AI’s Great Divide: Who’s Winning at Goldman’s TMT — and Who’s Still Being Asked to Prove It?

Investors crowned AI infrastructure and chipmakers as the clear winners at Goldman Sachs’ Communacopia + Technology conference, while many software vendors faced blunt questions about monetisation. $ORCL ’s contract surprise and $AVGO ’s lofty targets reinforced an infrastructure-first narrative, leaving traditional software firms to prove AI delivers paying customers and durable revenue.

Key Points

1. $ORCL flagged a 359% jump in future contract revenue tied to an OpenAI deal.

1. Investor attention concentrated on $NVDA /OpenAI and data-layer plays like Databricks, Snowflake, and MongoDB.

1. Goldman’s semiconductor thesis focuses on AI GPUs/ASICs, HBM/advanced packaging, and EDA tools.

1. Investors repeatedly demanded clear, paying AI use cases from software vendors.

Lead and atmosphere: what happened — and how the room reacted

Goldman’s marquee TMT gathering made the market’s message plain: build the compute and data plumbing and investors will reward you.

$NVDA and OpenAI sessions filled the main hall and overflow rooms, a vivid signal of capital focus.

By contrast, conventional software vendors endured pointed investor queries demanding evidence of revenue tied to AI features.

Proof points: Oracle, data infrastructure, and monetisation examples

$ORCL ’s disclosure — a 359% jump in future contract revenue largely linked to an OpenAI deal — served as a headline proof point. Databricks announced a $1 billion financing round and said its AI product annualised revenue tops $1 billion, adding credibility.

Snowflake and MongoDB’s strong year-to-date performances underscored market preference for firms turning data into model-ready fuel. Google Cloud’s Thomas Kurian noted $GOOGL has already generated “billions” from AI services, a straight monetisation example.

Software monetisation and market response

Across panels, investors asked one blunt question: how do customers pay for your AI?

Twilio highlighted enterprise AI agent revenue from text-to-voice and conversational tools; Grindr pointed to AI-driven subscription upsells. But many legacy software vendors lack multi-year, contract-backed proof, leaving their market stories vulnerable.

The conference amplified a bifurcation: clear monetisation equals investor reward; fuzzy pilots bring valuation risk.

Semiconductor thesis and Broadcom’s ambition

Goldman laid out a concentrated chip bull case built on three pillars: AI compute (GPUs/ASICs), advanced packaging & HBM, and EDA tools.

Applied Materials emphasised HBM and advanced packaging equipment as long-term growth vectors.

Cadence said AI-assisted EDA adoption accelerates design productivity and budget penetration.

$AVGO pitched up to $120bn in AI revenue by FY2030, while Goldman’s nearer-term FY2025 estimate is about $20bn.

Equipment, EDA and the “behind-the-scenes” winners

Equipment makers and EDA vendors are the unsung enablers of the AI build-out.

Without advanced etch, packaging and design tools, chipmakers cannot meet cloud-scale performance and power goals.

Applied Materials highlighted GAA/BPD nodes and HBM equipment expansion as central to future growth.

Cadence noted cloud and systems customers now contribute materially to revenue, supporting sustained EDA demand.

Market reaction and price moves — who’s been rewarded so far

The market has rewarded infrastructure and data plays.

Snowflake and MongoDB’s rallies reflect investor confidence in monetisation trajectories. Databricks’ funding and revenue signals further validate investor appetite for data-layer plays.

Software firms without clear, paid AI use cases risk underperformance until they prove ARR growth tied to AI products.

What to watch and structural implications

Look for repeatable, multi-year commercial contracts and rising ARR explicitly tied to AI products as primary validation signals.

On the supply side, watch HBM capacity expansions, advanced packaging rollouts, major ASIC wins with cloud providers, and EDA adoption metrics.

If these indicators align, the capital rotation toward compute, memory, packaging, and design software looks durable.

That structural shift will influence valuations, M&A, and corporate strategy across the industry.

The final word: hype versus repeatable economics

Goldman’s TMT conference reinforced a practical investor mantra: hype is not enough — monetisation is mandatory.

The event made clear that today’s largest investor flows favour the economic plumbing of AI: chips, memory, packaging, and data infrastructure.

For software vendors, the imperative is to move from demos to dollars.

The winners will be those who deliver measurable ROI and recurring revenue at scale; the rest risk being left on the wrong side of a market realignment.