Salesforce's Marc Benioff says Google's Gemini 3 just blew past ChatGPT: 'I'm not going back'

Salesforce CEO Marc Benioff publicly said he is abandoning OpenAI's ChatGPT in favor of Google's newly launched Gemini 3, calling its gains in reasoning, speed and multimodal capabilities an “insane” leap after two hours of use. High-profile tech figures including OpenAI CEO Sam Altman and Andrej Karpathy have also signaled positive early impressions, underscoring Gemini 3’s role in intensifying competition among Google, OpenAI and Anthropic. The release positions Google and DeepMind as stronger competitors in the AI arms race and could influence enterprise AI partnership preferences and sentiment toward major AI platform providers.

Analysis

Market structure: Google/DeepMind and AI compute suppliers (NVDA, AMD) are the primary beneficiaries as improved model quality raises willingness-to-pay for enterprise deployments; expect Google Cloud share gains of 200–400 bps over 6–12 months in AI-heavy RFPs if Google converts pilot wins. Incumbents with exclusive OpenAI ties (Microsoft/Azure) face pricing and share pressure in enterprise LLM procurement, though existing contracts and integration lock-in will blunt near-term churn.

Risk assessment: Tail risks include a regulatory antitrust probe into Google/DeepMind (probability 10–20% over 12–24 months) or export-control driven GPU supply shocks that could widen hardware margins by 15–30% or stall deployments. Short-term (days–weeks) sentiment volatility will spike around benchmark disclosures and CEO endorsements; medium-term (3–12 months) depends on measurable enterprise deal flow; long-term (2+ years) hinges on model safety/regulatory frameworks and total addressable market expansion.

Trade implications: The demand shock favors AI-infrastructure equities and selective cloud exposure—anticipate NVDA revenue upgrades and GOOGL cloud contract announcements within 3–9 months; implied volatility on NVDA/GOOGL should rise 10–25% around product benchmarks. Use levered, time-boxed option structures to capture upside while capping downside given adoption lags and procurement cycles.

Contrarian angles: Consensus equates demo quality with immediate enterprise displacement; that overlooks migration costs (data pipelines, fine-tuning, security) that typically delay platform switches 6–12 months. Historical parallels (early Bard enthusiasm) show hype can be mean-reverting; regulatory backlash or safety incidents could quickly re-rate winners, creating asymmetric risk/reward for near-term longs.

Key Decisions for Investors

• Establish a 2–3% long position in GOOGL (Alphabet) sized to portfolio risk over a 6–12 month horizon; target +15–25% upside if Google converts AI pilot wins to enterprise contracts within 6 months, cut at a 12% loss or if no material deal announcements in 6 months.
• Allocate 3–4% to NVDA via a 3-month call spread (buy 10% OTM, sell 25% OTM) to express higher AI compute demand while limiting premium spend; take profits at a 25–30% return or stop at a 15% loss.
• Implement a relative-value pair trade: long GOOGL 2% / short MSFT 2% for 3–6 months to play platform share shifts; tighten position if Microsoft announces exclusive OpenAI renewals or if the pair diverges by >8%
• Buy a 3-month MSFT 7% OTM put sized 1% of portfolio as insurance against rapid Azure share losses or surprise OpenAI contract developments; unwind if put loses 50% of premium or if MSFT announces continued exclusive commercial terms.