DeepSeek has rattled the U.S.-led AI ecosystem with its latest model, shaving hundreds of billions in chip leader Nvidia’s market cap. While the sector leaders grapple with the fallout, smaller AI companies see an opportunity to scale with the Chinese startup.
Several AI-related firms told CNBC that DeepSeek’s emergence is a “massive” opportunity for them, rather than a threat.
“Developers are very keen to replace OpenAI’s expensive and closed models with open source models like DeepSeek R1…” said Andrew Feldman, CEO of artificial intelligence chip startup Cerebras Systems.
The company competes with Nvidia’s graphic processing units and offers cloud-based services through its own computing clusters. Feldman said the release of the R1 model generated one of Cerebras’ largest-ever spikes in demand for its services.
“R1 shows that [AI market] growth will not be dominated by a single company — hardware and software moats do not exist for open-source models,” Feldman added.
Open source refers to software in which the source code is made freely available on the web for possible modification and redistribution. DeepSeek’s models are open source, unlike those of competitors such as OpenAI.
DeepSeek also claims its R1 reasoning model rivals the best American tech, despite running at lower costs and being trained without cutting-edge graphic processing units, though industry watchers and competitors have questioned these assertions.
“Like in the PC and internet markets, falling prices help fuel global adoption. The AI market is on a similar secular growth path,” Feldman said.
Inference chips
DeepSeek could increase the adoption of new chip technologies by accelerating the AI cycle from the training to “inference” phase, chip start-ups and industry experts said.
Inference refers to the act of using and applying AI to make predictions or decisions based on new information, rather than the building or training of the model.
“To put it simply, AI training is about building a tool, or algorithm, while inference is about actually deploying this tool for use in real applications,” said Phelix Lee, an equity analyst at Morningstar, with a focus on semiconductors.
While Nvidia holds a dominant position in GPUs used for AI training, many competitors see room for expansion in the “inference” segment, where they promise higher efficiency for lower costs.
AI training is very compute-intensive, but inference can work with less powerful chips that are programmed to perform a narrower range of tasks, Lee added.
A number of AI chip startups told CNBC that they were seeing more demand for inference chips and computing as clients adopt and build on DeepSeek’s open source model.
“[DeepSeek] has demonstrated that smaller open models can be trained to be as capable or more capable than larger proprietary models and this can be done at a fraction of the cost,” said Sid Sheth, CEO of AI chip start-up d-Matrix.
“With the broad availability of small capable models, they have catalyzed the age of inference,” he told CNBC, adding that the company has recently seen a surge in interest from global customers looking to speed up their inference plans.
Robert Wachen, co-founder and COO of AI chipmaker Etched, said dozens of companies have reached out to the startup since DeepSeek released its reasoning models.
“Companies are [now] shifting their spend from training clusters to inference clusters,” he said.
“DeepSeek-R1 proved that inference-time compute is now the [state-of-the-art] approach for every major model vendor and thinking isn’t cheap – we’ll only need more and more compute capacity to scale these models for millions of users.”
Jevon’s Paradox
Analysts and industry experts agree that DeepSeek’s accomplishments are a boost for AI inference and the wider AI chip industry.
“DeepSeek’s performance appears to be based on a series of engineering innovations that significantly reduce inference costs while also improving training cost,” according to a report from Bain & Company.
“In a bullish scenario, ongoing efficiency improvements would lead to cheaper inference, spurring greater AI adoption,” it added.
This pattern explains Jevon’s Paradox, a theory in which cost reductions in a new technology drive increased demand.
Financial services and investment firm Wedbush said in a research note last week that it continues to expect the use of AI across enterprise and retail consumers globally to drive demand.
Speaking to CNBC’s “Fast Money” last week, Sunny Madra, COO at Groq, which develops chips for AI inference, suggested that as the overall demand for AI grows, smaller players will have more room to grow.
“As the world is going to need more tokens [a unit of data that an AI model processes] Nvidia can’t supply enough chips to everyone, so it gives opportunities for us to sell into the market even more aggressively,” Madra said.
A logo hangs on the building of the Beijing branch of Semiconductor Manufacturing International Corporation (SMIC) on December 4, 2020 in Beijing, China.
After trading on Thursday, the company reported a first-quarter revenue of $2.24 billion, up about 28% from a year earlier. Meanwhile, profit attributable to shareholders surged 162% year on year to $188 million.
However, both figures missed LSEG mean estimates of $2.34 billion in revenue and $225.1 million in net income, as well as the company’s own forecasts.
During an earnings call Friday, an SMIC representative said the earnings missed original guidance due to“production fluctuations” which sent blended average selling prices falling. This impact is expected to extend into the second quarter, they added.
For the current quarter, the chipmaker forecasted revenue to fall 4% to 6% sequentially. Gross margin is also expected to fall within the range of 18% to 20%, compared to 22.5% in the first quarter.
Still, the first quarter saw SMIC’s wafer shipments increase by 15% from the previous quarter and by about 28% year-on-year.
In the earnings call, SMIC attributed that growth to customer shipment pull in, brought by changes in geopolitics and increased demand driven by government policies such as domestic trade-in programs and consumption subsidies.
In another positive sign for the company, its first-quarter capacity utilization— the percentage of total available manufacturing capacity that is being used at any given time— reached 89.6%, up 4.1% quarter on quarter.
“SMIC’s nearly 90% utilization rate reflects strong domestic demand for semiconductors, likely driven by smartphone and consumer electronics production,” said Ray Wang, a Washington-based semiconductor and technology analyst, adding that the demand was also reflected in the company’s strong quarterly revenue growth.
Meanwhile, the company said in the earnings call that it is “currently in an important period of capacity construction, roll out, and continuously increasing market share.”
However, SMIC’s first-quarter research and development spending decreased to $148.9 million, down from $217 million in the previous quarter.
Amid increased demand, it will be crucial for SMIC to continue ramping up their capacity, Simon Chen, principal analyst of semiconductor manufacturing at Informa Tech told CNBC.
SMIC generates most of its revenue from older-generation semiconductors, often referred to as “mature-node” or “legacy” chips, which are commonly found in consumer electronics and industrial equipment.
The state-backed chipmaker is critical to Beijing’s ambitions to build a self-sufficient semiconductor supply chain, with the government pumping billions into such efforts. Over 84% of its first-quarter revenue was derived from customers in China.
“The localization transformation of the supply chain has been strengthened, and more manufacturing demand has shifted back domestically,” a representative said Friday.
However, chip analysts say the chipmaker’s ability to increase capacity in advance chips — used in applications that demand higher levels of computing performance and efficiency at higher yields — is limited.
This is due to U.S.-led export controls, which prevent it from accessing some of the world’s most advanced chip-making equipment from the Netherlands-based ASML.
Nevertheless, the chipmaker appears to be making some breakthroughs. Advanced chips manufactured by SMIC have reportedly appeared in various Huawei products, notably in the Mate 60 Pro smartphone and some AI processors.
In the earnings call, the company also said it would closely monitor the potential impacts of the U.S.-China trade war on its demand, noting a lack of visibility for the second half of the year.
Phelix Lee, an equity analyst for Morningstar focused on semiconductors, told CNBC that the impacts of U.S. tariffs on SMIC are limited due to most of its revenue coming from Chinese customers.
While U.S. customers make up about 8-15% of revenue on a quarterly basis, the chips usually remain and are consumed in Chinese products and end users, he said.
“There could be some disruption to chemical, gas, and equipment supply; but the firm is working on alternatives in China and other non-U.S. regions,” he added.
SMIC’s Hong Kong-listed shares have gained over 32.23% year-to-date.
Close-up of a hand holding a cellphone displaying the Amazon Pharmacy system, Lafayette, California, September 15, 2021.
Smith Collection | Gado | Getty Images
Amazon is expanding its online pharmacy to fill prescription pet medications, the company announced Thursday.
The company said it has added “hundreds of commonly prescribed pet medications” to its U.S. site, ranging from flea and tick solutions to treatments for chronic conditions.
Prescriptions are purchased via Amazon’s storefront and must be approved by a veterinarian. Online pet pharmacy Vetsource will oversee the dispensing and delivery of medications, said Amazon, adding that items are typically delivered within two to six days.
Amazon launched its digital drugstore in 2020 with the added perk of discounts and free delivery for Prime members. The company has been working to speed up prescription shipments over the past year, bringing same-day delivery to a handful of U.S. cities. Last October, Amazon set a goal to make speedy medicine delivery available in nearly half of the U.S. in 2025.
The new pet medication offerings puts Amazon into more direct competition with online pet pharmacy Chewy, as well as Walmart, which offers pet prescription delivery.
Amazon Pharmacy is part of the company’s growing stable of healthcare offerings, which also includes One Medical, the primary care provider it acquired for roughly $3.9 billion in July 2022. Amazon’s online pharmacy was born out of the company’s 2018 acquisition of online pharmacy PillPack.
Coinbase agreed to acquire Dubai-based Deribit, a major crypto derivatives exchange, for $2.9 billion, the largest deal in the crypto industry to date.
The company said Thursday that the cost comprises $700 million in cash and 11 million shares of Coinbase class A common stock. The transaction is expected to close by the end of the year.
Shares of Coinbase rose nearly 6%.
The acquisition positions Coinbase as an international leader in crypto derivatives by open interest and options volume, Greg Tusar, vice president of institutional product, said in a blog post – which could allow it take on big players like Binance. Coinbase operates the largest marketplace for buying and selling cryptocurrencies within the U.S., but has a smaller share of the global crypto market, where activity largely takes place on Binance.
Deribit facilitated more than $1 trillion in trading volume last year and has about $30 billion of current open interest on the platform.
“We’re excited to join forces with Coinbase to power a new era in global crypto derivatives,” Deribit CEO Luuk Strijers said in a statement. “As the leading crypto options platform, we’ve built a strong, profitable business, and this acquisition will accelerate the foundation we laid while providing traders with even more opportunities across spot, futures, perpetuals, and options – all under one trusted brand. Together with Coinbase, we’re set to shape the future of the global crypto derivatives market.”
Tusar also noted that Deribit has a “consistent track record” of generating positive adjusted EBITDA the company believes will grow as a combined entity.
“One of the things we liked most about this deal is that it’s not just a game changer for our international expansion plans — it immediately diversifies our revenue and enhances profitability,” Tusar told CNBC.
The deal comes at a time when the crypto industry is riding regulatory tailwinds from the first ever pro-crypto White House. Support of the industry has fueled crypto M&A activity in recent weeks. In March, crypto exchange Kraken agreed to acquire NinjaTrader for $1.5 billion, and last month Ripple agreed to buy prime broker Hidden Road.
Don’t miss these cryptocurrency insights from CNBC Pro: