GS Johnstone TMT Spec Sales Communacopia Technology Conference Day 1 wrap 9 Sep 2026

GS Johnstone - TMT Spec Sales Communacopia Technology Conference Day 1 wrap 9 Sep 2026 Sean Johnstone · Goldman Sachs · TMT Spec Sales Wed 9 Sep 2026, 2:33am ET

GS Communacopia + Technology Conference – Day 1

Below thoughts from the names that I saw – for today agenda and more detailed summaries from GIR please see -

Day 1 of Communacopia reinforced a broadly bullish view on AI infrastructure, software and enterprise adoption, with the central theme being the shift from experimentation to production use cases (its only Day 1 so lets see how rest of week shapes up). Across OpenAI, Alphabet, SAP, Snowflake and Databricks, the debate is moving from model capability to cost per task, governance, model choice and workflow automation, with enterprises increasingly looking for model-agnostic platforms that can route work to the best and cheapest model while protecting data. OpenAI and Google both sounded highly confident on demand and monetisation, while SAP (notable more upbeat than last year, investors will debate the multiple given the bounce), Snowflake and Databricks positioned themselves as critical enterprise AI control layers built around agents, governed data, consumption-based pricing and open ecosystems (will be interesting to see more software names this week as investor debate here certainly 2 way and asking how much more than can bounce without revenue acceleration). On infrastructure, Broadcom, Arista, Infineon and Nebius all pointed to strong AI demand and improving visibility, with constraints shifting toward supply, power, data-centre capacity and execution rather than end demand. The broader conclusion is that AI is accelerating across the stack — from semicap and networking to enterprise software, payments and personal agents (while not at the conference Instinct and personal AI agents getting some debate that caused online travel names to roll.) Also note later today Apple new launch with the new Apple CEO. Company Notes OpenAI

OpenAI bullish on both demand and monetisation, highlighting Astra as its largest-ever training run and pointing to frontier-level performance at materially lower cost. Customers are increasingly focused on cost per task completed rather than cost per token, while enterprise adoption is accelerating, agentic products now have 25m users, and usage intensity remains well ahead of peers. Despite compute constraints, particularly in voice, revenue per GPU is rising and unit costs are falling. Strategically, OpenAI sees cybersecurity as the biggest 12-month opportunity, with life sciences the longer-term value-based monetisation opportunity. Pricing is expected to shift from seats to consumption and ultimately outcome- based models, while consumer momentum remains strong at c.1bn users, with subscription upsell and ads already contributing meaningfully. Prometheus

Prometheus was deliberately light on detail, but the message was centred on AI for the physical economy — using AI as an engineering and manufacturing tool to build complex products faster. The company discussed the idea of an “artificial general engineer”, with likely early use cases in jet engines, rockets and

complex casting. The broader read-across is that life sciences is becoming an increasingly crowded AI opportunity, with both OpenAI and Google also repeatedly highlighting it as a major strategic focus. Certainly not the bearish near term catalyst for DSY I had expected. Alphabet

Alphabet emphasised that its AI advantage comes from a vertically integrated stack spanning TPUs, Gemini models, data, security and applications. Management argued this delivers superior price- performance and lets Google capture value across multiple AI monetisation layers. Enterprise adoption of Gemini is strong, with over 90% of Fortune 100 companies using Gemini Enterprise as an agentic platform to automate workflows and reason over private data. Google Cloud is scaling the opportunity through industry-specific solutions, system-integrator partnerships, proprietary data integrations and forward-deployed engineers helping customers build frontier AI use cases. SAP

SAP was notably more confident than last year, with Christian Klein sounding upbeat on…