Trusted Editorial content, reviewed by leading industry experts and seasoned editors. Ad Disclosure

American Bitcoin Corp., a mining company linked to US President Donald Trump’s family, has completed one of the largest hardware purchases in the crypto sector this year.

Reports have disclosed that the firm paid around $314 million for 16,290 Antminer U3S21EXPH units from Chinese mining giant Bitmain.

Massive Purchase Ahead Of Tariffs

The order locks in a huge amount of high-performance ASIC machines capable of producing about 14.02 exahashes per second in combined hashing power.

This level of output could boost American Bitcoin’s share of global mining capacity. The company originally had an option for as many as 17,280 units but decided to move fast to avoid price hikes from US tariffs on Chinese-made mining gear.

Industry trackers say orders of this scale have been rare for American miners in 2025. While the firm has not revealed where the rigs will be deployed, sources familiar with the matter said they plan to distribute them across multiple large sites to reduce operational risks.

Bitmain’s Antminers at Hut 8’s Vega facility in Texas. Source: The Miner Mag.

Political And Trade Pressures

The deal came just before the Trump administration began enforcing tariffs on imported Chinese mining hardware. The policy covers a wide range of technology goods, including ASIC miners, and is aimed at pushing production back to the US. However, critics say these tariffs could raise operating costs for domestic miners.

Jaran Mellerud, CEO of BTC mining firm Hashlabs, warned that higher costs could cut into profitability. He warned that steep price increases could raise mining costs in the US to a point where demand collapses, blaming regulators he viewed as ineffective.

Bitmain Eyes US Expansion

Bitmain, which controls about 80% of the global ASIC market according to a University of Cambridge study, is adjusting its operations in response to the tariffs.

The company plans to open its first US-based ASIC production site in early 2026. By the end of this year, it also expects to set up a new headquarters in Texas or Florida.

The aim is to make its products accessible to US customers at reasonable prices and escape import taxes by manufacturing locally.

Industry experts opine that this action would prompt other industry leaders such as MicroBT and Canaan to explore the possibility of diverting some production capacity to North America.

Although the complete implications of the tariffs on the mining supply chain are still uncertain, this recent acquisition indicates the depth of stakes for manufacturers and operators alike.

For American Bitcoin, the $314 million order indicates faith in the profitability of the industry in spite of fluctuating Bitcoin prices and increased competition.

For Bitmain, it’s an indication that being able to bend with political and economic gusts will be the ticket to maintaining its dominance of the US market.

Featured image from Pexels, chart from TradingView

Editorial Process for bitcoinist is centered on delivering thoroughly researched, accurate, and unbiased content. We uphold strict sourcing standards, and each page undergoes diligent review by our team of top technology experts and seasoned editors. This process ensures the integrity, relevance, and value of our content for our readers.

You can see it from the stock price: Not a lot of people are taking Advanced Micro Devices (AMD 3.35%) seriously right now as a competitor to its much larger rival, Nvidia. Even though many stocks involved in artificial intelligence (AI) have been hot buys, Advanced Micro Devices, also known as AMD, has been lagging behind.

Over the past 12 months, shares of AMD are down 20%, while Nvidia’s have risen by around 14% (as of June 23). And if you stretch out the time frame even further, the difference becomes wider.

AMD was a bit late in launching AI accelerators, but it may finally be closing the hardware performance gap and possibly even catching up to Nvidia. It recently announced its latest AI chips, and one person who is excited about them is none other than OpenAI CEO Sam Altman.

Image source: Getty Images.

Altman is a fan and will be a buyer of the new chips

AMD CEO Lisa Su unveiled the Instinct MI400 line of chips a few weeks ago. They’ll start shipping to customers next year. The chips can be tied together in a rack solution called “Helios,” which Su says “functions like a single, massive compute engine.” OpenAI, the company behind ChatGPT, will use the new chips, and Altman is excited about their potential. “It’s gonna be an amazing thing.”

The hope at AMD is that these new chips will take some market share from Nvidia, which has been far and away the leading AI chipmaker in the market thus far. With the MI400s earning a favorable endorsement from Altman, and considering AMD’s plans to price them aggressively, faster growth could be ahead for the chipmaker.

AMD’s growth rate has already been improving

In recent quarters, AMD’s results have been looking better and its growth rate has started to accelerate again.

AMD Operating Revenue (Quarterly YoY Growth) data by YCharts.

If its growth rate can remain strong, that could attract the types of AI investors who are specifically seeking out riskier stocks with lots of upside potential. AMD’s valuation of $210 billion is just a fraction of Nvidia’s $3.5 billion market cap. But while Nvidia has been a beast in the AI sphere, investors today may be wondering just how much higher it can go given that it’s already among the three most valuable companies in the world. This is where AMD may be a more compelling option.

However, for AMD to win over investors, it will need to not only grow its sales but its bottom line as well; currently, the stock trades at more than 90 times its trailing earnings, and even its forward earnings multiple (based on analysts’ consensus estimates) is 34, which is still a bit high. The S&P 500‘s forward price-to-earnings ratio is 23.

Setting prices for its AI chips at aggressively low levels could help them win market share, but it may also hurt AMD’s margins, which is why it will be imperative for investors to pay close attention to AMD’s gross margins to see if they are declining as it rolls out its newest chips.

Is AMD stock a buy right now?

Shares of AMD have been rallying over the past few months, and the business has a lot of potential to rise in value over the years. By working closely with Altman and the OpenAI team, and gaining feedback on its chips, AMD can ensure that its products meet the needs of key customers and hyperscalers.

AMD looks like a good stock to buy right now as its growth rate has been improving, and the best may still be to come as it starts rolling out its Instinct MI400 chips. Although its margins may take a hit in the near term, proving that its chips can be formidable alternatives to Nvidia’s chips will arguably be much more important to AMD in the long run.

Auradine Inc., a US-based Bitcoin miner manufacturer, announced today that it will announce its expanded portfolio of mining products at the Bitcoin 2025 conference in Las Vegas.

“Our goal is to democratize access to Bitcoin mining and enable innovative integration,” said CEO and co-founder of Auradine Rajiv Khemani. “Whether you’re running a megawatt container or building a small form factor heater miter miter for your home, it will help you succeed by providing tips, systems and support. This new chapter is about providing efficient innovation, expansion and operations to miners.”

The new ASIC offering, designed for both industrial and small deployments, supports customizable form factors and is already adopted by operators such as Mara Holdings, Futurebit, and Deep South Operating. Along the chip, Auragin continues to produce a wide range of mining rigs to support a variety of deployment needs.

“Aurazine’s ability to provide both high-performance chips and scalable infrastructure is consistent with Mala’s mission to stay at the forefront of Bitcoin mining,” said Mala Holdings Chief Technology Officer Ashu Swami. “We are pleased with our partnership with Aurazine with our cutting-edge engineering capabilities and innovation.”

Auradine’s modular 1 MW container units were developed in collaboration with Fog Hashing and Fbox and are designed to accommodate 100-200 miners each. Merkle Standard, the first to deploy the system, reported increased energy efficiency and operational flexibility.

“We were the first to deploy Aurazine container solutions and quickly exceeded our expectations,” said the COO of Merkle Standard Monty Stahl. “The combination of performance, energy efficiency and modular design gives you the flexibility to expand operations faster and smarter than traditional infrastructures allow. This is a kind of innovation that mining has long required.”

The recent $153 million Series C funding provides flexible mining infrastructure and supports the push to supply ASIC chips for third-party integration. The company also plans to extend its hardware expertise to AI and network through its Auralinks initiative.

“We were one of the first to try out Aurazine’s ASIC chips and were immediately impressed with the support and customizations our team provided,” added LLC Brock Tompkins, CEO of Deep South Operating. “It helps miners like us to be scalable and efficient while raising the standard for what decentralized mining looks like.”

Forward-looking: Apple’s M5 chip is expected to make its debut in the MacBook Pro and iPad Pro later this year, but Apple is already deep into development on the M6 and M7 processors. Apple is advancing its custom silicon ambitions, with a sweeping roadmap that spans next-generation Macs, its first smart glasses, and a powerful new AI server chip. Sources familiar with the company’s plans told Bloomberg that Apple is building on its recent momentum in chip design, aiming to set new standards in performance and efficiency across its entire product lineup.

Apple will continue to invest heavily in custom silicon as it strives to lead not only in device performance but also in AI and wearable technology – two areas where it has attempted to make advances but has faced various challenges. The company is developing new chips to power everything from next-generation Macs to smart glasses and AI servers, aiming to maintain its tightly integrated hardware and software ecosystem.

The Apple M6 SoC, known internally as “Komodo,” is likely to arrive in 2026, followed by the M7, codenamed “Borneo,” in 2027, if Apple maintains its annual release cadence. These chips are rumored to represent a major leap in computing power, with configurations as high as 256 CPU cores and 640 GPU cores in future generations – an exponential increase compared to the current M3 Ultra, which tops out at 32 CPU and 80 GPU cores.

Alongside these chips, Apple is reportedly working on an advanced system-on-chip for Macs, codenamed “Sotra.” While details remain scarce, this chip could be destined for high-end desktop models, further diversifying Apple’s silicon offerings and targeting professional users who want top-tier performance.

In a parallel effort, Apple is preparing to enter the still nascent but competitive smart glasses market. The company’s first glasses, powered by a specialized chip called N401, are being designed to rival Meta’s Ray-Ban smart glasses.

Drawing on the energy-efficient architecture of the Apple Watch, the N401 chip is optimized for low power consumption and is engineered to control multiple cameras integrated into the glasses.

These non-augmented reality glasses will feature cameras, microphones, and on-board AI, enabling users to take photos, record video, translate languages, and receive contextual information about their surroundings.

Features under consideration include environmental scanning, object identification, and navigation assistance. The glasses’ processor is targeted for mass production in late 2026 or 2027, suggesting a product launch within the next 2 to 4 years.

Meanwhile, Apple’s ambitions for AI are expanding into the realm of large-scale infrastructure. In partnership with Broadcom, the company is developing its first AI server chip, codenamed “Baltra,” which is slated to enter mass production as early as 2026 or 2027.

This chip is being designed to handle Apple Intelligence requests in the cloud, replacing the current reliance on Mac-grade processors such as the M2 Ultra.

Apple is reportedly testing Baltra in configurations with 2, 4, 6, or even 8 times the CPU and GPU cores of the M3 Ultra, meaning the chip could boast up to 256 CPU cores and 640 GPU cores at its highest. Such a leap could position Apple as a major player in AI infrastructure, enabling faster and more efficient cloud-based intelligence for its devices.

These developments are part of a broader strategy within Apple’s hardware technology group, which is also working on chips for camera-equipped AirPods and Apple Watch models, with launches targeted around 2027.

The company’s silicon roadmap also includes upgrades to its in-house modem technology, with the C2 and C3 modems expected to improve wireless connectivity in future iPhones.

Masthead credit: Basic Apple Guy

Qualcomm is nearing a big decision. It’s reportedly gearing up to release a new, midrange Snapdragon chip. While this “budget” SoC is rumored to be powered by Arm cores — not the custom Oryon cores powering other Elite chips — it’s unclear whether the company will still use “Elite” branding. We’ve seen reports use both Snapdragon 8s Elite and Snapdragon 8s Gen 4 to refer to the unannounced mobile chip.

The problem? The custom Oryon CPU cores are what make Snapdragon 8 Elite chips, well, “elite.” Let’s review how we got here.

The switch from Arm to Oryon

Last year, Qualcomm finally achieved its breakthrough moment. It launched the Snapdragon X lineup of laptop and desktop-class systems-on-a-chip (SoCs) in June 2024 that would eventually disrupt the stronghold competitors AMD and Intel had on personal computing. Then, Qualcomm used what it learned to develop the Snapdragon X platform and brought it to mobile with the Snapdragon 8 Elite chipset.

By now, I’ve had the chance to review and test multiple Snapdragon X and Snapdragon 8 Elite laptops and phones, respectively. They’re both as good as advertised. Snapdragon X brought excellent power efficiency and responsiveness to Windows laptops, and Snapdragon 8 Elite brought massive year-over-year performance improvements to Android smartphones.

Oryon cores are the secret to Qualcomm’s major success. The company used its own custom Oryon cores instead of off-the-shelf designs from Arm to make the Snapdragon X, Snapdragon X Plus, and Snapdragon X Elite SoCs. Essentially, it followed the same blueprint Apple used when designing its own silicon.

A quick look at Qualcomm’s PC processors reveals that chips with Oryon cores don’t always use the Elite brand. That’s reserved for Qualcomm’s best PC chips, while budget versions with slower clock speeds and/or fewer cores use names like Snapdragon X and Snapdragon X Plus.

But we’ve never seen a Qualcomm chip branded with the Elite tag that doesn’t use custom Oryon cores. Although there are conflicting reports about whether the next Snapdragon 8s mobile processor will get a “Gen 4” or “Elite” brand name, the right option seems clear.

Qualcomm should name the unreleased midrange chip Snapdragon 8s Gen 4 to avoid consumer confusion and keep the Elite brand sacred for the company’s standout processors — those with Oryon cores.

What will Qualcomm’s next midrange chip be called?

Right now, there’s no official word on what Qualcomm will call the next Snapdragon 8s chip. One rumor from reputable leaker Yogesh Brar on X (formerly Twitter) earlier this month said that the SoC was “set to launch soon,” referring to it as Snapdragon 8s Elite. More recently, a tip from Digital Chat Station this week on Weibo called the unreleased chip the Snapdragon 8s Gen 4.

Although both rumored names are only speculation at this point, one seems more fitting than the other.

That’s because current leaks claim to reveal the upcoming Snapdragon 8s chip’s core structure, and custom Oryon cores are absent. According to Digital Chat Station, this processor has a model number of SN8735. It is said to be headlined by a 3.21GHz Cortex-X4 Prime Core, plus three more 3.01GHz Cortex-A720 cores and two 2.80GHz Cortex-A720 cores. The SoC is also expected to include two 2.02GHz Cortex-A720 efficiency cores.

If this configuration turns out to be accurate, the Snapdragon 8s chip will use standard Arm designs rather than Qualcomm’s best Oryon cores. This isn’t a bad thing, because the Snapdragon 8 Elite chip is expensive, and the “s” variant is supposed to be cheaper. However, it does mean that the Elite branding wouldn’t exactly make sense for this particular SoC.

Qualcomm is rumored to make the right move

Qualcomm is a very large company operating in many different verticals, from cellular modems to smartphone and desktop chips. Under those circumstances, it’s easy for naming schemes to get convoluted. Just look at what Intel and Sony have done over the years, releasing products like the Intel Core i9 14900K or the Sony WH-1000XM5.

Luckily, Qualcomm seems to be aware of how important branding is for consumer clarity. At MWC Barcelona 2025, it introduced a new naming scheme for its vast portfolio of products. Qualcomm will be used for general products, like modems, with the Snapdragon brand serving consumer products and the Dragonwing brand covering the enterprise market.

Qualcomm may be set up for an interesting decision. On the one hand, using the name Snapdragon 8s Elite would make it clear that the to-be-released chip is part of the same generation as the Snapdragon 8 Elite. On the other hand, using the name Snapdragon 8s Gen 4 would make it clear that the chip isn’t using custom Oryon cores.

Only time will tell which name Qualcomm will end up using. Personally, I’m voting for the next Snapdragon 8s SoC to be called the Snapdragon 8s Gen 4. While the Elite nomenclature is an excellent brand, it’s too good to be diluted. Currently, every Elite chip features Qualcomm’s custom Oryon cores, and I think it should stay that way.

Big if true: The rumors about Arm’s ambitions to enter the chip manufacturing business appear to be true after all. Industry sources now claim the UK-based designer is in need of a highly skilled workforce and has begun poaching top talent from clients.

Arm Holdings is currently on a recruiting spree, with the ultimate goal of entering the chip production industry independently. Reuters has quoted several sources familiar with the matter, all of whom confirm Arm’s “sudden” interest in the manufacturing side of the microchip industry.

Arm’s core business revolves around licensing its namesake instruction set architecture to other silicon designers and chip foundries around the world. The company has been around since the 1980s, initially selling home computers with RISC processors under the Acorn Computers brand. Virtually all modern smartphones feature processors based on the Arm ISA, but Arm earns money through licensing deals, rather than from actual device sales.

The UK-based designer is clearly trying to change the status quo. Two Reuters sources revealed that Arm has been reaching out to its licensing partners, attempting to recruit executives and potential employees from other positions to help build its new chipmaking business. The news agency was able to examine a copy of a “note” sent by an Arm recruiter to an executive at one of its customers.

The note states that Arm wants to transform its licensing-only business model into a company that sells its own silicon products. The new chips would focus on “AI enablement in the data center” and other computing devices, according to the note. Arm has also reached out to other chip designers in Silicon Valley, seeking to hire new talent and experts in the area.

Over the past few years, Arm has adopted a much more proactive approach to the chipmaking business. In 2022, the company sued Qualcomm to terminate their licensing agreement, but was later forced to abandon the effort. During the legal dispute over Qualcomm’s licensing rates, Arm CEO Rene Haas stated that the company had never been involved in building chips.

Sources say Arm began attempting to hire new managers from its customers as early as November. Haas’ testimony occurred just a few weeks later. Regardless, Arm’s attempt to start designing and selling its own chips could send a shockwave through the entire mobile industry.

Historically, Arm Holdings has been considered a neutral player in the market, while silicon providers like Broadcom and Qualcomm have raked in billions by developing Arm-based custom CPUs. If Arm were to abandon its neutral stance, the chipmaking industry might react in unpredictable ways.