Contrary Research Rundown #103
Using the upcoming Cerebras IPO as a pulse on the semiconductor market, plus new memos on Monte Carlo, Alchemy, and more
Contrary Research celebrated its second anniversary this week, reaching over 1 million written words, 75K+ subscribers, and almost 400 companies covered!
Research Rundown
Cerebras Systems is an AI chip manufacturer known for creating the world's largest chip, the Wafer-Scale Engine (WSE). This chip is designed to address the inefficiencies of traditional GPUs in AI training by integrating memory directly within the processor, reducing latency and memory bottlenecks. This week, Contrary Research has put together a 12K word deep dive into Cerebras’ business, and how it represents a pulse on the broader semiconductor market!
The company's flagship product, the WSE, integrates memory directly within the GPU, addressing the memory bottleneck problem that traditional GPUs face. As a company, Cerebras doesn’t sell individual WSE chips but offers integrated computing systems that can be connected to form large clusters, providing near-linear performance scaling and significantly reducing the complexity of distributed training. For example, a single Cerebras system can deliver the computing performance of an entire room of servers with tens to hundreds of GPUs!
Cerebras has raised over $720 million in total funding from firms like Sequoia, Benchmark, and Coatue and, in August 2024, the company kicked off the process to work towards going public. With Nvidia skyrocketing in the public markets, an AI revolution everywhere you look, and Taiwan (and TSMC) increasingly in the geopolitical crosshairs, there is a huge swath of speculation about the broader semiconductor market. Cerebras sheds light on all of it.
One key consideration for Cerebras is its unique approach to AI computing. The company’s chips are designed specifically for AI workloads and offers superior performance and installation simplicity compared to traditional GPU-based systems. Cerebras' technology also allows for faster AI model training and inference, servicing the growing need for more efficient and powerful AI computing solutions.
The company’s success highlights the industry's shift towards specialized AI hardware, moving beyond traditional GPUs. In particular, Cerebras’ expansion into AI inference reflects the growing importance of efficient inference solutions in the AI ecosystem. And Cerebras’ focus on efficiency, speed, and simplified deployment aligns with the industry's key focus areas.
The shift from training to inference is one key example of how Cerebras as a company is like a pulse on the broader AI hardware market. By proving its capability to optimize models for specific hardware, Cerebras could become a crucial player in connecting AI training and inference processes, potentially improving deployment efficiency across the industry.
But any leading capability in AI is a moving target. In August 2024, Cerebras enabled inference capabilities on its latest systems, becoming the world’s fastest AI inference provider. Just a month later, in September 2024, both Groq and SambaNova made strides in faster and faster inference before Cerebras reclaimed the title. Suffice it to say, it's a rapidly evolving race.
And any breakthroughs in AI run up against the natural limits of the technology’s existing boundaries. On the technological front, Cerebras is grappling with both current limitations and future uncertainties. The company has reached the maximum possible chip size given ASML's EUV equipment constraints, potentially slowing future innovation to incremental improvements. This limitation is compounded by manufacturing challenges inherent in producing such large chips, increasing the probability of defects and adding complexity to the production process.
With the growth of AI model size and computational requirements straining data center power capacity, even leading to Microsoft pushing towards reopening Three Mile Island’s nuclear plant, these types of limitations are creating an infrastructure bottleneck that could slow the adoption of technology across the board. Whether Nvidia will see its products unseated by the likes of Cerebras and other emerging players, and how the technological infrastructure will shape the future of AI hardware, one thing is for certain:
Cerebras is like a mirror, reflecting each changing tailwind and potential pitfall as people rush to build the necessary hardware to take advantage of the promise of AI.
You can read the full memo on Cerebras here!
Cerebras Systems is an AI chip manufacturer that builds what the company claims to be “the world’s largest chip”. To learn more, read our full memo here and check out some open roles below:
AI Infrastructure Test Engineer - Sunnyvale, CA
Design Automation Engineer - Sunnyvale, CA
Monte Carlo is a platform for end-to-end data observability – a term coined by its CEO – to help organizations monitor abnormal patterns in their data. To learn more, read our full memo here and check out some open roles below:
Senior Backend Engineer - Remote (US)
Senior Frontend Engineer - Remote (US), Portland, Seattle, San Francisco, Vancouver, New York
Culdesac is a real estate developer and property management company that was founded with the goal of reimagining American cities to be built for people, not cars. To learn more, read our full memo here and check out some open roles below:
There are no job openings at the moment but you can check here.
Alchemy is a blockchain developer platform that offers the tools and infrastructure needed to build, scale, and rapidly iterate on blockchain applications. To learn more, read our full memo here and check out some open roles below:
Senior Software Engineer (API Infrastructure) - New York or San Francisco
Senior Full Stack Engineer (Wallet Services) - New York or San Francisco
Check out some standout roles from this week.
Warp | Remote (US and Canada) - Software Engineer, Senior Product Designer
Moment | New York City - Software Engineer (Backend, Trading - Multiple Levels), Software Engineer (API -Backend)
Rize | San Francisco - Founding Software Engineer (Product -All levels), Founding Software Engineer (Durable Workflows Platform - Senior / Staff level), Founding Software Engineer (Frontend Infrastructure - Senior / Staff level)
Endeavor | San Francisco - Senior Software Engineer, Artificial Intelligence Engineer, Forward Deployed Engineer
Moab | New York City - Business Operations Manager
Pylon | San Francisco - Customer Success Manager, Product Designer, (New Grad) Software Engineer, Software Engineer, Account Executive
Constellation signed a 20-year deal with Microsoft to restart the Three Mile Island nuclear plant and launch the Crane Clean Energy Center, providing carbon-free power to Microsoft's data centers.
Hippocratic AI, a company building generative AI-based healthcare agents to help address healthcare staffing shortages, recently raised an additional $17 million from Nvidia's venture arm, Greycroft, and 7Wire Ventures, as part of an extended Series A funding round.
OpenAI is raising $5-7 billion in a massive funding round, with a minimum investment of $250 million required from each investor, and tech giants like Microsoft, Nvidia, and Apple expected to contribute $2-3 billion in the deal.
In addition, OpenAI is considering removing the profit cap on its for-profit subsidiary, which would allow early investors to earn even bigger returns, but could raise questions about the company's non-profit mission.
Elon Musk claims the FAA is unfairly targeting SpaceX while neglecting safety issues at Boeing, putting astronaut lives at risk.
Bill Gates believes that if he were to start Microsoft again, he would focus the company on AI in order to rival industry leaders like OpenAI and Google, stating that "Today, somebody could raise billions of dollars for a new AI company [that's just] a few sketch ideas.”
LinkedIn is quietly using your profile data, posts, and other content to train its AI models, including those used for its various AI features - but EU users are automatically opted out of this data collection.
A breakup of Google's adtech business could "boost innovation in the space, reducing costs for advertisers and increasing revenue for publishers".
James Cameron expressed concerns about AI making it "hard to write science fiction" as the technology is advancing so rapidly, with ideas taking a minimum of 3 years to reach the screen.
Regent Seven Seas Cruises, a leading ultra-luxury cruise line, has finished installing Starlink on all six of its ships, providing guests with high-speed, unlimited WiFi as part of the voyage fare.
The Library of Congress, with its 180 million digital items, has become a "training data playground" for AI companies looking to develop and train their large language models without the risk of copyright infringement.
The global AI safety summit will include technical experts from each member's AI safety institute to discuss priority work areas and advance collaboration on AI safety.
Every single member of 23andMe's board of directors has resigned from the company.
OpenAI is actively monitoring and threatening to ban users who try to probe the inner workings and "reasoning trace" of its latest AI model, o1, in order to maintain a competitive advantage and control the narrative around its capabilities.
There is a strong case for challenger banks, that want to truly challenge the large incumbents, to become a licensed bank.
The Federal Reserve cut interest rates by a half-point, the first rate cut since the start of the pandemic.
ICONIQ published a 2024 State of AI Report which details the current landscape of AI with key takeaways for future outlook and potential.
Snap launched a new version of its Spectacles AR glasses. The company is not selling these new Spectacles directly to consumers. Instead, they are distributing them to AR developers who apply and pay a $99 monthly subscription fee.
A former Bytedance employee stated, “During my employment at ByteDance, American users’ content on TikTok was censored by ByteDance employees, including ByteDance’s China-based employees.”
The Pilot Tesla Semi fleet has driven over 7.5 million kilometers to date. A singular Tesla Semi has driven over 400,000 real-world kilometers in less than 18 months, all at the max gross weight limit.
Brendan Frey, the founder of Deep Genomics, offers a blunt assessment of AI's failures in drug discovery. "AI has really let us all down in the last decade when it comes to drug discovery…We've just seen failure after failure."
Elon Musk claims the Blindsight device from Neuralink could enable even those who have lost both eyes and their optic nerve to see, provided their visual cortex is intact, including those blind from birth.
If the five major tech companies (Google, Microsoft, Meta, Apple, and Amazon) were considered a country, their "location-based" emissions in 2022 would rank them as the 33rd highest-emitting country.
The cost of intelligence is plummeting. The price of GPT-4 equivalent intelligence has dropped ~240x in 18 months, from $180/million tokens to less than $1/million tokens.
Mumsnet, a popular UK parenting forum, is considering legal action against OpenAI for allegedly scraping its 6 billion-word dataset without permission.
Process supervision significantly outperforms outcome supervision for training models to solve problems from the challenging MATH dataset. Providing feedback on each intermediate reasoning step, rather than just the final result, is a more effective approach for training reliable AI models.
There are two scaling laws for AI — one for training model size, and one for the "thinking" process during inference. This means AI capabilities can continue to improve dramatically through both larger models and more "thinking" power, leading to rapid advancements in the coming years.
Huntress has reached $100M ARR while protecting over 120K SMBs.
Tesla produced their one-hundred millionth 4680 cell across all of their factories.
Contrary Research is growing our team, hiring a Research Analyst. Come help us build the best starting place to understand any private tech company. To apply, email research@contrary.com.
We’re thrilled to be hosting our first NYC Tech Talk of the year featuring eng leads and founders from Ramp, Warp, Railway, Together AI & Moment. It's an evening built by engineers for engineers — each company will live demo their latest product innovations for leading builders in NYC. Register here for a chance to join.
At Contrary Research, we’re building the best starting place to understand private tech companies. We can't do it alone, nor would we want to. We focus on bringing together a variety of different perspectives.
That's why we're opening applications for our Research Fellowship. In the past, we've worked with software engineers, product managers, investors, and more. If you're interested in researching and writing about tech companies, apply here!