "speculative red flags" in private valuations (e.g., Cursor at $29B) compared to profitable, cheaper public peers.