Ideas
S&P 500 range-bound until 7,000 breakout.
The S&P 500 has been stuck in a wide 6,000-7,000 box because there is no new AI catalyst and Big Tech has not yet clearly shown AI monetization. Taver sees downside as protected by ample liquidity unless there is a rate or liquidity shock, but he says a decisive break above 7,000, ideally with short covering, would trigger stronger upside volatility; near the top he would stay conservative and accumulate near the bottom of the range.
Core AI winner; buy dips below box.
Nvidia is the core listed AI winner and trades in a roughly $170-$190 box. Taver argues AI capex must send at least 40-50% of spending to Nvidia, its forward P/E is only in the low 20s versus a historical 30x-plus, and it has excess cash, rising buybacks, and M&A optionality. Even if revenue growth slows to 30%, he still wants to own it and would buy dips rather than chase strength.
Next AI shortage favors materials and equipment.
The next AI bottleneck may move beyond GPUs into CPUs, hard drives, and power, but Taver sees no appealing CPU stock and says power infrastructure has already rallied. He therefore favors semiconductor materials, components, and equipment, where valuations are volatile but many companies show actual numbers and 50% year-over-year growth; he would allocate only about 5-10% and diversify.
Small bets on cooling and optical.
Intel's CEO said copper/air cooling is no longer enough and the industry must move to liquid immersion cooling and optical/optical communications. Taver likes these AI infrastructure niches because actual revenue and 50% year-over-year growth are showing up, but he would diversify into them with only 5-10% position sizes because valuations are volatile.
Hidden AI semiconductor with doubling revenue.
Credo Technology, a smaller semiconductor-related company, preannounced that 2026 revenue will double versus 2025. Taver uses this as evidence that hidden, underfollowed AI semiconductor names with strong earnings guidance can produce outsized gains like SanDisk if investors do their research and hold through earnings.
Top AI winner with fortress cash flow.
Alphabet/Google is one of Taver's two preferred listed AI winners because its cash generation is strong. Its 100-year bond issue attracted huge demand, showing investors will fund its AI capex; he cites management's view that data-center investment can return five times the spend, and Google can pause investment for a year or two and still earn money, unlike Oracle, which is tied to loss-making OpenAI.
Oracle AI tied to loss-making OpenAI.
Oracle is a concern relative to Google because its AI/data-center buildout is tied to OpenAI, which remains deeply loss-making and may not reach profitability until 2029-30. That makes Oracle's AI-related debt and capex exposure riskier, even though its bond demand has been strong.
Mega IPOs threaten mega-cap liquidity.
Mega IPOs in 2025-26, including SpaceX, xAI, OpenAI, and Anthropic, could absorb enormous liquidity and make it harder for the S&P 500 to break above 7,000. If a giant such as SpaceX enters the Nasdaq 100, existing mega-cap technology names like Nvidia, Apple, Google, and Tesla may be mechanically sold to make room, so Taver would not indiscriminately accumulate mega-cap tech this year and would treat those liquidity-driven sell-offs as buying opportunities.
AI power demand supports infrastructure.
AI power demand is expected to create power-infrastructure shortages. Although the area has already rallied, Taver later names power infrastructure as one of the non-overextended areas to accumulate through split purchases while managing cash around volatility.
This 3PRO TV (삼프로TV) video, published February 09, 2026,
features Taver
discussing SPY, NVDA, Semiconductor materials/components/equipment, Liquid cooling / immersion cooling, Optical communications, CRDO, GOOGL, ORCL, XLK, QQQ, Power infrastructure.
9 trade ideas extracted by AI with direction and confidence scoring.
Speakers:
Taver
· Tickers:
SPY,
NVDA,
Semiconductor materials/components/equipment,
Liquid cooling / immersion cooling,
Optical communications,
CRDO,
GOOGL,
ORCL,
XLK,
QQQ,
Power infrastructure
