A Strategist Says Big Tech Is ‘Completely Underappreciated' With a PEG Under 1 as Everyone Bails

Venu Krishna argues the Mag 7 ex-Tesla trade is mispriced: a PEG <1 with ~30% Q1 earnings growth and multiples cut 5–10%, while lifting his S&P 500 target to 7,800 (based on 21% earnings growth this year). He highlights chip/AI visibility with hyperscaler capex projected at $1.2T by 2028 (~$250B above consensus), citing Alphabet Q1 revenue of $109.9B (+21.8% YoY) and Google Cloud growth of 63% with backlog over $460B. Key risk is higher-for-longer rates—10Y Treasury at 4.38% and Core PCE in the 90.9th percentile—because discount rates could compress valuation of distant AI earnings.

Analysis

The cleanest read is not “buy tech” but “buy the de-rated cash generators that fund the AI buildout.” If hyperscaler capex stays locked in, the second-order winner is not just the chip complex; it is the platform owners that can still compound earnings while absorbing the investment cycle, because the market is currently pricing them like mature software utilities rather than scarce growth assets. That makes GOOGL and META the best asymmetry: they have the most room for multiple repair if earnings stay intact, while their AI monetization is still underappreciated relative to the spend headline.

NVDA is still a structural beneficiary, but the setup is less asymmetric than the market assumes because the stock already discounts a large share of the capex runway. The more interesting risk is that higher server/memory/storage inflation keeps real rates sticky, which would hit the longest-duration names first; in that regime, AAPL and AMZN are more vulnerable than GOOGL/META because they combine richer multiples with less obvious near-term AI monetization per dollar of incremental investment. Over 1-3 months, the trade is about style reversal and positioning; over 6-18 months, it is about whether AI capex becomes a margin tax on the ecosystem or a durable demand engine.

The consensus is probably underweight the durability of hyperscaler spend but overweight the idea that every AI beneficiary deserves the same multiple. If yields drift higher, the market may keep rewarding the closest pick-and-shovel exposure while punishing the “buying the picks” names, which argues for relative value rather than outright beta. The thesis breaks if 10-year yields move decisively above recent highs or if next earnings/capex guides show any hesitation in cloud and platform monetization; that would turn this from a rerating story into a slow multiple compression trap.

Key Decisions for Investors

• Initiate a long GOOGL / short AAPL pair trade over the next 1-2 weeks; thesis is that the market is paying a premium for slower growth and underpaying for the cheapest large-cap AI compounder. Stop if AAPL reaccelerates services growth or if GOOGL cloud/backlog momentum fades.
• Add META on weakness as a 3-6 month mean-reversion trade; it has the best combination of earnings leverage and valuation reset among the platform names. Falsifier: any guide that shows AI capex rising faster than monetization, which would cap rerating.
• Use NVDA as a tactical not strategic long only on dips, not breakouts; the capex visibility supports the stock, but the easy multiple expansion may already be behind it. Prefer call spreads over cash equity if entering above recent highs to limit downside from rate shocks.
• Avoid chasing AMZN at current levels unless AWS growth sustains at the current inflection for another quarter; it is the most capex-intense balance between monetization and reinvestment. If rates move higher, AMZN is the first platform name I would trim.
• Set a watch trigger on the 10-year above 4.5%; if breached, reduce long-duration tech exposure and rotate toward GOOGL/META versus AAPL/AMZN rather than holding the basket outright.