Goldman Sachs is urging investors to shift their focus from Nvidia to “Phase 3” AI companies that are actively developing AI-enabled revenues and integrating artificial intelligence into their core business models. While Nvidia has recovered from its DeepSeek-related losses, the investment bank’s chief investment strategist David Kostin believes the next major opportunity lies with companies monetizing AI applications rather than infrastructure providers.
The AI adoption landscape is rapidly evolving. During the most recent Q4 2024 earnings season, 50% of S&P 500 companies mentioned “AI” in their quarterly calls, a dramatic increase from under 20% in 2023. However, actual implementation remains nascent—only 6% of US companies currently incorporate AI into their business models, with another 10% planning adoption within six months, according to US Census Bureau data. This gap between interest and implementation creates what Goldman Sachs views as a prime entry point for investors.
Goldman’s “Phase 3” classification distinguishes these companies from “Phase 1” (Nvidia) and “Phase 2” (AI infrastructure firms). Recent market performance supports this thesis, with AI beneficiaries outperforming infrastructure companies when indexed against the equal-weight S&P 500. Interestingly, DeepSeek’s emergence—which initially triggered market volatility—may actually benefit Phase 3 companies by dramatically reducing inference costs, the computational expenses required to run AI models in real-world applications.
Software companies dominate Goldman’s Phase 3 basket, representing 15 of 27 identified stocks, with an additional five from IT services. Sumali Sanyal, managing director at Xponance, notes that increased competition from new entrants like DeepSeek will “open up the opportunity set for many smaller companies” as costs decline. The firm sees particular adoption momentum in human resources, customer relationship management, and cybersecurity applications.
Nine of Goldman’s Phase 3 picks actually declined in 2024 while the S&P 500 rose over 23%, potentially offering value opportunities. The list includes MongoDB (down 43%), Adobe (down 25%), Snowflake (down 22%), and Zscaler (down 19%). These companies are projecting strong 2025 sales growth, ranging from 9% to 29%, as they integrate AI capabilities. GitLab reports customers seeing productivity improvements up to 50% with its AI tools, while Adobe surpassed $1 billion in its AI-related business book, and Snowflake now has over 1,000 deployed AI use cases with more than 3,200 accounts using its AI features.
Key Quotes
Those costs are going to come down dramatically and that’s going to open up the opportunity set for many smaller companies.
Sumali Sanyal, managing director and senior PM for systematic global equities at Xponance, explained how DeepSeek and similar competitors will reduce inference costs, democratizing AI access beyond mega-cap companies and creating opportunities for smaller firms to adopt and monetize AI technologies.
Going forward, the expectation is that the software and services companies will start adopting generative AI to enhance their existing models, design new industry applications, and start improving coding efficiency.
Sanyal outlined the expected trajectory for Phase 3 companies, emphasizing that software and services firms will lead AI monetization by integrating generative AI into their core offerings rather than building entirely new products from scratch.
Customers are realizing tangible value with it. One recently told us that 40% of their developers are saving more than 2 hours per week with Duo and another customer is seeing productivity improvements of up to 50%.
GitLab’s earnings commentary demonstrates concrete ROI from AI implementation, providing evidence that Phase 3 companies are delivering measurable business value rather than just experimental features, which supports Goldman’s investment thesis.
We hit several new milestones with our AI innovations, enabling us to add more than $2 billion in Digital Media net new ARR and surpass $1 billion in the ending book of business for Adobe Experience Platform and native apps.
Adobe’s earnings statement quantifies significant revenue generation from AI features, showing that established software companies can successfully monetize AI capabilities at scale, validating the Phase 3 opportunity with concrete financial metrics.
Our Take
Goldman Sachs’ Phase 3 framework represents a sophisticated evolution in AI investment strategy that recognizes we’re entering the monetization era. The counterintuitive insight here is that DeepSeek—initially perceived as threatening to established AI players—may actually accelerate the Phase 3 opportunity by commoditizing infrastructure and reducing costs for application-layer companies. The 44-percentage-point gap between companies discussing AI and those actually implementing it reveals massive untapped potential. What’s particularly compelling is that many Phase 3 stocks underperformed in 2024 precisely because they were investing in AI capabilities rather than immediately monetizing them. Companies like MongoDB, Snowflake, and GitLab are now positioned to capture returns on those investments. The concentration in software—particularly in HR, CRM, and cybersecurity—makes strategic sense as these are areas where AI can deliver immediate, measurable productivity gains. This isn’t speculative technology; it’s practical business transformation with quantifiable ROI, making the Phase 3 thesis more grounded than previous AI investment waves.
Why This Matters
This analysis signals a critical inflection point in the AI investment landscape, marking the transition from infrastructure buildout to practical application and monetization. For nearly two years, AI investing has been dominated by chipmakers and cloud infrastructure providers, but Goldman Sachs’ Phase 3 thesis suggests the value creation is shifting to companies that can demonstrate actual revenue generation from AI capabilities.
The timing is particularly significant given the current adoption gap. With only 6% of companies actively using AI despite 50% discussing it, early movers in the Phase 3 category could capture substantial market share before competition intensifies. The DeepSeek development, rather than undermining the AI thesis, may actually accelerate democratization by reducing barriers to entry through lower inference costs.
For businesses and investors, this represents a broadening of the AI opportunity beyond mega-cap tech stocks. Software companies integrating AI into existing products—particularly in HR, CRM, and cybersecurity—are showing tangible productivity gains and customer adoption. The fact that many Phase 3 stocks underperformed in 2024 while building AI capabilities could present attractive entry points for those believing in the monetization wave ahead.
Recommended Reading
For those interested in learning more about artificial intelligence, machine learning, and effective AI communication, here are some excellent resources:
Recommended Reading
Related Stories
- Wall Street Asks Big Tech: Will AI Ever Make Money?
- The AI Hype Cycle: Reality Check and Future Expectations
- Jensen Huang: TSMC Helped Fix Design Flaw with Nvidia’s Blackwell AI Chip
- How Companies Can Use AI to Meet Their Operational and Financial Goals
- Legendary Investor Jeremy Grantham Predicts Stock Market Crash and AI Bubble Burst by 2024