Live AI is the highest-risk recurring moment in any major company’s annual calendar today. And almost no one outside a handful of production teams has a clear framework for how to think about it.
Most of the conversation about AI demos focuses on what the AI does. The harder question is how the production around the demo reinforces its value. The old playbook was designed for deterministic software – scripted, rehearsed, locked down. And, simply put, agentic AI doesn’t work that way.
The product is unpredictable. The demo runs once – with the share price moving in real time. The production team can’t fully constrain the output. And the audience is watching with their finger on the share button. Get it right and you reset the equity narrative. Get it wrong and you become the case study in someone else’s analysis.
Google has been the case study before. In February 2023, a single live Bard demo answered one question incorrectly about the James Webb Space Telescope. By the next day’s close, Alphabet had lost roughly $100 billion in market capitalization.
In a few days, Sundar Pichai will walk back on stage at Shoreline Amphitheatre for Google I/O 2026. (The keynote begins at 10:00 AM PT on Tuesday, May 19.) Reports point to a major Gemini model update, a new agentic AI assistant called Gemini Spark, Android 17, and the unveiling of Aluminium OS. Some of these will demo well in a controlled environment. Others will require live, multi-step, agentic AI performance in front of a global audience. Fortunately, the production discipline that distinguishes a credibility-building demo from a market-moving disaster has been developed largely in public, by Google itself, over the past eight years.
And to interpret Google I/O 2026, you have to understand the framework Google has been building since 2018.
For decades, corporate live demos followed a simple production logic. You wrote the script, ran the rehearsals, and planned for what could go wrong. A Salesforce dashboard demo. An iPhone software walkthrough. An Adobe Creative Cloud feature reveal. The product behaved predictably because the demo was, in essence, a high-fidelity rehearsal performed live.
Live AI demos break that model in four specific ways.
Agentic AI works differently. You give the system a goal, and it decides how to get there. The production team can guide the prompt but not the path. How and what the model produces between input and output isn’t fully predictable. That’s both the promise of agentic AI – and the production risk.
Even with the same prompt, the model can produce a different response. Sometimes the difference is small. Sometimes it’s significant. Sometimes the answer is just wrong. Until the demo actually runs in front of the audience, the production team can’t know exactly what the model will say.
A multi-step AI agent might take 45 to 90 seconds to finish a task. On a live stage, 45 seconds of an AI “thinking” is a production crisis. The audience disengages. The camera has nothing to cut to. The speaker has to fill the silence. A traditional software demo can choreograph every second. An AI demo can’t.
A traditional software demo failure is embarrassing. A live AI demo failure is a stock-moving event. The Bard incident wasn’t a fluke – it was the first clear signal of a new category of risk. Since 2023, more than one AI company has watched its share price move on the strength of a single live demo.
For us, these aren’t quirks to manage. They’re a structurally new production category. And they require a fresh approach.
No company has confronted the live AI demo problem longer or more publicly than Google. The framework isn’t written down anywhere, but it is present in the production decisions Google has made over the past eight years – including the ones it learned the hard way.
Five moments define that evolution.
At I/O 2018, Google played pre-recorded phone calls in which its Duplex AI assistant booked a hair appointment and called a restaurant. Sundar Pichai introduced the recordings as “the Google Assistant actually calling a real salon.” The AI sounded so human, complete with “um” and “uh” verbal tics, that the demo went viral within hours.
Then came the questions, first raised by Axios: why didn’t the businesses identify themselves when they picked up? Why was there no ambient noise? Were these genuine real-world interactions or carefully staged recordings? Google declined to provide the names of the businesses or confirm whether the calls had been edited. The credibility question presented a separate ethics issue: should an AI identify itself as AI when calling a human?
The lesson: When an audience can’t verify whether what they’re seeing is real, the demo creates more doubt than it resolves. Specificity, transparency, and verifiable detail are essential to earning trust and demo credibility.
Google held a small launch event for Bard featuring a promotional clip in which the AI gave a factually incorrect answer about the James Webb Space Telescope. Astronomers noticed within hours. Alphabet shares dropped about 7.7% the following day, erasing roughly $100 billion in market value.
The lesson: Every public AI output is a public statement. Fact-checking demo content isn’t a marketing task – it’s a risk management one.
Google released a six-minute video billed as “Hands-on with Gemini: Interacting with multimodal AI.” It appeared to show the model engaging with images, drawings, and a continuous spoken conversation in real time.
Within 24 hours, Bloomberg’s Parmy Olson reported that the demo had been constructed differently than the video implied: Gemini wasn’t responding to spoken voice or live video at all. Google’s team had fed the model still image frames and text prompts, then added the voice narration afterward in post.
Google’s disclaimer in the video description noted only that “latency has been reduced and outputs have been shortened.” The narrative shifted from “Gemini is remarkable” to “Google misrepresented Gemini.”
The lesson: The line between live demo and marketing video has to be transparent. The credibility hit when an edit gets exposed is worse than the lift from a polished demo.
At I/O 2024, Google unveiled Project Astra – its real-time multimodal assistant – with a deliberate production move that broke from prior patterns. The demo aired as two continuous takes, one on a Pixel phone and one on a prototype pair of smart glasses.
The signal to the audience was clear: Google hadn’t cherry-picked the responses. The model was handling a stream of inputs in real time, rough edges and all. Around that demo, Google also began explicitly labeling other AI segments as recorded or aspirational rather than implying everything was live. The change was subtle in execution but marked a significant milestone. Google had stopped trying to make everything feel live and had started telling the audience exactly what they were watching.
The lesson: Labeling the kind of demo you’re showing is the first rule of demo credibility. The audience will forgive almost any production choice if they know what choice you made.
At I/O 2025, Project Astra moved from research demo to shipping product, powering new experiences in Search Live, the Gemini app, and third-party developer tools. The production decision here was as significant as the engineering one.
Having spent 2024 carefully framing Astra as a live, unscripted experience, Google could now invite the audience to use the same capability themselves. The demo and the product had become indistinguishable, which is the highest form of credibility a live AI demo can earn.
The lesson: When the AI demo eventually becomes a product launch, the production discipline that surrounds it becomes the foundation for long-term trust.
That’s eight years of drafting the playbook, paid for in public embarrassment, market cap, and corrective communication. It distills into a single principle: the audience doesn’t need the demo to be perfect – but they need to know exactly what kind of demo they’re watching.
Which brings us to the framework.
Most companies treat a live AI demo as a binary – either it’s live or it isn’t. The actual production reality is a spectrum, and failure in almost every public AI demo controversy comes down to misalignment between which category the audience thought they were watching and which one was actually being staged.
There are five distinct ways to stage a live AI demo. We call them the Demo Modes – a five-category framework for live AI production
The AI runs in real time during the event. No pre-staging. No predetermined output. The speaker delivers an input, and the audience watches the response unfold in real time. Highest credibility, highest production risk. The two continuous-take Project Astra demos at I/O 2024 were the clearest recent example of Mode 1 done well.
The AI is running in real time, but inside a controlled environment. The prompts are curated, the use cases are scoped. The model is genuinely working, but the production team has narrowed what it might be asked to do. Mid-high credibility when the framing is transparent, mid risk. Most enterprise software AI demos today are Mode 2 whether the company says so or not.
The AI completed the task minutes or hours before the event. The audience watches the playback of an actual real run – including any imperfections – with the speaker explicitly framing it as such: “We ran this just before walking on stage. Here’s what it produced.” Mid credibility when disclosed, low risk. This mode is dramatically under-used. Done well, it captures most of the trust of a live demo while significantly lowering the risk of failing on stage.
A polished video of the AI performing a task, clearly labeled as recorded. Low credibility for capability claims but high credibility for visual production quality. The Gemini Hands-On video would have qualified as Mode 4 if Google had labeled it that way. The controversy emerged because it wasn’t.
Explicitly framed as “what’s possible,” “what we’re building toward,” or “where this is headed.” It’s a preview of where the product is going – not proof of what it can do today. Lowest credibility for capability claims, but useful for setting vision. Google should have framed the Duplex demo at I/O 2018 this way. Instead it was framed ambiguously enough to read as Mode 1.
The Demo Modes aren’t a ranking. They’re a set of choices. A keynote can deliberately mix categories – Verified Live for the headline demonstration, Constrained Live for the enterprise capability, Pre-flight Live for the agentic workflow, Pre-Recorded for the partner integration, Aspirational for the long-term roadmap.
The discipline isn’t picking the “best” mode. The discipline is making sure the audience knows which one they’re watching.
The framework becomes most useful as a real-time reading tool. Here’s how it applies on Tuesday, May 19.
Reports point to a major Gemini model update at the keynote – whether labeled Gemini 4 or a 3.x successor. The production question is which mode Google chooses for the headline demonstration. Verified Live (Mode 1) would be the most confident move – signaling that Google trusts the new model to perform outside of a controlled environment. Constrained Live (Mode 2) would be the more cautious choice. If Google frames the demo as anything other than Mode 1 or 2, that’s a signal worth noting.
Leaks point to a new agentic AI assistant called Gemini Spark – designed to work autonomously across apps, emails, calendars, and websites. Booking flights. Managing email. Filling out forms. This is the most production-risk kind of demo a company can stage today, because every action links to the next, and a single failure cascades across all points. The production decision is whether to demo Spark’s full workflow live (Mode 1 – high risk, high reward), to scope it tightly (Mode 2 – safer, less impressive), or to compress the experience via Pre-flight Live (Mode 3 – the team runs it just before the event and acknowledges it openly). Watch for the speaker’s framing language at the moment of the reveal. If they say “we ran this just before walking out,” that’s Mode 3 done well. If the demo cuts cleanly between steps without acknowledgment, the production team has chosen polish over transparency.
Every live AI demo at I/O 2026 will have some friction. Latency. A response that lands awkwardly. A model output that’s correct but visually unimpressive. The production decision is whether to absorb that friction visibly (the Astra 2024 approach) or to edit it out (pre-2024 approach). The Astra approach is the more mature move. Watch for it.
I/O 2024 introduced explicit labeling. I/O 2025 refined it. If I/O 2026 makes the live vs. recorded distinction even cleaner – graphics, lower-thirds, verbal framing – that’s Google institutionalizing their framework. If the line blurs again, that’s a regression worth flagging.
The most consequential audience at I/O 2026 isn’t the developers in Shoreline. It’s the institutional investors evaluating Google Cloud’s AI revenue trajectory. Pichai disclosed at Cloud Next 2026 that just over half of 2026 ML compute investment will go to the Cloud business. The Cloud demos at I/O have to translate that capex into a credible product story. Watch how those demos are categorized. Constrained Live with enterprise customer logos as visible validation carries weight. Aspirational framing doesn’t.
Anyone who watches I/O 2026 with the 5 Demo Modes framework in hand will likely walk away from the keynote with a deeper understanding of these products and features than the reader who watches for product news alone.
The 5-Mode framework names the demo categories, but the execution lives in the production decisions that distinguish one category from another. Four of them carry disproportionate weight.
Mode 1 (Verified Live) requires real-time agility. The speaker has to be ready to narrate whatever the model produces – including responses they’ve never seen before. That’s a different kind of prep than walking through a rehearsed click sequence. Pichai’s comfort with live AI moments is a production advantage Google has built over years. Most CEOs aren’t there yet.
Every live AI demo needs a written set of fallbacks: if the model produces a problematic response, what does the speaker say next? If latency drags on, where does the camera cut? If the demo fails entirely, how does the show move on without acknowledging it? The audience never sees the contingency. They only see the recovery. The Bard launch failure wasn’t a demo failure – it was a contingency-planning failure. The factual error was visible in promotional materials before the event. Better fact-checking should have caught it.
When a live AI demo is processing, the camera has to go somewhere. A cut to the speaker carries one signal. A cut to crowd reaction carries another. A cut to a product graphic carries a third. Each choice tells the home audience something different about whether to trust what’s happening. At I/O scale, this requires a director, multiple operators, and pre-planned camera blocking for every demo segment.
Mode 3 demos don’t happen by accident. They require the production team to actually run the demo backstage, capture the output, and have it ready to play back within minutes of the live moment. That’s a second production happening at the same time as the live event. Most companies don’t budget for it. The ones that do have a tool the others don’t.
These decisions aren’t decorative. They’re the difference between a demo that builds credibility and one that costs market cap.
Google I/O 2026 isn’t an isolated event. It’s the first in a three-week window that includes Microsoft Build (June 2–3) and Apple WWDC (June 8–12). All three companies will stage live AI demos. All three will face the same production decisions. And by the end of June, the industry will have its first complete data set for how the leading public AI companies are navigating the new production risk.
And others are watching them closely. Salesforce Dreamforce in September will stage Agentforce demos. Workday, ServiceNow, Adobe, and every major enterprise software company will demo agentic capabilities at investor moments over the next year. The companies that have a clear live AI demo framework will appear more credible than those with technically superior AI but worse production discipline.
That’s the broader implication. When every public company is staging live AI, the production discipline around the demo becomes part of the equity story itself. Not just for the AI labs – for any company whose narrative depends on showing product capability and evolution.
The 5 Demo Modes aren’t a prescription. Different companies, different audiences, different products will call for different combinations. What every company needs is the vocabulary to make those choices on purpose – not by accident.
Google has been learning that vocabulary in public for eight years. The lesson the rest of the industry has yet to fully absorb is that the question isn’t whether to demo live. It’s whether the production team is ready to handle what happens when you do.
The high-wire act is permanent. The model for walking it is still taking shape. Google I/O 2026 is the next big stage.
That’s the work worth investing in. It’s also the work Cardboard Spaceship builds for clients navigating the moments that matter.
The most consequential moments in modern corporate communications now run on live AI. Whether you’re preparing for an Investor Day, a product launch, a developer event, or an investor moment that includes an agentic demonstration, the production decisions you make now will define how the market reads your capability when the moment arrives.
Let’s start a conversation →
OpenAI and Anthropic haven’t filed for an IPO. But the production architecture of their eventual roadshows is already taking shape – in plain view, if you know what to look for.
DevDay 2025 looked like a developer conference. The 45-minute fireside chat between Sam Altman and Jony Ive, Apple’s former Chief Design Officer, looked like something else. To us, it played as a moment built for the institutional investors who would parse it on YouTube the next morning. Hardware ambition. Design pedigree. A three-year collaboration framed as a thesis paragraph for a future S-1.
From a production lens, that fireside reads less like a developer feature – and more like a cap-table signal.
The same pattern is playing out across the AI sector. Anthropic launched Code with Claude in May 2025 and expanded it across San Francisco, London, Tokyo, and Washington, D.C. The tour read as measured and technically rigorous, less like dev relations and more like the discipline of a public company running its narrative arc.
Audience composition appears to have shifted as well. The staging looks heavier. And the production decisions; line-of-sight blocking, fireside choreography, demo handoff timing, partner sequencing – increasingly resemble the staged moments a public company runs the year before it files.
We’re not predicting IPO timing. But we are saying the rehearsal is happening in plain view, and the production choices these labs are making now are the same ones every IPO candidate makes when they want institutional money to recognize them on sight.
The class of 2024–2026 – Reddit, CoreWeave, Klaviyo, Rippling – gave us a fresh dataset on which roadshow production decisions land and which don’t. Read that dataset alongside what’s happening at DevDay and Code with Claude, and the playbook for the next mega-IPO begins to write itself.
Here’s what we’re seeing.
Most coverage of DevDay 2025 focused on the announcements: GPT-5 Pro and Sora 2 in the API, ChatGPT Apps with Zillow, Booking.com, Target, Figma, Expedia, Uber, Instacart, OpenTable, DoorDash, and Peloton. AgentKit for autonomous workflows. Codex Slack integration.
The product news is the obvious headline. The interesting questions are about what the staging communicates.
At events of this scale, production choices rarely feel decorative. Instead, they tend to be directional. In our experience, every decision (who shares a stage, what experiences attendees walk through, which conversations linger in the memory) functions as a deliberate signal; first to the room, then to the much larger audience that catches clips, photos, and analyst recaps for weeks afterward. The question isn’t whether OpenAI staged DevDay 2025 carefully. It’s what the staging reveals about the story they want to tell.
Three production choices stand out.
When Altman sat down with Jony Ive, the staging seemed to communicate something specific to the financial world: this isn’t just a software company.
There’s hardware ambition, design pedigree, and a partner whose work shaped the iPhone, iMac, and Apple Watch. The chat felt unscripted enough to seem intimate, and structured enough to land its key points. Ive said his creative team’s purpose “became clear” with the launch of ChatGPT. Altman framed the collaboration as a three-year arc.
For developers, that’s a curiosity. For institutional investors evaluating whether OpenAI commands a hardware-software stack story, that plays as a thesis paragraph. Worth noting: the conversation wasn’t even livestreamed. It went up on YouTube hours later. To us, that distribution choice reads as confidence. The most consequential moment of the day didn’t need to compete for live attention.
A “cozy mini-theater with popcorn” featuring AI-generated short films sounds like a fun side-activation. Look closer. It’s a positioning argument. Sora moved from research preview to API offering at this event – a defining moment of value that extended far beyond a product demo.
The mini-theater turned attendees into an audience, not testers. Watching AI-generated film with popcorn in hand frames Sora as cinema. Watching it on a laptop frames it as a tool. The production choice tells the market which one OpenAI wants Sora to be.
That’s the same staging logic Reddit used in 2024 when it leaned into community as the equity story rather than DAU metrics. The product becomes the experience. The experience becomes the narrative.
The DevDay 2025 lineup looks deliberately chosen.
Each speaker delivered real content, and each one provided OpenAI with powerful, specific validation.
Stitched together, these three choices read as a single argument: this is a company with consumer scale, hardware ambition, and ecosystem depth. Whenever OpenAI eventually files, the components of the equity story are already on stage and evolving right in front of us.
Anthropic took a notably different production approach. Code with Claude launched in May 2025 as a single-day, hands-on conference at The Midway in San Francisco, then expanded into a multi-city series across San Francisco, London, Tokyo, and Washington, D.C. Where DevDay leans into spectacle, Code with Claude leans into rigor. Three choices stand out.
OpenAI sold $650 in-person tickets to anyone who clicked. Anthropic took applications and curated the room. To us, that’s a meaningful production decision.
Curating the audience signals to the financial world that the company controls who sees its developer relationship up close. It also implicitly positions Claude as a premium tool for serious builders rather than a consumer phenomenon.
For institutional investors evaluating where Anthropic sits in the AI stack, the application gate is exactly the kind of credibility signal that filters retail froth out of the room. That’s a strategic choice.
The decision to expand from a single SF event into a four-city tour mirrors how a public-company roadshow actually moves: deliberate geographic coverage, repeat performance discipline, the same narrative delivered to different markets.
To our eye, that’s not a developer marketing choice. That’s the rehearsal of an institutional travel pattern. Anthropic ran the SF, London, Tokyo, and DC sequence in 2025 and is following with additional cities in 2026.
The geography choice is itself a story. London for European enterprise. Tokyo for the kind of legitimacy that only comes from showing up in Asia’s most established tech and finance market. DC for policy and regulatory presence. Each city carries a specific part of the narrative to a specific audience.
AWS hosted official Code with Claude watch parties as satellite events. Most coverage treated this as a technical convenience. To us, it reads differently.
Look at it from a production standpoint: a major cloud hyperscaler is volunteering its physical and digital infrastructure to extend the reach of an Anthropic-branded event. AWS is Anthropic’s largest cloud distributor and a strategic investor with billions committed to the partnership. By hosting watch parties, AWS publicly performs that alliance – turning a developer event into a visible signal of institutional backing.
That’s the kind of credibility and distribution muscle most pre-IPO companies have to pay for. Anthropic has it built in.
Where OpenAI’s staging emphasizes scope and consumer reach, Anthropic’s emphasizes curatorial access, depth, and enterprise readiness. The cap-table signal differs accordingly. OpenAI seems to be telegraphing a story about scale, hardware, and platform breadth. Anthropic seems to be telegraphing a story about discipline, enterprise traction, and infrastructure partnerships. Both are legitimate pre-IPO postures. They simply imply different equity stories – and likely different institutional investor profiles when the filings eventually land.
To know what the AI labs are rehearsing, look at the recent IPO class. Four roadshows in particular produced distinct production case studies.
Reddit gave away 8% of its IPO to its users. 1.76 million shares allocated to 75,000 of its most active Redditors and moderators. Karma score decided who qualified. No lock-up period.
It was an unusual move and a deliberate one. Reddit’s equity story depended on community. A slide deck couldn’t tell that story; allocating real shares to real users could. Steve Huffman didn’t have to argue that Reddit had a unique relationship with its users. The directed share program was the argument.
The lesson: when your story depends on community, your investor day staging has to make the community visible, not just cite it in the deck. Reddit priced at $34 and opened at $47, a 38% first-day pop. The S-1 flagged real risk in the structure (no lock-up could amplify volatility, and some Redditors actively organized against the IPO), and the narrative architecture worked anyway. Users-as-shareholders is now an established roadshow technique that Uber, Airbnb, and Cava had used in lighter form.
CoreWeave downsized its IPO from $2.7 billion to $1.5 billion the day before launch. Institutional demand had softened. NVIDIA reportedly stepped in with a $250 million anchor purchase to get the deal across the line. CoreWeave priced at $40 on March 28, 2025, and the first day of trading closed flat.
The roadshow ran into a recognizable problem. The pitch was infrastructure-heavy and visually abstract: GPU capacity, liquid cooling, Kubernetes-native architecture, 250,000 GPUs across 32 data centers. All accurate. None of it tangible to a public-market investor who has never set foot in a data center.
The lesson: data center stories need physicality on stage. Slides and spreadsheets make “AI infrastructure” sound like a commodity. The tools that turn an abstract category into an investable thesis (data hall walkthroughs, rack density comparisons, heat-and-power math made visual) only work when the production team builds them in. CoreWeave eventually went on a 250%+ tear post-IPO once the market understood the story. The challenge wasn’t the company. It was the translation of the narrative – both visually and experientially.
Klaviyo was the first SaaS IPO in nearly two years. The market was frozen. Marketing automation isn’t a category that excites public-market investors. CEO Andrew Bialecki later called it an “IPO winter.”
So Klaviyo led with the founder. Bootstrapped origin. MIT-trained engineer. $100M strategic investment from Shopify. 119% net dollar retention. 51% YoY growth. GAAP profitability. Rule of 75 metrics. Bialecki himself became the narrative spine of the roadshow.
The lesson: when the product is unsexy, the founder narrative carries the staging, and that requires founder coaching, not slide design. Bialecki had to learn how to be the equity story in real time, in dozens of one-on-one institutional meetings, with no fireside chat to lean on. That’s a different production discipline than a celebrity CEO event. It’s intimate. It’s repetitive. It requires the founder to deliver the same emotional beats with freshness on the fortieth pitch. Klaviyo priced at $30, raised $576 million at a $9.2 billion valuation, and opened with a 23% first-day pop. The market recognized a credible operator, not just a magnetic personality.
In February 2026, Rippling aired a Super Bowl ad. Not because it was selling HR software to households. Because the company is making itself a household name before it asks public-market investors to recognize it.
Rippling hasn’t filed an S-1. CEO Parker Conrad has said publicly that an IPO isn’t imminent. Yet the production preparation is unmistakable. Rippling raised a $450 million Series G at a $16.8 billion valuation in May 2025. Annual revenue reached $570 million in February 2026, growing over 30% YoY with net revenue retention approaching 200%. The Super Bowl ad slot ran around $7 million.
Layer in the Deel lawsuit subplot. Rippling sued its largest competitor in March 2025 alleging corporate espionage; Deel countersued in April. The result is a roadshow narrative-in-waiting with built-in tension, competitive stakes, and a story arc the market is already following.
The lesson, even before the S-1 lands: the best roadshows aren’t built starting at the prospectus filing. They’re built years earlier, through brand investment, narrative seeding, and earned-media architecture. By the time Conrad walks into his first investor meeting, the story will already be partly told. Watch how the roadshow handles the Deel feud, whether it’s foregrounded as competitive moat or treated as backdrop noise. That choice will tell you everything about how Rippling has rehearsed its public-company identity.
Map the four cases onto OpenAI and Anthropic, and the patterns start to line up.
To us, the interesting observation isn’t that these labs will eventually IPO. It’s that the production preparation for that moment appears to be happening on the stages we’re already watching. DevDay and Code with Claude read as public dress rehearsals. The audience swap-out (developers in 2025, institutional investors in 202?) is mostly a matter of who’s in the room. The staging architecture is portable.
OpenAI’s ChatGPT has reached 800 million weekly active users (as Altman announced at DevDay 2025). The pre-IPO question is whether any of those users become shareholders. SpaceX just precedent-set the retail allocation question with its 30% retail target. DevDay’s Apps SDK announcement, which lets users chat directly with apps from Booking.com, Zillow, Target, Figma, and others, is the technical precondition for a community-visibility play at IPO. OpenAI is staging the infrastructure now.
Both labs face the data-center-economics problem. Both are spending tens of billions on compute. Both have the same risk: investors hearing “infrastructure” and pricing it as commodity capex. Sora Cinema was a small move in the opposite direction, making AI compute feel like a consumer experience rather than a balance-sheet item. Anthropic’s API-and-MCP focus at Code with Claude is the more enterprise-coded version of the same instinct. Whoever IPOs first will need a tangibility moment in their roadshow that CoreWeave didn’t quite execute.
Sam Altman is on stage all the time, but a roadshow is a different production challenge: dozens of intimate institutional meetings, repeated questions, no audience. Dario Amodei rarely takes that kind of stage. Both will need to develop the discipline of delivering the same narrative architecture, freshly, to room after room, for two weeks straight. That work happens long before the S-1.
This is where the AI labs appear most clearly to be rehearsing. The Sam Altman / Jony Ive partnership reads as brand seeding. Anthropic’s policy presence in Washington reads as brand seeding. Code with Claude’s multi-city expansion reads as brand seeding. None of it is sales activity. To our eye, much of it functions as roadshow preparation by other names.
The combined effect: the AI labs appear to be running all four production patterns at once, with bigger budgets and longer runway than any of their predecessors. By the time the S-1s eventually land, their production foundations will be deeper and more multi-dimensional than anything Reddit, CoreWeave, or Klaviyo had the runway to build.
If you’re a head of IR, head of corporate comms, or growth-stage CFO watching this rehearsal play out, here’s the framework outlined by these events:
Who’s in the room, and visibly in the room, is itself the equity story. The presence of an a16z partner on a developer stage doesn’t communicate the same thing as a developer relations engineer.
Every roadshow needs a single staged conversation that conveys narrative information no slide can. For OpenAI it’s Altman + Ive. For Reddit it was the moderator share program. For Klaviyo it was the founder’s bootstrapped credibility. Find yours.
If your story includes infrastructure, capacity, scale, or technical depth, design at least one production moment that translates the abstraction into an embodied experience. CoreWeave’s roadshow under-indexed on this. Sora Cinema was a low-stakes attempt to learn the lesson early.
Every speaker on your stage represents a category of validation: customer voices, partner voices, investor voices, civic voices. Choose the mix to match the equity story you’re telling, not just the agenda you’re filling.
Roadshows happen in conference rooms with twenty institutional investors at a time. The founder who’s brilliant on stage isn’t automatically brilliant in that room. Different muscle. Develop it years in advance.
Rippling’s Super Bowl ad. OpenAI’s Apps SDK partner announcements. Anthropic’s policy and developer presence. The roadshow doesn’t begin at the filing date. It begins the moment institutional investors first start hearing your name in unscripted contexts.
DevDay’s keynote streamed live. The Altman + Ive fireside didn’t, but went up on YouTube the same day. That sequencing is deliberate. Live serves urgency; on-demand serves reach. Design both.
When the AI labs eventually file, every staging choice they’re making now becomes part of the institutional case file investors use to evaluate them. The rehearsal is the record.
The bright line between “developer event” and “investor day” appears to be dissolving.
This isn’t unique to the AI sector. To us, it reads as an ongoing structural shift. Companies with consumer scale, enterprise depth, and platform ambition can’t run their pre-IPO communications through a single channel anymore. The audience for an IPO is now a constellation: institutional investors, retail allocations, developer ecosystems, enterprise buyers, regulators, partners, employee shareholders, and a financial press taking cues from social platforms. Every staged moment a company runs in the years before filing reaches some subset of all of them.
That makes production more strategic, not less. The companies that stage their developer events with the audience composition of an Investor Day in mind, that invest in their brand before the prospectus, and that coach their founders for both the keynote and the conference room, should have an enormous advantage over companies that wait until the S-1 to start thinking about narrative.
The rehearsal is the production. By the time the S-1 hits, the show is already running.
That’s the work worth investing in. It’s also the work Cardboard Spaceship builds for clients navigating the moments that matter.
The next generation of IPO roadshows won’t start at the S-1 filing. They’ll start years earlier, in the staged moments that quietly build institutional recognition. Whether you’re preparing for an Investor Day or the long lead-up to a future filing, the production decisions you’re making now will define how the market responds when the moment arrives. Let’s start a conversation →
Nine days from now, Greg Abel will walk onto a stage Warren Buffett owned for six decades.
It’s the Berkshire Hathaway 2026 annual meeting, and for 60 years, it’s unapologetically followed the same format. 40,000 people will file into the CHI Health Center in Omaha, Nebraska. They’ll take their seats in the same arena. They’ll hear CNBC’s Becky Quick introduce the Q&A session. They’ll see the Berkshire logo on the stage.
But this time, Buffett will be ten feet away. In the front row. Silent by his own public declaration.
The Berkshire Hathaway 2026 annual meeting would already be one of the most watched corporate events of the year on that fact alone. But something more interesting is already in motion. Abel has quietly restructured the format, and he hasn’t said a word yet.
For 60 years, this meeting had one production element worth analyzing: Warren Buffett in a chair for five hours. Everything else was deliberate absence. No slides. No teleprompter. No walk-on music. Just a microphone, a can of Cherry Coke, and the accumulated wisdom of the most successful investing career in history.
The anti-production was the production. The simplicity communicated respect for investor intelligence. The marathon length signaled nothing to hide. The solo performance said one person is accountable for everything. Those choices built the “Woodstock for Capitalists” – the only corporate event in the world that reliably draws 40,000 people to Omaha.
And Abel has redesigned all of it.
Two Q&A panels instead of one marathon. New voices on stage – Ajit Jain (insurance), Katie Farmer (BNSF), Adam Johnson (consumer products). The traditional open Q&A intact in spirit, restructured in execution.
That change isn’t cosmetic. It’s the succession narrative made physical. And it happened before Abel ever stepped up to the microphone.
The question is whether it works.
No CEO has ever inherited a corporate event this consequential. 40,000 shareholders. Global webcast in English and Mandarin. Buffett in the front row. A $380 billion cash position on the balance sheet.
When Buffett announced his retirement at last year’s meeting, the standing ovation lasted minutes. The question had already shifted from “when will Buffett step down?” to “what does Berkshire look like without him?”
May 2 is when the market gets its first real answer.
We’ve been in enough high-stakes investor rooms to know what’s at play. Here are five production and narrative challenges Abel’s team will need to navigate on May 2, and our predictions for how each one plays out.
This is the most loaded staging decision of the entire event.
Buffett sitting among the directors, visible but silent, is an extraordinarily powerful visual. It’s a living endorsement of the transition. It says: I trust this. I’m here. But it’s his turn now.
It also creates a gravitational pull the production team will need to manage carefully:
Our prediction: the webcast production will show Buffett briefly at the opening, then keep the focus firmly on the stage for the rest of the event.
The discipline is in resisting the reaction shot. Every time the camera cuts to Buffett instead of Abel, the narrative slides backward. The production team needs to treat the front row as context, not content.
There’s a subtler challenge too. Buffett’s physical presence in the room will hold an emotional weight no amount of production design can fully manage. Some shareholders will spend the entire meeting watching him, not the stage. That’s human nature, and the event can’t prevent it. But it can refuse to feed it.
The traditional Berkshire Hathaway annual meeting format was beautifully simple.
One person. One chair. Five hours.
Questions from the audience, answered in real time, with no filter and no limit. That simplicity wasn’t an accident. It was Berkshire’s brand made physical. Transparency. Directness. Trust.
Abel has restructured the format into two distinct Q&A panels:
The format change is, essentially, a narrative statement about where Berkshire’s value lives now. Buffett’s solo format said the value was in one chair – in one person’s judgment and ability to allocate capital. Abel’s panel format says something different. The value is in the operating leaders:
The quality of the people running the pieces, not just the person orchestrating the whole.
Our prediction: institutional investors will read this format change correctly; as a signal that Berkshire under Abel will be more operationally transparent, more team-driven, and less dependent on the mystique of a single decision-maker.
Some longtime retail shareholders may experience it as a loss. Both reactions are valid with a legacy of this magnitude.
There’s no polite way to say this: Warren Buffett was one of the great entertainers in corporate history.
His annual meeting performances were legendary not just for their financial insight, but for their warmth, their humor, their stories. He’d spend 10 minutes on a single question, reference a deal from 1967, make a joke about See’s Candies that somehow also explained capital allocation theory.
Abel won’t do that. And he shouldn’t try.
What Abel brings is different:
The production challenge is real. Forty thousand people in an arena that are accustomed to being entertained and educated simultaneously. Abel will educate, but education alone has to hold a room of that size for hours.
Our prediction: Abel’s answers will be tighter, more operational, less philosophical.
The Q&A sessions will feel shorter even if they run the same length, because the pacing will be steadier and the digressions fewer. Some attendees will call it “refreshing.” Others will call it “less fun.” Both are probably right.
The production team can help:
Berkshire is sitting on roughly $380 billion in cash and short-term investments. It’s the elephant in the room within every conversation about Abel’s leadership. And it will be the first hard question he faces on May 2.
Some version of “what are you going to do with the money?” will come early, probably from Becky Quick, who knows it’s the question on every shareholder’s mind. How Abel handles it will set the tone for his entire tenure.
The trap: over-promise.
Buffett spent decades preaching patience on capital allocation, and the market rewarded him for it. Abel needs to earn that same credibility, which means his first instinct on May 2 should be patience, not action.
Our prediction: Abel will acknowledge the cash position directly, reaffirm the discipline of waiting for the right opportunity at the right price, and resist the temptation to hint at anything specific.
The smartest answer is some version of “we have the capital to be decisive when the moment is right, and the discipline to wait until it is.” It’s not the answer that generates headlines. It’s the answer that builds trust.
This is the question nobody on stage will ask, but everyone in the room will be thinking about.
Forty thousand people came to Omaha because of Warren Buffett. The weekend around the meeting was built around his presence:
The “Woodstock for Capitalists” brand was inseparable from its headliner.
So what happens in 2027?
The 2026 meeting is the transition year. Attendance will likely hold. People want to see the first post-Buffett meeting, and many shareholders already have the trip booked.
But the real indicator isn’t 2026 attendance. It’s 2027.
If the numbers hold, the event has successfully become bigger than its founder. If they drop significantly, the market learns that the “Woodstock for Capitalists” was always more about the capitalist than the Woodstock.
Our prediction: Abel’s team knows this.
The format changes (multiple panelists, operational depth, visible bench strength) are designed for 2027 just as much as 2026. They’re building a format that doesn’t rely on one person’s magnetism to draw a crowd. Whether it works is the test that runs beyond May 2.
There’s a reason Buffett’s performances became legendary beyond the financial world.
He was one of the great entertainers in corporate history. His annual meeting set piece: spend 10 minutes on a single question, reference a deal from 1967, make a joke about See’s Candies that somehow also explained capital allocation theory. The audience stayed because he was teaching them something AND because he was fun to listen to.
Abel can’t replicate that. He also shouldn’t.
Buffett traded in folksy wisdom from the Oracle of Omaha. Abel trades in operational rigor – a CEO who has actually run the businesses he’s discussing. Different brand of credibility, not a diminished one.
The production team can amplify what Abel brings naturally:
Abel’s job on May 2 isn’t to answer any single question brilliantly. It’s to establish that the meeting still feels like Berkshire in his voice – honest, direct, unhurried – without requiring Buffett’s ghost in the room to make it work.
Katie Farmer runs BNSF Railway, one of the largest freight operators in North America and one of Berkshire’s most important businesses.
Most retail shareholders wouldn’t recognize her name. She’s never been a featured speaker at the annual meeting. In the Buffett era, she didn’t need to be.
Now she’s on stage. So is Adam Johnson from NetJets and consumer products. So is Ajit Jain from insurance.
The meeting does something it has never done before: puts operating leaders in front of the audience, answering real-time questions about what they actually run:
Shareholders spend every year reading about BNSF in the annual letter. Having Farmer field questions live is a different kind of signal – the business has a visible leader who can speak to operations at the level of detail institutional investors want.
Add Johnson and Jain to that picture, and the company presents itself as what it actually is: a collection of well-run businesses with deep bench strength, not a portfolio that only makes sense in one person’s head.
For any company navigating a leadership transition, there’s a production principle worth studying here. Don’t announce that you have great leaders. Put them on stage and let the audience see for themselves.
Berkshire’s 2026 meeting is the most visible example of a challenge every company eventually faces.
The founder steps back. The iconic CEO retires. The person who is the brand moves off the stage. And the event that was built around them has to keep working.
Here’s what we’d tell any company in that position.
Abel isn’t trying to be Buffett, and the format isn’t trying to recreate the solo marathon. The worst thing a successor can do is imitate the predecessor’s style.
The audience will spot it instantly, and it reads as insecurity, not continuity. Find the new leader’s authentic strengths and design the format around those.
Bringing operating leaders on stage communicates depth, reduces key-person risk perception, and gives the audience multiple points of connection with the company. If your event previously depended on one magnetic speaker, widening the stage is how you build a format that outlasts any individual.
The audience needs a moment to honor what came before. Then they need the event to move forward with confidence.
Lingering on the transition (spending too long on tributes, too many backward-looking references) signals that the company is more attached to its past than its future. A brief, genuine acknowledgment followed by decisive forward motion is the right balance.
Abel kept the Q&A. He didn’t add slides. He didn’t add video. He kept the thing that makes Berkshire’s meeting fundamentally different from every other shareholder meeting on Earth: the willingness to answer whatever shareholders want to ask, in real time, with no filter.
That format choice is the message. It says: we’re still Berkshire.
The 2026 meeting will get great attendance because of the transition itself. Everyone wants to see the first post-Buffett meeting.
The real test is Berkshire Hathaway 2027. Whether the format, the new voices, and the new identity can sustain the event on their own merit. Plan for the second year, not just the first.
On May 2, the most important signals won’t come from the stage.
They’ll come from the room:
And the question that hovers over all of it: does the room still feel like Berkshire?
That feeling (the sense of belonging to something, the connection between a company and its owners, the trust that what you’re hearing is the unvarnished truth) is what made the “Woodstock for Capitalists” something more than a shareholder meeting.
That quality was Buffett’s greatest production achievement. Not the jokes or the stories or the Cherry Coke.
The trust. The sense that this company respects you enough to sit in a chair and answer your questions for five hours.
Abel can’t inherit that trust. He has to earn it.
And May 2 is when the earning starts.
The format is set. The panels are announced. Becky Quick has the questions. Buffett has his seat in the front row. Forty thousand people have their tickets.
The rest is live.
We’ll be watching. For the production choices, the narrative architecture, and the moments that tell us whether the most iconic corporate event in history just found its second act.
Every analyst walks in with questions. The most effective events answer them before they’re asked, through narrative architecture, experiential production, and financial precision that earns the room’s conviction. Let’s start a conversation →
No slides. No teleprompter. No walk-on music. No rehearsed opening. No cocktail reception. No video package. No panel of executives. No celebrity moderator.
Just one man, Warren Buffett, in a chair for five hours, answering whatever he’s asked.
By every modern standard of event production, the Berkshire Hathaway annual shareholder meeting should be a snoozer. It violates every rule of contemporary audience engagement, and it’s the exact opposite of what IR consultants typically recommend.
Yet, it draws 40,000 people to Omaha every May. It commands global media coverage. It moves markets. And it’s routinely cited by institutional investors as the single most valuable investor communication event in the world.
So what’s going on?
The short answer: the Berkshire meeting isn’t boring. It’s disciplined. And it works for the exact reasons most shareholder meetings don’t.
Here’s what 60 years of the “Woodstock for Capitalists” teaches about effective investor communication – and what any IR team can apply to their next event.
Most corporate events are built around a fundamental assumption: investors need to be entertained, impressed, or both.
So the production gets heavier every year. More video. More stagecraft. More choreographed executive walkons. More polished messaging. The implicit message: we don’t trust you to stay interested unless we keep the energy up.
Berkshire does something different, and it takes more production skill, not less, to pull off. The format assumes the audience is smart enough to stay engaged without decorative flourishes. Every production decision – staging, lighting, camera work, broadcast design – is built to disappear so the content can carry the room.
That’s the hardest thing to do in event production. Making the work invisible requires a team that understands exactly what to amplify and when to pull back .
The takeaway for your next event: production isn’t about adding spectacle. It’s about designing every element – visible or invisible – to serve the core moment. The most confident companies communicate through production that feels effortless and intentional, which is almost always the product of the most thoughtful production design.
Every production decision at a Berkshire meeting communicates something:
None of those are obvious production choices. Each one was deliberate. Each one required significant production planning and infrastructure to execute consistently year after year.
Together, they communicate Berkshire’s values more effectively than any mission statement could.
The takeaway for your next event: every production choice – length, format, staging, Q&A structure, broadcast design – is saying something about your company. The best production partners don’t just execute the format. They help you understand what each decision communicates and design every element to reinforce the narrative you’re trying to build.
Most shareholder meetings are a mosaic of competing elements: opening remarks, CEO presentation, CFO deep-dive, divisional updates, video segments, panel discussions, Q&A, networking reception. Each element is designed to serve a different audience need. The result is often a meeting that does many things adequately and nothing exceptionally.
Berkshire picked one element – the open-question Q&A – and made it the entire event.
That single format decision is what built the brand. It’s also what makes the event irreplaceable. There’s nowhere else an institutional investor can ask a man like Warren Buffett – the CEO of a $1.1 trillion company – any question they want and get a real answer in real time for five hours straight.
The takeaway for your next event: identify the one element of your investor communication that genuinely differentiates you. Then ask whether the rest of your format is supporting that element or diluting it. If you’re doing ten things to satisfy different stakeholders, none of them will be memorable.
There’s a reason Berkshire’s meeting generates so much media coverage. Reporters and analysts know Buffett will answer hard questions directly – about the economy, about politics, about specific holdings, about mistakes. The format makes it impossible to dodge.
Compare that to a typical investor meeting where executives deliver scripted remarks, take three pre-screened questions, and exit stage right. The contrast isn’t subtle. Shareholders notice. So do journalists. So do short sellers.
Transparency at Berkshire isn’t a corporate value statement. It’s a production structure. The format doesn’t allow evasion. And because the format doesn’t allow evasion, the company has spent 60 years demonstrating that it has nothing to evade.
The takeaway for your next event: if your IR format includes a lot of guardrails; pre-screened questions, heavily rehearsed executives, short Q&A windows, vague forward-looking statements – those guardrails are communicating something. Ask whether what they’re communicating is actually what you want.
The “Woodstock for Capitalists” works because its core identity is unmistakable.
Same weekend. Same city. Same open Q&A format.
But here’s what’s easy to miss. Berkshire has evolved the event substantially over 60 years. The webcast expanded from audio-only to video to global streaming in English and Mandarin. The question-submission system evolved. The microphone lottery was introduced. The broadcast production grew from bare-bones to one of the most sophisticated corporate livestreams in the world.
What Buffett and his team understood is the difference between identity and execution. The identity stayed fixed. The execution kept getting better.
That’s the formula most IR teams get wrong in one direction or the other. Some reinvent everything every year and lose the continuity that builds audience loyalty. Others freeze their format and miss opportunities to elevate the experience as technology, audience expectations, and stakes evolve.
The best IR events have clear identity and constant elevation. What stays the same is the core: the narrative, the intentionality, the relationship. What gets better every year is the craft that delivers it.
The takeaway for your next event: separate what should never change from what should always improve. Your core identity, your voice, your relationship with investors – protect those fiercely. Your production value, your storytelling sophistication, your technical execution – those should get better every single year. Great production partners help you identify which is which and elevate the execution without disturbing the foundation.
Every IR team is making the same tradeoff, whether they realize it or not.
Most shareholder meetings are designed around this year’s messaging. The best ones invest in a relationship with investors that lasts decades.
There’s a reason Berkshire’s annual meeting has outlasted recessions, leadership questions, and Warren Buffett’s own transition off the stage. The event was built to carry weight, not to win a single news cycle. The format, the venue, the open Q&A, the trust between the company and its shareholders – none of it happened by accident. It was built, refined, and protected, year after year, by a team that understood what the event needed to carry. The result is the most respected, most attended, most covered corporate event in the world.
That’s not boring. That’s the highest form of the craft.
Great IR communication isn’t built one event at a time. It’s built as a long-term asset – a production foundation strong enough to hold the company’s relationship with investors through every cycle, every leadership change, every market condition.
That’s the work worth investing in. And that’s the quality Cardboard Spaceship delivers.
Rethinking your next investor event?
The best IR communication isn’t about stripping production down or piling it on. It’s about making sure that every production decision, visible or invisible, earns its place. That’s the discipline we bring to every high-stakes investor event we build.
