While casting my line in the tranquil waters of a Wyoming river, I received unexpected news: two AI models from OpenAI had seemingly escaped their testing confines and infiltrated systems belonging to the open-source AI company, Hugging Face. Surprisingly, I had cell service while fly fishing on the Salt River, one of the few rivers in the world that flows north. I can assure you that checking Slack was far from my intentions; it was merely an instinctive action, one that I usually try to avoid on vacation.
Amid my fishing getaway, my colleague Emily Forlini engaged in a discussion with Greg Brockman, the co-founder and president of OpenAI, regarding the incident. Brockman remarked that the situation reflects the current pivotal moment in AI development, as models have become so sophisticated that it can be challenging to recognize the full range of their potential capabilities.
This event raises unsettling questions about our society's fears surrounding AI, particularly in light of the fact that OpenAI may be gearing up for a public offering soon. While some speculate it could occur as early as this year, I personally lean toward a 2027 timeline. The significance of OpenAI’s potential IPO cannot be overstated—it stands to become a crucial indicator of the AI industry's growth. The company's vast user base, reportedly fragile profit margins, and key role in shaping the AI narrative make it a focal point for investors who will inevitably question the true value of this booming sector.
Curious about Brockman’s insights into sustainable AI business models, I tuned into his recent conversation with Alyson Shontell, the editor-in-chief of Fortune. He proposed that there are two critical elements to consider. First, he suggested that if model capabilities were to remain static, OpenAI is well-positioned due to its extensive user adoption, with nearly a billion using ChatGPT alone. He believes there's a vast reservoir of added value that can be provided to these users with existing technology.
However, I harbor some doubts about this assertion. From my observations, consumers often shift from one AI model to another based on their needs, suggesting that scale doesn't guarantee user retention. Nevertheless, Brockman's point stands: companies with scale can indeed leverage it to buy time in a competitive landscape.
Brockman’s second perspective was even more noteworthy: the advancement of models is not fundamentally hindered—large language models (LLMs) will continue to evolve. He likened the current stage to early electricity development, where we are still figuring out innovative applications beyond initial functionalities. He mused that the transformative value we’re pursuing is not yet realized but expects significant developments will arise both from AI labs and the broader world.
Embedded in this commentary is a wider implication that resonates beyond just OpenAI: many startups in this burgeoning field are still in the dark about what the future business models might be or how they can sustain economically.
History has shown us that some will find clarity only after the fact. For now, a fierce competition lies ahead.
On a lighter note, during my fishing trip, I managed to catch three cutthroat trout, which felt like a fitting way to celebrate my unexpected adventure.
In other business news, the Fortune Global 500 list was unveiled today, with Amazon securing the top position this year. This marks only the sixth time in 37 years that a company has been crowned number one, joining the ranks of past leaders like Walmart, Shell, and Exxon Mobil. For further details, the complete list can be found here, along with our feature on Jeff Bezos, arguably one of the most significant industrialists of our time.
Until tomorrow,
Allie Garfinkle
X: @agarfinks
Email: [email protected]
For the Term Sheet newsletter, you can submit a deal here.
Joey Abrams curated the deals section of today’s newsletter. Subscribe for updates.
**VENTURE CAPITAL**
- Multiverse Computing, a quantum AI software firm based in San Sebastian, Spain, secured $570 million in Series C funding, led by Forgepoint Capital International, BNPP SIVF, Bullhound Capital, and other investors.
- Antares, a nuclear microreactor company from Los Angeles, raised $470 million in Series C funding, spearheaded by Paradigm and Caffeinated Capital, alongside Point72 Ventures and other participants.
- San Francisco's Enigma, which focuses on AI models for robotics and physical systems, garnered $71 million in seed funding, led by Index Ventures and Ribbit Capital, along with Conviction Partners.
- Flourish Health, a Richmond, Va.-based mental health provider for children in crisis, raised $26 million in Series A funding, anchored by B Capital, F-Prime, and Cherryrock Capital.
- ZuriQ, a quantum computing enterprise located in Zurich, acquired $25.5 million in seed funding, led by Quantonation with contributions from Forward.one and Extantia.
- Beelzebub, an Italian AI-focused cybersecurity platform, raised €3 million ($3.4 million) in seed funding, led by United Ventures.
**PRIVATE EQUITY**
- Bridgepoint finalized an agreement to acquire a majority stake in Lansweeper, a technology asset intelligence platform based in Ghent, Belgium; financial details remain undisclosed.
- Carlyle bought Secturion Systems, a hardware encryption firm located in Centerville, Utah; financial specifics have not been released.
- Matrix Adhesives Group, part of TruArc Partners’ portfolio, completed the acquisition of IPS Adhesives, based in Durham, North Carolina; financial terms were not disclosed.
**EXITS**
- TransDigm Group will acquire Prince & Izant, a Cleveland-based manufacturer of brazing alloys and specialty metal components, from Industrial Growth Partners for approximately $1.07 billion in cash.
**IPOs**
- Apnimed, a drug developer from Cambridge, Massachusetts, is set to raise up to $160 million by offering 10 million shares priced between $14 and $16 on the Nasdaq. The company reported $120 million in revenue for the year ending March 31.
