Axios Pro Rata: Remembering Om | Axios Pro Rata Summary

1. Key Themes

Theme 1: Gaming Data as a Novel AI Training Moat

The most capital-intensive deal of the issue centers on an unorthodox data strategy — using gaming content to train world models and action models, rather than conventional web-scraped or synthetic data.

"GI's bet is that gaming — both gameplay video and the player inputs that produced it — can help build both world models and large action models faster and cheaper than by other training techniques."

At $320M Series A and a $2.3B valuation, the market is clearly pricing this thesis as credible.

Theme 2: AI Expanding Into the Physical Economy

Goldman Sachs, in a report shared exclusively with Axios, frames the current moment as a structural shift — AI moving from software into real-world, physical applications.

"The next phase of AI adoption and investment is expanding into the physical economy, according to a Goldman Sachs report shared exclusively with Axios."

This signals that investors should be tracking AI companies targeting industrial, logistics, and physical-world use cases, not just software productivity.

Theme 3: AI Infrastructure & Reliability as an Emerging Subsector

Multiple investments this issue point to a maturing AI stack where reliability, observability, and stress-testing are becoming standalone investment categories — not just features bolted onto models.

Scaled Cognition raised $100M Series A focused specifically on AI reliability.
Patronus raised $50M Series B for AI agent stress-testing.
Sazabi raised $8M seed for engineering observability.

No single quote summarizes all three, but the Patronus entry notes backing from Datadog and Samsung alongside VCs — a strong signal that enterprise operators are co-investing in reliability infrastructure.

Theme 4: Frontier Lab Formation Window Is Closing (But Not Closed)

The General Intuition raise carries a broader meta-signal about the perceived difficulty of starting new foundation model labs.

"It shouldn't have been possible to start a frontier lab in 2025. The doors were shut, they said. But thanks to Vinod [Khosla], and his relentless ability to believe in founding teams, it was." — Pim de Witte, CEO of General Intuition

This is a rare acknowledgment that the window for new entrants at the frontier level is nearly shut — making General Intuition's raise a notable exception, not a trend.

Theme 5: OpenAI IPO Pushback Signals Private Market Preference

OpenAI is reportedly deferring its public offering, consistent with a broader pattern of large AI companies staying private longer.

"OpenAI is 'leaning toward' going public in 2027 instead of this year, per the NY Times."

For investors, this extends the period of value capture in private markets and may pressure secondary markets for OpenAI equity.

2. Contrarian Perspectives

Contrarian Take 1: Geopolitical Oil Shocks Are Losing Their Economic Punch

The conventional wisdom is that Middle East conflict = oil disruption = economic pain. The data here challenges that:

"The Iran war hasn't wreaked nearly as much havoc on the U.S. economy as prior oil supply disruptions, for two big reasons: (1) The rise of U.S. oil production (thus reduced import reliance). (2) We use less oil to produce electricity and have more efficient vehicles, plus the U.S. economy has shifted to services from manufacturing."

The implication for investors: energy markets and geopolitical risk premiums may be structurally less correlated to U.S. economic outcomes than historical models would suggest. Traditional "flight to safety" trades around oil shocks may be increasingly mispriced.

Contrarian Take 2: Gaming Is Serious AI Training Infrastructure, Not Entertainment

The mainstream view treats gaming as a consumer entertainment vertical. General Intuition's $2.3B valuation frames it as critical AI training infrastructure — an entirely different asset class.

"GI's bet is that gaming — both gameplay video and the player inputs that produced it — can help build both world models and large action models faster and cheaper than by other training techniques."

If validated, this repositions gaming data archives and game engine companies as strategic AI infrastructure assets — a lens few investors are currently applying.

Contrarian Take 3: New Frontier AI Labs CAN Still Be Built, With the Right Backer

The consensus is that the frontier AI lab era is over for new entrants — compute costs and incumbents have locked the door. General Intuition's founder explicitly pushes back:

"It shouldn't have been possible to start a frontier lab in 2025. The doors were shut, they said. But thanks to Vinod [Khosla], and his relentless ability to believe in founding teams, it was."

The contrarian read: the barrier isn't insurmountable — it's a conviction and relationship barrier more than a pure capital barrier. The right lead investor can still unlock the formation of new frontier labs.

3. Companies Identified

General Intuition

Description: AI lab using gaming content (gameplay video + player inputs) for AI model training
Why mentioned: The "BFD" (Big F***ing Deal) of the issue — $320M Series A at $2.3B valuation, already working on Series B; backed by Khosla, General Catalyst, Hedosophia, Bezos Expeditions, Innovation Endeavors, Nico Rosberg
Quote: "GI's bet is that gaming — both gameplay video and the player inputs that produced it — can help build both world models and large action models faster and cheaper than by other training techniques."

Kalshi

Description: Prediction markets platform
Why mentioned: In talks to raise at a $40B valuation, per the FT — a massive implied scale for an event contracts platform
Quote: "Kalshi is in talks to raise funds at a $40b valuation, per the FT."

Scaled Cognition

Description: Mountain View-based AI lab focused on reliability
Why mentioned: $100M Series A led by Khosla Ventures — notable as a pure-play bet on AI reliability as a category
Quote: "Scaled Cognition, a Mountain View, Calif.-based AI lab focused on reliability, raised $100m in Series A funding."

Patronus

Description: AI agent stress-testing startup
Why mentioned: $50M Series B with strategic backing from Datadog and Samsung alongside top VCs — signals enterprise validation of AI testing infrastructure
Quote: "Patronus, an AI agent stress-testing startup, raised $50m in Series B funding. Greenfield Partners led, joined by Notable Capital, Lightspeed, Datadog, and Samsung."

Gigaom

Description: Early tech media company founded by Om Malik
Why mentioned: Cited as a pioneering tech media venture and context for Om Malik's legacy
Quote: "He even pinch-hit on my original newsletter and was an enthusiastic career counselor."

OpenAI

Description: Leading AI research company
Why mentioned: Reportedly pushing IPO to 2027, extending its private market window
Quote: "OpenAI is 'leaning toward' going public in 2027 instead of this year, per the NY Times."

Reformation

Description: Womenswear retailer owned by Permira
Why mentioned: Filed for IPO, estimated ~$200M raise; $507M revenue and $13M net income in 2025 — a data point on consumer brand IPO viability
Quote: "It reports $13m in net income on $507m in revenue for 2025 and plans to list on the NYSE (REF)."

iMerit

Description: San Jose-based AI model training company
Why mentioned: Acquired by ExlService for up to $310M — a significant exit for a company that had raised only ~$25M, representing a strong return multiple for backers including YC
Quote: "ExlService (Nasdaq: EXLS) agreed to buy iMerit, a San Jose, Calif.-based AI model training company, for up to $310m. iMerit had raised around $25m."

Clearwater Analytics

Description: Boise-based investment and accounting software maker
Why mentioned: $8.4B take-private by Permira and Warburg Pincus completed — one of the largest PE software take-privates referenced
Quote: "Permira and Warburg Pincus completed their $8.4b take-private buyout of Clearwater Analytics."

Adobe

Description: Public software company (Nasdaq: ADBE)
Why mentioned: Acquiring Topaz Labs (video and image enhancement AI models) — signals continued Adobe AI content tool consolidation
Quote: "Adobe agreed to acquire Topaz Labs, an Addison, Texas-based provider of video and image enhancement models."

Elroy Air

Description: Byron, CA-based cargo drone startup
Why mentioned: Going public via reverse merger at $800M pre-money valuation — notable defense/logistics drone SPAC-style exit
Quote: "Elroy Air, a Byron, Calif.-based cargo drone startup, agreed to go public via a reverse merger with Columbus Circle Capital Corp. II at a $800m pre-money valuation."

Bumble

Description: Dating app (Nasdaq: BMBL), Blackstone-backed
Why mentioned: Hired Morgan Stanley to explore a sale — potential M&A target in consumer social
Quote: "Bumble (NAS: BMBL), a dating app backed by Blackstone, hired Morgan Stanley to explore a sale, per Reuters."

ON Semiconductor / Synaptics

Description: ON Semi (Nasdaq: ON) acquiring Synaptics (Nasdaq: SYNA)
Why mentioned: $7B all-stock semiconductor consolidation deal
Quote: "ON Semiconductor agreed to acquire Synaptics for $7b in stock."

Chess.com

Description: Online chess platform
Why mentioned: CVC Capital Partners agreed to invest, with General Atlantic as existing backer — PE interest in niche digital sports/gaming platforms
Quote: "CVC Capital Partners agreed to invest in online chess platform Chess.com, whose existing backers include General Atlantic."

4. People Identified

Om Malik

Description: Pioneer internet entrepreneur, journalist, tech media founder (Gigaom), and venture capitalist
Why mentioned: Subject of the newsletter's lead tribute following his death
Quote: "Om was an early internet entrepreneur who soon became one of the nascent industry's leading journalists. He blogged. He reported. He founded a tech media company — serving as both inspiration and competition for many of us — before later becoming a venture capitalist."

Pim de Witte

Description: CEO of General Intuition
Why mentioned: Led the founding of a frontier AI lab despite consensus that this was no longer possible; quoted directly on the improbability of the raise
Quote: "It shouldn't have been possible to start a frontier lab in 2025. The doors were shut, they said. But thanks to Vinod [Khosla], and his relentless ability to believe in founding teams, it was."

Vinod Khosla

Description: Founder, Khosla Ventures
Why mentioned: Named explicitly as the believer who made General Intuition's frontier lab formation possible; also led Scaled Cognition's Series A — notably active in AI infrastructure bets
Quote: "Thanks to Vinod [Khosla], and his relentless ability to believe in founding teams, it was [possible]."

Katie Fehrenbacher

Description: Axios Pro reporter; first employee of Gigaom
Why mentioned: Provided a firsthand remembrance of Om Malik, offering insight into his methodology and mentorship style
Quote: "Om had a unique way of looking at technology, using his knowledge of the underlying infrastructure to predict how people would use it."

Kara Swisher

Description: Veteran tech journalist and media entrepreneur
Why mentioned: Provided remembrance of Om Malik, underscoring his industry-wide influence
Quote: "From the start of the Internet age, Om was always my inspiration for so many things. While we took different paths, he forged the trail we all followed."

Marianne Lake

Description: Outgoing CEO of JPMorgan Consumer & Community Banking
Why mentioned: Her retirement removes her from the Jamie Dimon succession race — significant for JPMorgan succession watchers
Quote: "Marianne Lake is retiring as CEO of JPMorgan's consumer and community banking business, taking her out of the Jamie Dimon succession race."

Novak Djokovic

Description: Professional tennis player
Why mentioned: Joined General Atlantic as a global strategic advisor — notable athlete-to-PE crossover
Quote: "Novak Djokovic, the star tennis player, joined General Atlantic as a global strategic advisor."

Adam Candeub

Description: General Counsel at the FCC
Why mentioned: Expected Trump nomination as next DOJ antitrust chief — relevant to tech M&A regulatory outlook
Quote: "President Trump is expected to nominate Adam Candeub, general counsel at the FCC, as the next DOJ antitrust chief."

5. Operating Insights

Insight 1: Use Infrastructure Knowledge to Predict Consumer Behavior

Om Malik's analytical approach, highlighted in remembrance, is a durable operating framework for anyone building or investing in tech:

"Om had a unique way of looking at technology, using his knowledge of the underlying infrastructure to predict how people would use it."

For operators and investors, the implication is clear: deep technical understanding of infrastructure layers is a predictive advantage — not just for engineering decisions, but for product strategy and market timing.

Insight 2: Unconventional Data Sources Can Unlock Cost and Speed Advantages in AI Training

General Intuition's thesis is operationally instructive for any AI company wrestling with training data costs:

"GI's bet is that gaming — both gameplay video and the player inputs that produced it — can help build both world models and large action models faster and cheaper than by other training techniques."

Operators building AI products should actively audit whether underutilized data domains (simulations, games, procedural environments) could substitute for expensive real-world data collection.

Insight 3: Community Building Is a Durable Competitive Advantage

A quieter lesson from the Om Malik tribute: his legacy was amplified not just by his writing, but by his deliberate investment in community.

"He was always there first, whether it was how to write about tech as a human endeavor, or how to throw amazing and substantive gatherings to bring the community together, to being righteous about the importance of holding leaders accountable."

For founders and operators, intentional community cultivation — events, mentorship, accountability — builds durable reputational moats that outlast any single product or company.

6. Overlooked Insights

Overlooked Insight 1: iMerit's Capital Efficiency Is a Standout Exit Signal

iMerit's acquisition by ExlService for up to $310M on only ~$25M raised is buried in the Liquidity Events section but represents a remarkably capital-efficient outcome:

"ExlService (Nasdaq: EXLS) agreed to buy iMerit, a San Jose, Calif.-based AI model training company, for up to $310m. iMerit had raised around $25m from British International Investment, the Michael & Susan Dell Foundation, YC, and Omidyar Network."

That's potentially a 12x+ return on invested capital — and a signal that AI data labeling and model training services companies, often overlooked in favor of model builders, can generate outsized exits with modest capital needs.

Overlooked Insight 2: Prediction Markets Are Approaching Major Institutional Scale

Kalshi's reported $40B fundraising valuation is noted with a single bullet and no further commentary — but the number deserves attention:

"Kalshi is in talks to raise funds at a $40b valuation, per the FT."

If accurate, this would make Kalshi one of the most valuable fintech companies in the U.S. — and signals that prediction markets, long dismissed as niche or legally precarious, may be crossing into mainstream institutional asset class territory.