- AI For All
- Posts
- The Beginning of the End
The Beginning of the End
OpenAI Raises $6.6 Billion
Hello readers,
Welcome to another edition of the AI For All newsletter! OpenAI has managed to raise $6.6 billion from investors at a $157 billion valuation. This may seem like a triumph, but there are conditions, and the road ahead is a tumultuous one. Let’s dive in!
The Beginning of the End
OpenAI has successfully raised $6.6 billion from investors at a $157 billion post-money valuation and also established a $4 billion revolving credit facility with JPMorgan Chase, Goldman Sachs, Morgan Stanley, and others, giving them access to $10 billion in liquidity. That’s right. A grossly unprofitable company with a half-working product that will only get more expensive to serve managed to raise billions from hype alone. However, this raise comes with some conditions.
Funding will convert into debt at a 9% interest rate if OpenAI does not become a for-profit within two years. The $4 billion revolving credit facility provides access to capital, but it’s still more debt. If OpenAI draws on it, they will need to repay it with interest. Since OpenAI loses billions of dollars a year, they actually need to raise even more money again and soon. OpenAI had some conditions of its own. Investors had to fork over at least $250 million and were asked not to invest in competitors.
Apple wisely chose not to invest in this round while SoftBank’s Masayoshi Son thought it prudent to invest $500 million. Son famously lost billions investing in WeWork. All this bailing out investment means the pressure is on for OpenAI to become a real company. OpenAI is attempting to undergo a metamorphosis. Like Google before them, they are going from engineering-led to management class ruled — a “real” company.
This is one explanation for all the departures. Add Tim Brooks to the list, co-lead of OpenAI’s Sora video generator, which is yet to be released. We should expect the price of ChatGPT to increase and more half-baked products. Apparently, OpenAI employees felt that o1 had been released prematurely. Ever determined to fail in OpenAI’s footsteps is Google, who are also planning a “reasoning” model.
The future of OpenAI hinges on an entirely hypothetical breakthrough, one that has nothing to do with their current product. It’s not even clear if OpenAI can raise the money it needs after they burn their $6.6 billion dollars, some of which may just be in the form of cloud credits from investor Microsoft. This may be the beginning of the end.
🔥 Rapid Fire
OpenAI announces multiple API updates and capabilities at DevDay 2024
OpenAI introduces Canvas, a new ChatGPT interface for writing and coding
Liquid AI unveils Liquid Foundation Models, an alternative to transformers
NVIDIA introduces NVLM, a family of open frontier-class multimodal LLMs
Accenture and NVIDIA form business group to scale enterprise AI adoption
Microsoft invests $4.7B into AI infrastructure and cloud capacity in Italy
White House announces competition for AI sustainable semiconductors
New studies from Uplevel, SSRN, and GitClear suggest AI not good for coding
Researchers find larger and more instructable LLMs become less reliable
New paper shows LLMs can be hacked using math despite safety measures
California governor vetoes AI safety bill that was potentially misrepresented
Google makes Gemini 1.5 Flash-8B production ready for developers via API
Google shares updates to Search with ads in AI Overviews and Google Lens
Cohere launches improved fine-tuning features with Command R 08-2024
Meta confirms it trains AI on images users ask Ray-Ban Meta AI to analyze
Cancer centers launch Cancer AI Alliance with AWS, NVIDIA, and Microsoft
Writer RAG tool: build production-ready RAG apps in minutes
RAG in just a few lines of code? We’ve launched a predefined RAG tool on our developer platform, making it easy to bring your data into a Knowledge Graph and interact with it with AI. With a single API call, writer LLMs will intelligently call the RAG tool to chat with your data.
Integrated into Writer’s full-stack platform, it eliminates the need for complex vendor RAG setups, making it quick to build scalable, highly accurate AI workflows just by passing a graph ID of your data as a parameter to your RAG tool.
📖 What We’re Reading
“GenAI is often seen as a miracle engine that promises to put early adopters on the track to greater efficiencies, lower costs, and quick wins. But most companies find that after a year of investing in endeavors that have underdelivered compared with expected results, the effect seems more like a collective spinning of wheels. The risk is that disappointed expectations can lead to disillusionment and lost momentum, when there are real benefits to be had.”
💻️ AI Tools and Platforms
LiveKit → Build real-time AI applications
Edge Impulse → AI for any edge device
Alinia → AI alignment for enterprises
Constructor → AI search and discovery
Synthflow → No-code AI phone calling