Is it time to buy tech, again? A flurry of good news from Broadcom may hold the answer

UBS expects 'tech+' revenue growth to accelerate to ~23% YoY in Q1 with earnings up ~30.4% vs 5.1% for the rest of the S&P 500, while the Nasdaq remains well below its record 23,958 (Oct 29, 2025) after a ~10% correction in late March. Broadcom signed a multi-year TPU supply deal with Google through 2031 and expanded Anthropic access to 3.5 GW of TPU-based AI compute starting in 2027 (up from 3 GW), as Anthropic reports a $30B revenue run-rate (vs ~$9B exiting 2025). Goldman and Wells Fargo analysts flag materially cheaper tech valuations and upgraded/positive stances, but geopolitical risk around Iran and higher oil prices keep the outlook volatile and defensive in the near term.

Analysis

Geopolitical risk has re-priced the optionality of a tech-led market recovery: when oil spikes persist for weeks, capital-intensive growth (ad monetization, consumer device upgrades) tends to lag while software and hyperscaler revenues prove stickier. That dynamic increases the value of vendors that monetize recurring cloud/AI consumption rather than one-off device cycles, which should translate to multi-quarter margin resilience even if headline equity multiples remain volatile.

Second-order winners are the integrated infra players that can sell a bundled stack (silicon + networking + software) into large-scale datacenters; they avoid the single-point-of-failure reorder cycle of fabless or pure-play suppliers and therefore capture a disproportionate share of incremental AI dollar per watt. Conversely, smaller vendors reliant on hyperscaler spot RFPs or on customers with weak balance sheets face lumpy revenue and inventory risk if hyperscaler pacing normalizes.

Key catalysts to watch by horizon: days–weeks — Iran headlines and an oil move >$90 would trigger risk-off and a re-test of recent lows; months — hyperscaler capex cadence updates and Q2 cloud guides will prove the sustainability of AI spend; 6–18 months — signs of compute overcapacity or falling ASPs for accelerators could compress valuations across semis. A dominant contrarian point: consensus is pricing AI as pure demand growth, not as a structural shift that will reallocate margin up the stack — if true, vendors with deeper software/IP will re-rate higher, while pure silicon commodity exposures will lag.

Key Decisions for Investors

• Long AVGO (equity or 12-month call spread): allocate 1.0–1.5% of portfolio to capture higher-margin infrastructure capture. Target 30–40% upside over 9–12 months; hard stop at -15%. Rationale: outsized ability to monetise bundled stack vs fabless peers.
• Pair trade — Long GOOGL + AMZN (equal $) / Hedge with 2% notional 3‑month S&P 5% OTM puts: timeframe 3–6 months. Reward: 20–35% if rotation into cloud accelerates; cost: ~2% for tail protection against geopolitics-driven drawdown.
• Relative trade — Long AVGO / Short MRVL (dollar‑neutral, 3–12 month horizon): target 20–30% relative outperformance if integrated supplier execution holds. Exit/stop: unwind if AVGO underperforms MRVL by 10% intraday or on material partnership reversal.
• Risk management — Buy 1–2 month S&P 5% OTM puts (or synthetic funded via selling 3% OTM calls) sized to cover 1–3% portfolio downside during headline risk windows. This is inexpensive insurance around immediate geopolitical tail risk and preserves upside exposure to a tech re-rate.